In this episode of Eureka!, McKinsey\u2019s podcast on innovation in life sciences, host Navraj Nagra and McKinsey partner Yigit Yorulmaz speak with Dr. Alex Condoleon, chief medical affairs officer for global medical engagement and impact at Pfizer. Condoleon, a physician executive who has led medical, innovation, and digital health efforts across Australia, China, Europe, and the United States, argues that medical affairs\u2019 longstanding ceiling has lifted and that the function can now be a genuine strategic partner in life sciences. He discusses what it takes to bring the right evidence to the right clinicians at the right moment, how to measure medical affairs\u2019 impact beyond commercial metrics, and what patient-first looks like at a global scale across a wide portfolio of disease areas. An edited version of their conversation follows.
\nA career built anticipating the industry\u2019s next move
\nNavraj Nagra: You\u2019ve had a distinctive journey into medical affairs, including leadership roles spanning medical, innovation, and digital healthcare. What were the pivotal moments that shaped your career direction, and what ultimately pulled you into this role at Pfizer?
\nAlex Condoleon: Looking back at the career trajectory I\u2019ve been on, I\u2019ve always believed that if you want to impact global health in a meaningful way, you need to get a step ahead of where the system is already going, not follow it. That has guided each of my career moves.
\nI started in the industry in Australia and moved to China. That wasn\u2019t because I had clarity on what China would become; it was because I had a conviction that China would play a significant role in healthcare. Fast-forward to today, and China\u2019s ambition is becoming clear in biotech and in AI, and that shift is going to impact global health.
\nFrom there, I moved to Europe, into my first dedicated role focused on consumer healthcare rather than the prescription business. That was me seeking to understand how patients were starting to take control of their health and what that could mean. In that role, almost a decade ago, I was seeing early direct-to-patient models. I remember meeting the team at what today we call Ro.com when they were in their infancy, and yet there was something about what they were building toward that I thought was going to matter. Fast-forward to today and we see platforms like Hims & Hers and Ro.com playing a dominant role in how people choose to access glucagon-like peptide-1 (GLP-1) drugs in the United States.
\nMy next pivot was moving to the United States, focused on enterprise digital transformation in medical affairs, and later building out a digital healthcare franchise. Again, it\u2019s a similar pattern: investing ahead of a change I believed was coming, so I could be ready for the moment. I think that moment is the one we\u2019re living through now, a moment of convergence. Many global powers are influencing what the healthcare landscape looks like; digital and AI are rising in healthcare delivery; and all of that is pushing the trend of patient empowerment and patient-guided care. When you see that kind of seismic opportunity come together, it\u2019s great to be in an organization like Pfizer, which seeks to do bold things, because being bold in healthcare is your way of staying relevant and achieving leadership.
\nI\u2019ve been very fortunate to come into a position focused on the role of data, analytics, and AI in optimizing how we think about medicine, science, and value, and how we communicate it and bring it to the forefront. I\u2019m hopeful those things make a meaningful difference to moving healthcare forward.
\nWhy the ceiling on medical affairs is gone
\nYigit Yorulmaz: Your remit in medical affairs involves overseeing a high-caliber global platform focused on medical engagement and impact. How does this, and your past experience, help you reimagine what medical affairs can look like? What excites you about building and leading a function that doesn\u2019t exist at any scale in many medical affairs organizations?
\nAlex Condoleon: I\u2019m excited by the role I get to do every day. Part of the reason is that I\u2019d reached a point where I genuinely felt medical affairs had hit a ceiling. What energizes me today is that I\u2019m more convinced than ever that ceiling is gone, and that medical affairs has the potential to be a true strategic partner in life sciences. We\u2019re in the process, as a function, of stepping into that role, and I think now is the moment.
\nThere are a few reasons for that belief. It starts with how healthcare systems, and the experience of healthcare professionals and patients engaging with the system, are fundamentally changing. In a world where the lives of healthcare professionals are chaotic and made more complex every day, the need for accurate, relevant, timely signals to guide how they manage patients is more important than ever. But it must be approached differently than in the past. That\u2019s exciting, but it will also challenge us: Compared to a decade ago, we need to reimagine our role, reinvent how we bring value to our target customer groups, and constantly listen to insight and feedback that helps us improve.
\nI say that partly because, as an industry, we\u2019re in a period of great productivity, with lots of new products coming to market. But there\u2019s a paradox: Every time we launch a new product and solve an unmet need, we also unintentionally add complexity to an already complex treatment landscape. That pushes the importance of thinking about what information healthcare professionals and patients need, and when. That\u2019s where there\u2019s an exciting role for digital platforms to enhance and magnify what we aspire to do, enriched by the AI advances we\u2019re seeing now.
\nSo although we\u2019re challenged to lead through a period where the pace of external change feels like it\u2019s outstripping our internal capacity to absorb it, that\u2019s fundamentally a world of opportunity for medical affairs teams. But to grab that opportunity, building the right data, analytics, and AI infrastructure is critical. We\u2019re really good with data in medical affairs, good at making sense of incomplete and imprecise data. We just need different techniques, and to get comfortable with the learning curve.
\nIf we can build those fundamentals into our capability mix, and lead in a way that resonates with teams, I think the way the external world and technology are advancing means something important for medical, and represents the pillars on which we\u2019ll step up as a function.
\nReaching physicians in a more complex landscape
\nNavraj Nagra: Just to add a further deep dive here: You mentioned the use of some of these technologies, these new ways of working, to better engage physicians and medical professionals. Can you give me an example of what you\u2019re doing at the moment to help physicians think about this more complex landscape, and how you\u2019re engaging with them differently compared to what was being done even ten or 20 years ago?
\nAlex Condoleon: That\u2019s a really important question. Given the data now available, we can, with a good degree of precision, understand the clinical realities physicians are facing, including where there are gaps in care that represent challenges and opportunities for medical teams working with the patients they oversee. Once we understand that, we have a need state we can seek to solve.
\nWe\u2019ve built the data and analytics capabilities that give us precision on who\u2019s facing a clinical challenge that aligns with where we think there\u2019s room for improvement. At the same time, we\u2019ve been able to be precise about the digital footprint of those healthcare professionals, which lets us rethink what content we need and how to make it relevant to the channels physicians are actually on. That\u2019s half the equation.
\nWhat\u2019s equally important is that there\u2019s now maturity in the healthcare landscape, letting us partner with medically focused digital channels to deploy content precisely, in the right form and at the right time, for target physicians. By bringing together the right customer segment, the right need state, and content adapted to how information is digested on the right channels, and by partnering with these platforms, we\u2019ve seen tangible improvement in care. That improvement doesn\u2019t replace what our field medical teams do, it\u2019s supplemental. It increases reach and magnifies what we\u2019re already doing through impactful field conversations.
\nThat\u2019s something we couldn\u2019t have fully believed two years ago would be possible, but it is and it works. It\u2019s an example of taking the discipline of medical affairs and matching it to the need state of the healthcare professionals you\u2019re trying to reach. Our traditional face-to-face work can now be magnified and elevated beyond that, and we\u2019ll only see acceleration of it.
\nOne portfolio, many specialized environments
\nNavraj Nagra: One thing we recognize is that there\u2019s such a wide portfolio across Pfizer, from vaccines to primary care to specialty medicines. When you think about patient needs, physician needs, and the scientific conversations across these categories, how does that vary? What are the differences in maturity of the approaches you\u2019ve discussed across those categories?
\nAlex Condoleon: That\u2019s a fascinating question, and one of the things I\u2019m really fortunate about in my role. Working across a portfolio like Pfizer\u2019s, you start to see different elements of the future of healthcare emerge, in different ways, in different specialties. One day I\u2019m hopeful they\u2019ll converge.
\nTake oncology, or rare diseases, where Pfizer has a significant focus. We\u2019re seeing the rapid rise of a precision, science-driven form of personalization. It was only a handful of decades ago that we defined cancers by their organ of origin, but today we have hundreds of subtypes based on histology, genetics, and biomarkers, and we can envisage further precision coming. We can take the same parallel with heart failure, which has been segmented as we understand more of the genetic contributors. So there\u2019s precision emerging in some areas, and it will only grow.
\nThen at another end of the spectrum, look at vaccines, or an area like GLP-1s, where the dynamic is very different: You\u2019re fundamentally engaging with the population and shaping patient choice and behavior. Patient literacy, ease of access, and convenience matter enormously to the choices people make. And then there\u2019s primary care and specialty care, where we\u2019re focused on understanding healthcare professional preferences, beliefs, practice patterns, and care gaps, so we can be as targeted and relevant as possible.
\nI call those out because one day those three thematics will converge into a great healthcare experience, where the science of precision spans all of healthcare, where people navigate their own healthcare guided by their choices, preferences, and literacy, and where the industry has learned to be as relevant and personalized as possible to the attitudes and behaviors of the healthcare professionals we engage with.
\nSo the breadth of our portfolio is a landscape of learning, one I\u2019m hopeful will lead to a convergence that brings more value to patients and healthcare professionals alike. Today you can\u2019t treat the portfolio as one-size-fits-all, but you can treat it as several test-and-learn environments, where concepts proven in one area show us when others will benefit from the same trends.
\nWhat to standardize, what to keep local
\nYigit Yorulmaz: Fascinating to hear you talk about that. It\u2019s absolutely true that different parts of healthcare are evolving in different ways, but when the portfolio is so diverse, medical affairs can\u2019t rely on a single playbook to determine what good looks like. So you have, as a leader, the choice of deciding what must be standardized\u2014whether platform capabilities, principles, or how you measure things\u2014and what must be bespoke: the content you use, how you\u2019re thinking about channel mix. How do you think about that variation across therapeutic areas or geographies, based on your experience? What\u2019s standardized, and what\u2019s unique?
\nAlex Condoleon: There\u2019s no easy answer to that. As we build new capabilities for medical affairs\u2014especially data, analytics, and AI capabilities\u2014scale matters. On the flip side, being relevant to different therapeutic areas or geographies also matters. Bridging that tension is a day-to-day exercise in judgment.
\nThere are areas where standardization and scale matter most. How we structure and govern data is best done with discipline at an enterprise level. The same goes for the analytics platforms and new professional skills we\u2019re bringing into teams; while we\u2019re maturing them and working out what value looks like, there\u2019s benefit to housing that learning centrally. The same is true for how we build AI models and partner with large language model (LLM) companies: At this early-learning stage, that fits into working out the foundational capabilities that represent the workforce of the future.
\nBut none of that matters if we can\u2019t make it relevant to the people we\u2019re seeking to bring value to. Central capabilities can be quicker to build in transforming moments, but ultimately they need to sit as close to the customer as possible. That means working closely with the teams you\u2019re partnering with, not being overprescriptive about the offering, and building understanding of a capability and a best practice while giving people the freedom to innovate beyond it. That\u2019s how we keep capabilities relevant, especially to medical stakeholders.
\nWe\u2019re in build mode right now. We might have ideas of a target operating model, but the key word is target, not today\u2019s state. Organizations aren\u2019t static, they\u2019re living things, and it\u2019s when we let that chemistry play out that we innovate beyond what\u2019s apparent today. So there\u2019s a place, especially in this digitally accelerated phase, for building capabilities that don\u2019t yet exist. But if we can\u2019t equally let teams apply those capabilities with freedom to operate in their own context, we\u2019ll lose value and struggle to translate healthcare know-how and tech know-how into medical affairs impact.
\nGoverning AI and digital initiatives
\nNavraj Nagra: You\u2019ve described what sounds like a plethora of AI and digital initiatives to drive medical engagement and impact. How are you governing the decision-making around this? How are you deciding where the ROI is good enough to give your teams the freedom to operate in that space, and where you don\u2019t believe there\u2019s going to be impact from certain digital interventions? What\u2019s the operating model behind these digital and AI initiatives?
\nAlex Condoleon: Firstly, executive sponsorship really matters when you\u2019re going through big transformative change like this. That\u2019s because there\u2019s a lot we don\u2019t know; we\u2019re going to make mistakes and learn through failing. If the goal is worth heading toward, we\u2019ll tolerate the learning that comes with it.
\nBeyond executive sponsorship, the operating model has to adapt to what you\u2019re building, especially when capabilities in digital and AI are emergent rather than matured. That means wanting as many stakeholders vested in the journey as possible, while respecting that your priority might not be theirs. That\u2019s why high-visibility steering groups matter; they give stakeholders a chance to stay close, understand where the transformation is heading, and be heard. At every layer of leadership, it matters, and wherever the bottleneck comes up, you need to understand it and have empathy with it.
\nI want to pressure-test that one layer down and talk about the fear people understandably have with AI: fear of job security, job change, reskilling, in an environment where it\u2019s not always obvious how those things will play out. I\u2019m more optimistic than most about what these tools can bring, but I understand why those perceptions exist. Good executive sponsors, steering committees, and project structures all mean something, but it all gets lost the moment you\u2019re out of touch with the feeling in the organization, and whether your pace of change is landing the way it\u2019s meant to.
\nOne more point, specific to data and AI: We\u2019re governing data on one hand, but we have to watch that we\u2019re not protecting it to build new power bases of leadership. That\u2019s one of the most important tensions we\u2019re navigating right now. We can\u2019t have full democratization of data\u2014there's important governance needed to protect the reasons it was created or licensed, and how it can be used. So we need governance, but also freedom to innovate. What gets in the way is when we control data for power rather than governance. That\u2019s a watch-out we need to keep course-correcting. But I\u2019m sure in five years we\u2019ll be more mature than we are now, and in ten years native to how we work. For now, these are areas that deserve double attention.
\nLeading change without losing trust
\nYigit Yorulmaz: I want to probe on that idea of change and how you enact it. You\u2019ve been emphasizing how technology alone isn\u2019t enough, and clearly we need to think about creating the right governance, operating model, and winning hearts and minds as part of the journey. But I want to ask about the hardest moments you\u2019ve reflected on. As an industry, we\u2019re now at the cusp of another wave of innovation with even greater potential. How do you think about the relationship between the degree of ambition, leaning forward, and doing change the right way for the culture and structural ways of working? Is there a balance? What learnings are you bringing into this new chapter of your leadership?
\nAlex Condoleon: A friend of mine once said innovation is like a potent therapy\u2014the dosage, the cadence, the sequencing, it all matters. I keep that top of mind. I\u2019ve probably got more battle scars than trophies on my wall.
\nThere\u2019s no perfect recipe to it. But when you get the mix right, you see teams energized, you create momentum, and people lean in. When you get it wrong, there\u2019s resistance and fatigue. Even the most ambitious vision means nothing if you don\u2019t win hearts and minds, and ripple that change outward.
\nA few thoughts occur to me, in dramatic periods like the one we\u2019re in. First, on the leader: Leading change today takes immense courage. Success is like a very fine wire that you have to walk, and it\u2019s got failure on either side of it. And you have to be prepared to fail if you want to succeed, something organizations don\u2019t always reward.
\nSecond, alignment top-down matters, because organization-wide transformative change can\u2019t make sense in isolation; it has to fit the totality of where the organization is heading. One-off bits of change don\u2019t make sense to people if they don\u2019t fit the collective effort. So the second thought is about organizational relevance to the level of ambition you\u2019re pursuing.
\nThird is demonstrating that it\u2019s not just a priority, but one you're willing to put resources and capabilities against. Those are the elements that actually make change happen. But all of that means nothing without trust in the system, and trust is often our afterthought rather than our first thought. We\u2019re at our best when the change we drive builds a great result for the company, but also a great culture\u2014a spirit of collective innovation, success, and celebration. That means staying focused on whether the change is relevant to the passion, purpose, and pride people bring to their work.
\nThere\u2019s a bit of political intelligence involved, which is really just understanding how to lead change relevant to multiple agendas. If you can make healthcare better and make people proud of their contribution, that\u2019s the beginning of a change journey set up for success.
\nNavraj Nagra: When we think about proving the impact of medical affairs, what does success look like for you, both for the work you\u2019re doing and for the medical affairs organization more broadly?
\nAlex Condoleon: I don\u2019t know how many people hold my view, but I hope soon everyone in medical affairs does. I believe that you absolutely can measure the impact of medical affairs. And if I get a little bit provocative, maybe the more relevant question is: Are we ready to take accountability for what the results tell us? Because, if we\u2019re going to talk about metrics and measuring impact, I think we need to be really clear on why we even want to answer those questions. And when we\u2019re at our best, I think we want to understand what our impact is solely and purely so we can push ourselves to do better. I say that not because we shouldn\u2019t be proud of the results we\u2019re already achieving, but because the needs in healthcare are so high that we should always want more.
\nAdditionally, measuring medical affairs\u2019 impact doesn\u2019t mean measuring everything we do. We should choose only a few things that really matter. Two lenses come to mind. First, it needs to be an externally relevant guide. We\u2019re a healthcare professional (HCP), healthcare system, and healthcare environment facing team, and everything we do\u2014the evidence we generate, the insight we bring, how we engage with stakeholders\u2014is about shifting healthcare. Being grounded in something externally based matters because it represents the full value chain of medical. Second, to help stakeholders at any level understand our measures of impact, we need to keep it simple and consistent; a dashboard with a thousand data points isn\u2019t easy to understand.
\nSo what tells a meaningful, externally grounded story that keeps pushing us further? Care gaps are one area I\u2019ve been a strong believer in\u2014understanding demonstrated practice behavior improvement, with a belief that over time it leads to patient outcome improvement. That represents the full value chain of medical: Wherever you sit in that chain, you\u2019ve contributed to something impacting how care is delivered. It\u2019s understandable at the highest level, and externally grounded. That doesn\u2019t mean it\u2019s the only data point further down; we need to understand how each team contributes to that roll-up metric, which is where we get a hierarchy of metrics feeding up to something that shows impact.
\nThat\u2019s a tangible, actionable example, relevant to the full value chain, one that says medical understands its ultimate impact, and teams can pursue an endless goal of driving better care. We\u2019re probably never going to reach zero care gap, but we can reach a cultural and professional endpoint, an enduring ambition to not give up. That represents the passion, precision, and value of this function, and can resonate externally and internally to elevate it.
\nWhat patient centricity actually looks like
\nYigit Yorulmaz: I love that what you\u2019re saying ties back to your clear passion for the patient, that ultimately the better outcomes are in service of patients across these different therapeutic areas. I want to probe on that a bit more. What does great look like when we\u2019re being truly patient-centric? It\u2019s a term everybody uses, and in some shape or form we are, but if we push ourselves to bring that to day-to-day for a global medical affairs organization, from its full remit, what could great look like in your vision for being patient-centric?
\nAlex Condoleon: It\u2019s a term we\u2019ve used for a long time, and we keep learning what it means in its full power. At its core, patient centricity starts as a cultural philosophy before it becomes an operational model. Many of us in medical affairs started as healthcare professionals, doctors, pharmacists, nurses, or allied health professionals, and we need to stay connected to that origin, ultimately impacting patient outcomes. The cultural symbols start with leaders and colleagues talking about the insights they\u2019ve learned from and for patients, the actions they\u2019re taking to improve patients\u2019 lives, and how they\u2019re measuring the contribution they\u2019re making. That cultural ethos is the important starting point for the operations to come alive.
\nOperationally, we often think about engaging healthcare professionals, especially key opinion leaders, to make our science relevant and applicable to how care is delivered. We can go a step beyond that now. Data and analytics bring us closer to the patient, giving us a more precise understanding of the patient journey and where the friction points are than we might have assumed before. It gives us a louder voice for what patients want, their need state, and the endpoints that matter most in their lives. Lastly, it shows us that patients aren\u2019t uniform; every individual is distinct. Injecting those reality checks into our strategies, priorities, and tactics can move us off the textbook approach to healthcare and into the meaningful part of it, the reason we entered this field in the first place.
\nThe role of AI in shaping medical affairs
\nNavraj Nagra: You\u2019ve described how you\u2019re seeing AI genuinely make a difference for medical affairs, clinicians, and patients, somewhat through a form of co-intelligence, where your teams work with AI to be more productive. Where is there still hype in the equation, though? We still see a lot of claims and counterclaims about what\u2019s possible and what\u2019s real. I would love your take from a global perspective.
\nAlex Condoleon: I genuinely believe AI is the most disruptive technological advancement of our generation, at least so far. There is hype, coming mostly from doomsdayers forecasting the downfall of society and workforces. But technologically, I don\u2019t think we\u2019re seeing hype. If anything, it\u2019s the opposite: The technology is ahead of what we\u2019re able to deploy. A decade ago, when I was pushing a digital agenda for medical affairs, the formula was overpromise, work hard, and learn from what didn\u2019t live up to the business case. Today it\u2019s the reverse: More is possible than we can deploy, and the external world is moving faster than we can internally absorb.
\nA few points to back that up: More than 10 percent of the world\u2019s population are using ChatGPT every week today. And one in four of those frequent users is searching on healthcare every week. Microsoft Copilot recently put out its report of usage in 2025, and the number one (mobile usage) search topic was health. So it\u2019s not only that AI is disruptive, it\u2019s that healthcare is front and center of this AI story. That\u2019s why I say there should be a positive call to action for stakeholders in life sciences, pharma, and medical affairs when it comes to AI.
\nAt the same time, AI has the potential to eliminate repetitive, lower-value work, freeing up our professional contribution to be maximized elsewhere. That means thinking through how AI augments human potential, lifting the co-intelligence, judgment, creativity, and discovery we bring. AI will give us productivity, and I think the real strategic question, the counterbalance to the doomsday hype question is: What are we going to do with the capacity that we free up in organizations? If all we do is drive productivity with AI, I don\u2019t think we promote new discovery. I don\u2019t think we change standards of care. I don\u2019t think we close care gaps beyond what we are doing today. All we do is achieve some productivity. I think the question for all of us is to get crisp on how we are going to reinvest the liberated resources so that we use the co-intelligence of AI to have an even bigger impact on healthcare than what we can imagine achieving today.
\nThat\u2019s the magic place, but getting there takes real strategic thinking. I actually think that second part\u2014where we redeploy resources to\u2014is as big a strategic question as any we\u2019ll have to answer as we build up AI capabilities.
", "summary": "In this episode of Eureka!, McKinsey\u2019s podcast on innovation in life sciences, host Navraj Nagra and McKinsey partner Yigit Yorulmaz speak with Dr. Alex Condoleon, chief medical affairs officer for global medical engagement and impact at Pfizer. Condoleon, a physician executive who has led medical, innovation, and digital health efforts across Australia, China, Europe, and the United States, argues that medical affairs\u2019 longstanding ceiling has lifted and that the function can now be a genuine strategic partner in life sciences. He discusses what it takes to bring the right evidence to the right clinicians at the right moment, how to measure medical affairs\u2019 impact beyond commercial metrics, and what patient-first looks like at a global scale across a wide portfolio of disease areas. An edited version of their conversation follows.
\nA career built anticipating the industry\u2019s next move
\nNavraj Nagra: You\u2019ve had a distinctive journey into medical affairs, including leadership roles spanning medical, innovation, and digital healthcare. What were the pivotal moments that shaped your career direction, and what ultimately pulled you into this role at Pfizer?
\nAlex Condoleon: Looking back at the career trajectory I\u2019ve been on, I\u2019ve always believed that if you want to impact global health in a meaningful way, you need to get a step ahead of where the system is already going, not follow it. That has guided each of my career moves.
\nI started in the industry in Australia and moved to China. That wasn\u2019t because I had clarity on what China would become; it was because I had a conviction that China would play a significant role in healthcare. Fast-forward to today, and China\u2019s ambition is becoming clear in biotech and in AI, and that shift is going to impact global health.
\nFrom there, I moved to Europe, into my first dedicated role focused on consumer healthcare rather than the prescription business. That was me seeking to understand how patients were starting to take control of their health and what that could mean. In that role, almost a decade ago, I was seeing early direct-to-patient models. I remember meeting the team at what today we call Ro.com when they were in their infancy, and yet there was something about what they were building toward that I thought was going to matter. Fast-forward to today and we see platforms like Hims & Hers and Ro.com playing a dominant role in how people choose to access glucagon-like peptide-1 (GLP-1) drugs in the United States.
\nMy next pivot was moving to the United States, focused on enterprise digital transformation in medical affairs, and later building out a digital healthcare franchise. Again, it\u2019s a similar pattern: investing ahead of a change I believed was coming, so I could be ready for the moment. I think that moment is the one we\u2019re living through now, a moment of convergence. Many global powers are influencing what the healthcare landscape looks like; digital and AI are rising in healthcare delivery; and all of that is pushing the trend of patient empowerment and patient-guided care. When you see that kind of seismic opportunity come together, it\u2019s great to be in an organization like Pfizer, which seeks to do bold things, because being bold in healthcare is your way of staying relevant and achieving leadership.
\nI\u2019ve been very fortunate to come into a position focused on the role of data, analytics, and AI in optimizing how we think about medicine, science, and value, and how we communicate it and bring it to the forefront. I\u2019m hopeful those things make a meaningful difference to moving healthcare forward.
\nWhy the ceiling on medical affairs is gone
\nYigit Yorulmaz: Your remit in medical affairs involves overseeing a high-caliber global platform focused on medical engagement and impact. How does this, and your past experience, help you reimagine what medical affairs can look like? What excites you about building and leading a function that doesn\u2019t exist at any scale in many medical affairs organizations?
\nAlex Condoleon: I\u2019m excited by the role I get to do every day. Part of the reason is that I\u2019d reached a point where I genuinely felt medical affairs had hit a ceiling. What energizes me today is that I\u2019m more convinced than ever that ceiling is gone, and that medical affairs has the potential to be a true strategic partner in life sciences. We\u2019re in the process, as a function, of stepping into that role, and I think now is the moment.
\nThere are a few reasons for that belief. It starts with how healthcare systems, and the experience of healthcare professionals and patients engaging with the system, are fundamentally changing. In a world where the lives of healthcare professionals are chaotic and made more complex every day, the need for accurate, relevant, timely signals to guide how they manage patients is more important than ever. But it must be approached differently than in the past. That\u2019s exciting, but it will also challenge us: Compared to a decade ago, we need to reimagine our role, reinvent how we bring value to our target customer groups, and constantly listen to insight and feedback that helps us improve.
\nI say that partly because, as an industry, we\u2019re in a period of great productivity, with lots of new products coming to market. But there\u2019s a paradox: Every time we launch a new product and solve an unmet need, we also unintentionally add complexity to an already complex treatment landscape. That pushes the importance of thinking about what information healthcare professionals and patients need, and when. That\u2019s where there\u2019s an exciting role for digital platforms to enhance and magnify what we aspire to do, enriched by the AI advances we\u2019re seeing now.
\nSo although we\u2019re challenged to lead through a period where the pace of external change feels like it\u2019s outstripping our internal capacity to absorb it, that\u2019s fundamentally a world of opportunity for medical affairs teams. But to grab that opportunity, building the right data, analytics, and AI infrastructure is critical. We\u2019re really good with data in medical affairs, good at making sense of incomplete and imprecise data. We just need different techniques, and to get comfortable with the learning curve.
\nIf we can build those fundamentals into our capability mix, and lead in a way that resonates with teams, I think the way the external world and technology are advancing means something important for medical, and represents the pillars on which we\u2019ll step up as a function.
\nReaching physicians in a more complex landscape
\nNavraj Nagra: Just to add a further deep dive here: You mentioned the use of some of these technologies, these new ways of working, to better engage physicians and medical professionals. Can you give me an example of what you\u2019re doing at the moment to help physicians think about this more complex landscape, and how you\u2019re engaging with them differently compared to what was being done even ten or 20 years ago?
\nAlex Condoleon: That\u2019s a really important question. Given the data now available, we can, with a good degree of precision, understand the clinical realities physicians are facing, including where there are gaps in care that represent challenges and opportunities for medical teams working with the patients they oversee. Once we understand that, we have a need state we can seek to solve.
\nWe\u2019ve built the data and analytics capabilities that give us precision on who\u2019s facing a clinical challenge that aligns with where we think there\u2019s room for improvement. At the same time, we\u2019ve been able to be precise about the digital footprint of those healthcare professionals, which lets us rethink what content we need and how to make it relevant to the channels physicians are actually on. That\u2019s half the equation.
\nWhat\u2019s equally important is that there\u2019s now maturity in the healthcare landscape, letting us partner with medically focused digital channels to deploy content precisely, in the right form and at the right time, for target physicians. By bringing together the right customer segment, the right need state, and content adapted to how information is digested on the right channels, and by partnering with these platforms, we\u2019ve seen tangible improvement in care. That improvement doesn\u2019t replace what our field medical teams do, it\u2019s supplemental. It increases reach and magnifies what we\u2019re already doing through impactful field conversations.
\nThat\u2019s something we couldn\u2019t have fully believed two years ago would be possible, but it is and it works. It\u2019s an example of taking the discipline of medical affairs and matching it to the need state of the healthcare professionals you\u2019re trying to reach. Our traditional face-to-face work can now be magnified and elevated beyond that, and we\u2019ll only see acceleration of it.
\nOne portfolio, many specialized environments
\nNavraj Nagra: One thing we recognize is that there\u2019s such a wide portfolio across Pfizer, from vaccines to primary care to specialty medicines. When you think about patient needs, physician needs, and the scientific conversations across these categories, how does that vary? What are the differences in maturity of the approaches you\u2019ve discussed across those categories?
\nAlex Condoleon: That\u2019s a fascinating question, and one of the things I\u2019m really fortunate about in my role. Working across a portfolio like Pfizer\u2019s, you start to see different elements of the future of healthcare emerge, in different ways, in different specialties. One day I\u2019m hopeful they\u2019ll converge.
\nTake oncology, or rare diseases, where Pfizer has a significant focus. We\u2019re seeing the rapid rise of a precision, science-driven form of personalization. It was only a handful of decades ago that we defined cancers by their organ of origin, but today we have hundreds of subtypes based on histology, genetics, and biomarkers, and we can envisage further precision coming. We can take the same parallel with heart failure, which has been segmented as we understand more of the genetic contributors. So there\u2019s precision emerging in some areas, and it will only grow.
\nThen at another end of the spectrum, look at vaccines, or an area like GLP-1s, where the dynamic is very different: You\u2019re fundamentally engaging with the population and shaping patient choice and behavior. Patient literacy, ease of access, and convenience matter enormously to the choices people make. And then there\u2019s primary care and specialty care, where we\u2019re focused on understanding healthcare professional preferences, beliefs, practice patterns, and care gaps, so we can be as targeted and relevant as possible.
\nI call those out because one day those three thematics will converge into a great healthcare experience, where the science of precision spans all of healthcare, where people navigate their own healthcare guided by their choices, preferences, and literacy, and where the industry has learned to be as relevant and personalized as possible to the attitudes and behaviors of the healthcare professionals we engage with.
\nSo the breadth of our portfolio is a landscape of learning, one I\u2019m hopeful will lead to a convergence that brings more value to patients and healthcare professionals alike. Today you can\u2019t treat the portfolio as one-size-fits-all, but you can treat it as several test-and-learn environments, where concepts proven in one area show us when others will benefit from the same trends.
\nWhat to standardize, what to keep local
\nYigit Yorulmaz: Fascinating to hear you talk about that. It\u2019s absolutely true that different parts of healthcare are evolving in different ways, but when the portfolio is so diverse, medical affairs can\u2019t rely on a single playbook to determine what good looks like. So you have, as a leader, the choice of deciding what must be standardized\u2014whether platform capabilities, principles, or how you measure things\u2014and what must be bespoke: the content you use, how you\u2019re thinking about channel mix. How do you think about that variation across therapeutic areas or geographies, based on your experience? What\u2019s standardized, and what\u2019s unique?
\nAlex Condoleon: There\u2019s no easy answer to that. As we build new capabilities for medical affairs\u2014especially data, analytics, and AI capabilities\u2014scale matters. On the flip side, being relevant to different therapeutic areas or geographies also matters. Bridging that tension is a day-to-day exercise in judgment.
\nThere are areas where standardization and scale matter most. How we structure and govern data is best done with discipline at an enterprise level. The same goes for the analytics platforms and new professional skills we\u2019re bringing into teams; while we\u2019re maturing them and working out what value looks like, there\u2019s benefit to housing that learning centrally. The same is true for how we build AI models and partner with large language model (LLM) companies: At this early-learning stage, that fits into working out the foundational capabilities that represent the workforce of the future.
\nBut none of that matters if we can\u2019t make it relevant to the people we\u2019re seeking to bring value to. Central capabilities can be quicker to build in transforming moments, but ultimately they need to sit as close to the customer as possible. That means working closely with the teams you\u2019re partnering with, not being overprescriptive about the offering, and building understanding of a capability and a best practice while giving people the freedom to innovate beyond it. That\u2019s how we keep capabilities relevant, especially to medical stakeholders.
\nWe\u2019re in build mode right now. We might have ideas of a target operating model, but the key word is target, not today\u2019s state. Organizations aren\u2019t static, they\u2019re living things, and it\u2019s when we let that chemistry play out that we innovate beyond what\u2019s apparent today. So there\u2019s a place, especially in this digitally accelerated phase, for building capabilities that don\u2019t yet exist. But if we can\u2019t equally let teams apply those capabilities with freedom to operate in their own context, we\u2019ll lose value and struggle to translate healthcare know-how and tech know-how into medical affairs impact.
\nGoverning AI and digital initiatives
\nNavraj Nagra: You\u2019ve described what sounds like a plethora of AI and digital initiatives to drive medical engagement and impact. How are you governing the decision-making around this? How are you deciding where the ROI is good enough to give your teams the freedom to operate in that space, and where you don\u2019t believe there\u2019s going to be impact from certain digital interventions? What\u2019s the operating model behind these digital and AI initiatives?
\nAlex Condoleon: Firstly, executive sponsorship really matters when you\u2019re going through big transformative change like this. That\u2019s because there\u2019s a lot we don\u2019t know; we\u2019re going to make mistakes and learn through failing. If the goal is worth heading toward, we\u2019ll tolerate the learning that comes with it.
\nBeyond executive sponsorship, the operating model has to adapt to what you\u2019re building, especially when capabilities in digital and AI are emergent rather than matured. That means wanting as many stakeholders vested in the journey as possible, while respecting that your priority might not be theirs. That\u2019s why high-visibility steering groups matter; they give stakeholders a chance to stay close, understand where the transformation is heading, and be heard. At every layer of leadership, it matters, and wherever the bottleneck comes up, you need to understand it and have empathy with it.
\nI want to pressure-test that one layer down and talk about the fear people understandably have with AI: fear of job security, job change, reskilling, in an environment where it\u2019s not always obvious how those things will play out. I\u2019m more optimistic than most about what these tools can bring, but I understand why those perceptions exist. Good executive sponsors, steering committees, and project structures all mean something, but it all gets lost the moment you\u2019re out of touch with the feeling in the organization, and whether your pace of change is landing the way it\u2019s meant to.
\nOne more point, specific to data and AI: We\u2019re governing data on one hand, but we have to watch that we\u2019re not protecting it to build new power bases of leadership. That\u2019s one of the most important tensions we\u2019re navigating right now. We can\u2019t have full democratization of data\u2014there's important governance needed to protect the reasons it was created or licensed, and how it can be used. So we need governance, but also freedom to innovate. What gets in the way is when we control data for power rather than governance. That\u2019s a watch-out we need to keep course-correcting. But I\u2019m sure in five years we\u2019ll be more mature than we are now, and in ten years native to how we work. For now, these are areas that deserve double attention.
\nLeading change without losing trust
\nYigit Yorulmaz: I want to probe on that idea of change and how you enact it. You\u2019ve been emphasizing how technology alone isn\u2019t enough, and clearly we need to think about creating the right governance, operating model, and winning hearts and minds as part of the journey. But I want to ask about the hardest moments you\u2019ve reflected on. As an industry, we\u2019re now at the cusp of another wave of innovation with even greater potential. How do you think about the relationship between the degree of ambition, leaning forward, and doing change the right way for the culture and structural ways of working? Is there a balance? What learnings are you bringing into this new chapter of your leadership?
\nAlex Condoleon: A friend of mine once said innovation is like a potent therapy\u2014the dosage, the cadence, the sequencing, it all matters. I keep that top of mind. I\u2019ve probably got more battle scars than trophies on my wall.
\nThere\u2019s no perfect recipe to it. But when you get the mix right, you see teams energized, you create momentum, and people lean in. When you get it wrong, there\u2019s resistance and fatigue. Even the most ambitious vision means nothing if you don\u2019t win hearts and minds, and ripple that change outward.
\nA few thoughts occur to me, in dramatic periods like the one we\u2019re in. First, on the leader: Leading change today takes immense courage. Success is like a very fine wire that you have to walk, and it\u2019s got failure on either side of it. And you have to be prepared to fail if you want to succeed, something organizations don\u2019t always reward.
\nSecond, alignment top-down matters, because organization-wide transformative change can\u2019t make sense in isolation; it has to fit the totality of where the organization is heading. One-off bits of change don\u2019t make sense to people if they don\u2019t fit the collective effort. So the second thought is about organizational relevance to the level of ambition you\u2019re pursuing.
\nThird is demonstrating that it\u2019s not just a priority, but one you're willing to put resources and capabilities against. Those are the elements that actually make change happen. But all of that means nothing without trust in the system, and trust is often our afterthought rather than our first thought. We\u2019re at our best when the change we drive builds a great result for the company, but also a great culture\u2014a spirit of collective innovation, success, and celebration. That means staying focused on whether the change is relevant to the passion, purpose, and pride people bring to their work.
\nThere\u2019s a bit of political intelligence involved, which is really just understanding how to lead change relevant to multiple agendas. If you can make healthcare better and make people proud of their contribution, that\u2019s the beginning of a change journey set up for success.
\nNavraj Nagra: When we think about proving the impact of medical affairs, what does success look like for you, both for the work you\u2019re doing and for the medical affairs organization more broadly?
\nAlex Condoleon: I don\u2019t know how many people hold my view, but I hope soon everyone in medical affairs does. I believe that you absolutely can measure the impact of medical affairs. And if I get a little bit provocative, maybe the more relevant question is: Are we ready to take accountability for what the results tell us? Because, if we\u2019re going to talk about metrics and measuring impact, I think we need to be really clear on why we even want to answer those questions. And when we\u2019re at our best, I think we want to understand what our impact is solely and purely so we can push ourselves to do better. I say that not because we shouldn\u2019t be proud of the results we\u2019re already achieving, but because the needs in healthcare are so high that we should always want more.
\nAdditionally, measuring medical affairs\u2019 impact doesn\u2019t mean measuring everything we do. We should choose only a few things that really matter. Two lenses come to mind. First, it needs to be an externally relevant guide. We\u2019re a healthcare professional (HCP), healthcare system, and healthcare environment facing team, and everything we do\u2014the evidence we generate, the insight we bring, how we engage with stakeholders\u2014is about shifting healthcare. Being grounded in something externally based matters because it represents the full value chain of medical. Second, to help stakeholders at any level understand our measures of impact, we need to keep it simple and consistent; a dashboard with a thousand data points isn\u2019t easy to understand.
\nSo what tells a meaningful, externally grounded story that keeps pushing us further? Care gaps are one area I\u2019ve been a strong believer in\u2014understanding demonstrated practice behavior improvement, with a belief that over time it leads to patient outcome improvement. That represents the full value chain of medical: Wherever you sit in that chain, you\u2019ve contributed to something impacting how care is delivered. It\u2019s understandable at the highest level, and externally grounded. That doesn\u2019t mean it\u2019s the only data point further down; we need to understand how each team contributes to that roll-up metric, which is where we get a hierarchy of metrics feeding up to something that shows impact.
\nThat\u2019s a tangible, actionable example, relevant to the full value chain, one that says medical understands its ultimate impact, and teams can pursue an endless goal of driving better care. We\u2019re probably never going to reach zero care gap, but we can reach a cultural and professional endpoint, an enduring ambition to not give up. That represents the passion, precision, and value of this function, and can resonate externally and internally to elevate it.
\nWhat patient centricity actually looks like
\nYigit Yorulmaz: I love that what you\u2019re saying ties back to your clear passion for the patient, that ultimately the better outcomes are in service of patients across these different therapeutic areas. I want to probe on that a bit more. What does great look like when we\u2019re being truly patient-centric? It\u2019s a term everybody uses, and in some shape or form we are, but if we push ourselves to bring that to day-to-day for a global medical affairs organization, from its full remit, what could great look like in your vision for being patient-centric?
\nAlex Condoleon: It\u2019s a term we\u2019ve used for a long time, and we keep learning what it means in its full power. At its core, patient centricity starts as a cultural philosophy before it becomes an operational model. Many of us in medical affairs started as healthcare professionals, doctors, pharmacists, nurses, or allied health professionals, and we need to stay connected to that origin, ultimately impacting patient outcomes. The cultural symbols start with leaders and colleagues talking about the insights they\u2019ve learned from and for patients, the actions they\u2019re taking to improve patients\u2019 lives, and how they\u2019re measuring the contribution they\u2019re making. That cultural ethos is the important starting point for the operations to come alive.
\nOperationally, we often think about engaging healthcare professionals, especially key opinion leaders, to make our science relevant and applicable to how care is delivered. We can go a step beyond that now. Data and analytics bring us closer to the patient, giving us a more precise understanding of the patient journey and where the friction points are than we might have assumed before. It gives us a louder voice for what patients want, their need state, and the endpoints that matter most in their lives. Lastly, it shows us that patients aren\u2019t uniform; every individual is distinct. Injecting those reality checks into our strategies, priorities, and tactics can move us off the textbook approach to healthcare and into the meaningful part of it, the reason we entered this field in the first place.
\nThe role of AI in shaping medical affairs
\nNavraj Nagra: You\u2019ve described how you\u2019re seeing AI genuinely make a difference for medical affairs, clinicians, and patients, somewhat through a form of co-intelligence, where your teams work with AI to be more productive. Where is there still hype in the equation, though? We still see a lot of claims and counterclaims about what\u2019s possible and what\u2019s real. I would love your take from a global perspective.
\nAlex Condoleon: I genuinely believe AI is the most disruptive technological advancement of our generation, at least so far. There is hype, coming mostly from doomsdayers forecasting the downfall of society and workforces. But technologically, I don\u2019t think we\u2019re seeing hype. If anything, it\u2019s the opposite: The technology is ahead of what we\u2019re able to deploy. A decade ago, when I was pushing a digital agenda for medical affairs, the formula was overpromise, work hard, and learn from what didn\u2019t live up to the business case. Today it\u2019s the reverse: More is possible than we can deploy, and the external world is moving faster than we can internally absorb.
\nA few points to back that up: More than 10 percent of the world\u2019s population are using ChatGPT every week today. And one in four of those frequent users is searching on healthcare every week. Microsoft Copilot recently put out its report of usage in 2025, and the number one (mobile usage) search topic was health. So it\u2019s not only that AI is disruptive, it\u2019s that healthcare is front and center of this AI story. That\u2019s why I say there should be a positive call to action for stakeholders in life sciences, pharma, and medical affairs when it comes to AI.
\nAt the same time, AI has the potential to eliminate repetitive, lower-value work, freeing up our professional contribution to be maximized elsewhere. That means thinking through how AI augments human potential, lifting the co-intelligence, judgment, creativity, and discovery we bring. AI will give us productivity, and I think the real strategic question, the counterbalance to the doomsday hype question is: What are we going to do with the capacity that we free up in organizations? If all we do is drive productivity with AI, I don\u2019t think we promote new discovery. I don\u2019t think we change standards of care. I don\u2019t think we close care gaps beyond what we are doing today. All we do is achieve some productivity. I think the question for all of us is to get crisp on how we are going to reinvest the liberated resources so that we use the co-intelligence of AI to have an even bigger impact on healthcare than what we can imagine achieving today.
\nThat\u2019s the magic place, but getting there takes real strategic thinking. I actually think that second part\u2014where we redeploy resources to\u2014is as big a strategic question as any we\u2019ll have to answer as we build up AI capabilities.
", "image": null, "banner_image": null, "date_published": "2026-10-25T16:00:00Z", "date_modified": "2026-10-25T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}, {"id": "https://www.mckinsey.com/insights/5265434b-be4d-4919-b340-b595e6ae68ec", "url": "https://www.mckinsey.com/featured-insights/mckinsey-on-books/author-talks-ai-is-advancing-is-our-judgment-keeping-up", "external_url": null, "title": "Author Talks: AI is advancing. Is our judgment keeping up?", "content_text": "", "content_html": "In this edition of Author Talks, McKinsey Global Publishing\u2019s Adam Volk chats with Tiffany Xingyu Wang, president and cofounder of Oasis Consortium, and World Economic Forum Young Global Leader, about The Trust Code: How to Unlock AI\u2019s Promise Without Losing Humanity (Wiley, September 2026). Wang examines why our instincts for trust can misfire when applied to increasingly capable AI systems. She argues that learning to calibrate trust can help individuals and organizations use AI with intention, preserve human judgment, and capture its benefits without sacrificing agency. An edited version of the conversation follows.
\nWhy did you write this book?
\nI wrote this book because I kept watching very smart people ask a question that\u2019s almost too big to be useful: \u201cDo we trust AI?\u201d AI\u2019s already here, and we trust it when we use our credit cards, because fraud detection is running behind the scenes. We trust it when we switch on the light, because various types of software are balancing the grid. But more and more often, AI models are drafting our work, and, increasingly, AI agents are acting on our behalf. The central question of our time is no longer whether to trust AI, but how much trust does a system deserve, under which conditions, and for which tasks?
\nOften, we see that being overly trusting exposes us to harm, and not trusting enough traps us in paralysis when it comes to innovation. It\u2019s only the calibrated trust, which this book advocates for, that enables us to move forward. Professionally, I spent over a decade building trust and safety systems for big enterprises.
\nOften in boardrooms, crisis rooms, and on conference stages, people ask me what trust means in the AI era. I never really had the time to give a holistic answer. When I began writing this book, I was pregnant with my second child, and my son was two years old. That\u2019s when the stakes became even more personal to me.
\nMy son talked to Alexa and rode in the Waymo with me regularly. None of this is remarkable because the next generation will not remember a time when AI was not already embedded in our daily life, making every decision with us and on our behalf.
\nWe may not control how fast AI is advancing, but one thing we still have the time and the power to control is how deliberately we give AI that authority. I searched for a book that could teach me that missing skill, but I couldn\u2019t find it. So I decided to write it.
\nWas there anything that surprised you in the research, writing, or response?
\nWe have a machine\u2014cognitive technology\u2014that sounds warm, very confident, very empathetic, but without any of those signals corresponding to understanding, responsibility, or intentionality that a human would have. This creates a profound evolutionary mismatch: the Stone Age brain we have evolved from is evaluating Space Age systems.
\nThat mismatch is one of the first surprises that I encountered during the overall process. It\u2019s very foundational to understanding many aspects the book covers. The second surprise came from writing the book itself. The faster AI moved, the more convinced I became that the book shouldn\u2019t be about what is happening with AI now but about how to think about whatever happens next.
\nAs you can imagine, it\u2019s terrifying for an author to write about what is probably the fastest-moving technology in history. I wondered whether the book would be outdated at the time it hit bookstores. I tried many ways and spent much effort to really go back to first principles when explaining any concept, any headline in this book, so that people could use it as an evergreen guide to decipher whatever comes next.
\nWhy is it so difficult to define trust when it comes to AI?
\nAfter speaking with many people while researching and interviewing, I found the distinction that cleared this up for me.
\nThere is trustworthiness, which is about the system itself. And we can talk about parameters that matter to AI technology. But in this context of AI, is it truthful enough? Is it competent enough? Is it also benevolently aligned with your personal goals or your enterprise goals well enough? Is it appropriate for that context? That\u2019s about the system.
\nTrust is a different thing. Trust is about us. Am I willing to rely on it?
\nAnother matter is calibration, and whether those two match. A technically very strong system can be under-trusted, and a very charming but unreliable system can be wildly over-trusted. That\u2019s what\u2019s happening everywhere now. Both [approaches] are actually failures.
\nInstead of asking, \u201cDo I trust AI?,\u201d what we should ask is, \u201cWhat vulnerability are you accepting today for what promise in the future and in what particular situation?\u201d Because the definition of trust is that you accept vulnerabilities today in hopes of a promise in the future. Trust remains a very vague sentiment until you can articulate and understand the vulnerabilities you accept today and the goals you want to achieve tomorrow for very specific conditions and situations.
\nBut once you correctly articulate the question and think through it\u2014some frameworks we talk about in the book\u2014trust becomes a decision. It becomes a decision that you can examine, measure, and improve upon.
\nWhen it comes to AI adoption, which areas of trust do you see leaders neglecting the most?
\nI really focus on what trust means through the lens of AI. There are four prerequisites for trust: competence, truthfulness, benevolence, and contextual fit. I chose those four because I constantly saw people mistake competence for trustworthiness.
\nIt\u2019s very important to understand that AI is so capable that it will sound very competent and aligned with you, while it has no reason to be benevolent to maximize your well-being. Understanding that pitfall allows you to have the sense of guardrails you need to constantly think about when using AI. Manifestation is the quadrant that is especially important for enterprise leaders now. Does our behavior match what we say we trust? There are so many research papers now being published that say that. We started pilots, but the ROI is low.
\nIt\u2019s very important to understand that a pilot is a statement of belief. What really shows a state of trust when an enterprise adopts AI or not is whether we change the workflow, whether we train the staff, or whether we change the incentives. Match what you say to what you do, and think about how you use the signals of behaviors to assess the level of trust you delegate to AI.
\nFinally, there are the unseen factors. It\u2019s important to understand that these technologies are meant to be warm, empathetic, and very confident. But they are not earning our trust by delivering the track record and the proven evidence, either by ROI or by reliability over time. They\u2019re simply triggering our trust.
\nHow can we hold AI players accountable when even they can\u2019t predict how their systems will behave?
\nIn today\u2019s AI age, accountability means owning the conditions around unpredictability, and that articulation is extremely important. There is a missing actor in this conversation about AI. We usually draw a map of two posts: AI labs that build models, and regulators that write the rules.
\nBut enterprises decide what gets bought and what\u2019s connected to real systems. We give access to real data and deploy it at scale. Consider any mature industry, for example, aviation. Airplane builders are responsible for designing and testing it, but airlines have the responsibility for how the plane is operated and who controls it.
\nOf course, the regulators set a floor nobody can opt out of. AI needs the same layer. Often, the power and responsibility of enterprises as a whole are underused and underrecognized in AI governance or AI accountability. In other words, enterprise leaders have more agency than we give credit for in this race of AI.
\nIf I\u2019m buying AI for an enterprise, I would stop asking, \u201cIs this model safe?\u201d Instead, I would say, \u201cShow me a failure and show me how you contained it. Show me how quickly I will know [about a failure], and show me what recourse I have if it costs me money or reputation.\u201d
\nEnterprise leaders need to understand the strong agency you have to decide what model you buy, how you deploy it, and what protection measures you take. Hopefully, when enterprises come together and make that the default, the norm, we can see the maturity of this industry.
\nLabs build capabilities and regulators set the floor, but it is the enterprise that decides how much authority we give it.
\nCan you give an example of how AI is expanding human judgment and perception?
\nI interviewed Professor Stephen Quake at Stanford, a pioneer of single-cell genomics. He described biology to me as historically 90 percent experimental and 10 percent computational. The ambition is to reverse it so that it\u2019s 90 percent computational and 10 percent experimental.
\nWe have already seen a glimpse of this with AlphaFold, which used AI to predict more than 20 million protein structures and to understand better how biology works. A cell takes it to a more complex, different level, because every cell has 20,000 protein-coding genes.
\nUnderstanding cells better means we can better understand our body, our health, our diseases, and our therapies. But when it involves running in 20,000 dimensions, that\u2019s beyond human brain capacity. That\u2019s something you can\u2019t do without AI.
\nBut that doesn\u2019t mean human judgment is stripped from the process. Instead, where the judgment is situated shifts. Arguably, the more AI can surface and test hypotheses for us\u2014more cheaply, quickly, and at scale\u2014the more important human judgment becomes. Because ultimately, humans must decide which hypotheses to pursue, which tasks and experiments to undertake, and how to apply results in ways that benefit humanity\u2014and weed out harmful applications, such as using that same technology for bioweapons.
\nBiology is a very clear example of how AI expands our perception that, as humans, we operate in three dimensions. When a task involves high-dimensional data, we need AI\u2019s help. But it makes our human judgment more important.
\nIf you move from science to art, it becomes controversial. In the creative world, many designers and creatives face this question: With AI on the rise, what is the value of creatives? To answer that question, I interviewed two masterminds of creativity and design: Brian Collins and Greg Hahn, both of whom run their own design advertising shops. They are at the top of their game, and no doubt both are using AI. The workflow is becoming like an hourglass. On the top, it\u2019s faster and cheaper to run variations and do research, and at the bottom, it is cheaper and faster to distribute it.
\nBut the narrow waist of the hourglass is our taste, our care for the audiences, and our judgment of what would work, and our understanding of cultures. These are the many things that make one design somehow click with you, and one advertising becomes viral, versus others simply that don\u2019t.
\nHuman judgment continues to be a scarce and valuable asset that we need to practice and own.
\nHow has your personal relationship with AI changed after writing this book?
\nMy personal relationship with AI is becoming intimate but unsentimental. I use AI every day by default, and I trust almost none of it. Writing this book didn\u2019t make me more or less cautious; it simply made me more specific. I started the process expecting to land somewhere on the optimist-to-pessimist spectrum, and I ended up rejecting the spectrum entirely.
\nI think binary trust is lazy trust. The question is, trust it to do what, under which conditions, and with what recourse? I don\u2019t really have a position on AI anymore. What\u2019s most important now is to have a practice, because trust is not practiced. It\u2019s not well-defined, but the most important skill is being able to protect yourself and your enterprise, and unlock AI\u2019s promise.
\nAs a parent, that\u2019s where trust matters most to me. My children talk to Alexa without hesitation. I\u2019m not trying to keep AI away from them. That ship sailed before they were born.
\nJudgment isn\u2019t taught to children in a lecture format. Children watch someone practice it, and then they learn it. I\u2019m trying to make sure my children see me make decisions and hear me say, \u201cLet\u2019s check it out.\u201d I\u2019m genuinely optimistic, but not because I think the systems will get safe on their own.
\nWe talked about accountability, the systematic thinking we need to have, and what agency enterprises and individuals need to take on. The variable that determines how this goes is not only the intelligence and capability of machines; it is also whether we, as humans, keep exercising that judgment in their presence.
\nUnfortunately, I see this less discussed and practiced, and that is the purpose of this book. I trust AI precisely, but not generally. I want my children to inherit the habit of deciding for themselves. That\u2019s really the trust code.
", "summary": "In this edition of Author Talks, McKinsey Global Publishing\u2019s Adam Volk chats with Tiffany Xingyu Wang, president and cofounder of Oasis Consortium, and World Economic Forum Young Global Leader, about The Trust Code: How to Unlock AI\u2019s Promise Without Losing Humanity (Wiley, September 2026). Wang examines why our instincts for trust can misfire when applied to increasingly capable AI systems. She argues that learning to calibrate trust can help individuals and organizations use AI with intention, preserve human judgment, and capture its benefits without sacrificing agency. An edited version of the conversation follows.
\nWhy did you write this book?
\nI wrote this book because I kept watching very smart people ask a question that\u2019s almost too big to be useful: \u201cDo we trust AI?\u201d AI\u2019s already here, and we trust it when we use our credit cards, because fraud detection is running behind the scenes. We trust it when we switch on the light, because various types of software are balancing the grid. But more and more often, AI models are drafting our work, and, increasingly, AI agents are acting on our behalf. The central question of our time is no longer whether to trust AI, but how much trust does a system deserve, under which conditions, and for which tasks?
\nOften, we see that being overly trusting exposes us to harm, and not trusting enough traps us in paralysis when it comes to innovation. It\u2019s only the calibrated trust, which this book advocates for, that enables us to move forward. Professionally, I spent over a decade building trust and safety systems for big enterprises.
\nOften in boardrooms, crisis rooms, and on conference stages, people ask me what trust means in the AI era. I never really had the time to give a holistic answer. When I began writing this book, I was pregnant with my second child, and my son was two years old. That\u2019s when the stakes became even more personal to me.
\nMy son talked to Alexa and rode in the Waymo with me regularly. None of this is remarkable because the next generation will not remember a time when AI was not already embedded in our daily life, making every decision with us and on our behalf.
\nWe may not control how fast AI is advancing, but one thing we still have the time and the power to control is how deliberately we give AI that authority. I searched for a book that could teach me that missing skill, but I couldn\u2019t find it. So I decided to write it.
\nWas there anything that surprised you in the research, writing, or response?
\nWe have a machine\u2014cognitive technology\u2014that sounds warm, very confident, very empathetic, but without any of those signals corresponding to understanding, responsibility, or intentionality that a human would have. This creates a profound evolutionary mismatch: the Stone Age brain we have evolved from is evaluating Space Age systems.
\nThat mismatch is one of the first surprises that I encountered during the overall process. It\u2019s very foundational to understanding many aspects the book covers. The second surprise came from writing the book itself. The faster AI moved, the more convinced I became that the book shouldn\u2019t be about what is happening with AI now but about how to think about whatever happens next.
\nAs you can imagine, it\u2019s terrifying for an author to write about what is probably the fastest-moving technology in history. I wondered whether the book would be outdated at the time it hit bookstores. I tried many ways and spent much effort to really go back to first principles when explaining any concept, any headline in this book, so that people could use it as an evergreen guide to decipher whatever comes next.
\nWhy is it so difficult to define trust when it comes to AI?
\nAfter speaking with many people while researching and interviewing, I found the distinction that cleared this up for me.
\nThere is trustworthiness, which is about the system itself. And we can talk about parameters that matter to AI technology. But in this context of AI, is it truthful enough? Is it competent enough? Is it also benevolently aligned with your personal goals or your enterprise goals well enough? Is it appropriate for that context? That\u2019s about the system.
\nTrust is a different thing. Trust is about us. Am I willing to rely on it?
\nAnother matter is calibration, and whether those two match. A technically very strong system can be under-trusted, and a very charming but unreliable system can be wildly over-trusted. That\u2019s what\u2019s happening everywhere now. Both [approaches] are actually failures.
\nInstead of asking, \u201cDo I trust AI?,\u201d what we should ask is, \u201cWhat vulnerability are you accepting today for what promise in the future and in what particular situation?\u201d Because the definition of trust is that you accept vulnerabilities today in hopes of a promise in the future. Trust remains a very vague sentiment until you can articulate and understand the vulnerabilities you accept today and the goals you want to achieve tomorrow for very specific conditions and situations.
\nBut once you correctly articulate the question and think through it\u2014some frameworks we talk about in the book\u2014trust becomes a decision. It becomes a decision that you can examine, measure, and improve upon.
\nWhen it comes to AI adoption, which areas of trust do you see leaders neglecting the most?
\nI really focus on what trust means through the lens of AI. There are four prerequisites for trust: competence, truthfulness, benevolence, and contextual fit. I chose those four because I constantly saw people mistake competence for trustworthiness.
\nIt\u2019s very important to understand that AI is so capable that it will sound very competent and aligned with you, while it has no reason to be benevolent to maximize your well-being. Understanding that pitfall allows you to have the sense of guardrails you need to constantly think about when using AI. Manifestation is the quadrant that is especially important for enterprise leaders now. Does our behavior match what we say we trust? There are so many research papers now being published that say that. We started pilots, but the ROI is low.
\nIt\u2019s very important to understand that a pilot is a statement of belief. What really shows a state of trust when an enterprise adopts AI or not is whether we change the workflow, whether we train the staff, or whether we change the incentives. Match what you say to what you do, and think about how you use the signals of behaviors to assess the level of trust you delegate to AI.
\nFinally, there are the unseen factors. It\u2019s important to understand that these technologies are meant to be warm, empathetic, and very confident. But they are not earning our trust by delivering the track record and the proven evidence, either by ROI or by reliability over time. They\u2019re simply triggering our trust.
\nHow can we hold AI players accountable when even they can\u2019t predict how their systems will behave?
\nIn today\u2019s AI age, accountability means owning the conditions around unpredictability, and that articulation is extremely important. There is a missing actor in this conversation about AI. We usually draw a map of two posts: AI labs that build models, and regulators that write the rules.
\nBut enterprises decide what gets bought and what\u2019s connected to real systems. We give access to real data and deploy it at scale. Consider any mature industry, for example, aviation. Airplane builders are responsible for designing and testing it, but airlines have the responsibility for how the plane is operated and who controls it.
\nOf course, the regulators set a floor nobody can opt out of. AI needs the same layer. Often, the power and responsibility of enterprises as a whole are underused and underrecognized in AI governance or AI accountability. In other words, enterprise leaders have more agency than we give credit for in this race of AI.
\nIf I\u2019m buying AI for an enterprise, I would stop asking, \u201cIs this model safe?\u201d Instead, I would say, \u201cShow me a failure and show me how you contained it. Show me how quickly I will know [about a failure], and show me what recourse I have if it costs me money or reputation.\u201d
\nEnterprise leaders need to understand the strong agency you have to decide what model you buy, how you deploy it, and what protection measures you take. Hopefully, when enterprises come together and make that the default, the norm, we can see the maturity of this industry.
\nLabs build capabilities and regulators set the floor, but it is the enterprise that decides how much authority we give it.
\nCan you give an example of how AI is expanding human judgment and perception?
\nI interviewed Professor Stephen Quake at Stanford, a pioneer of single-cell genomics. He described biology to me as historically 90 percent experimental and 10 percent computational. The ambition is to reverse it so that it\u2019s 90 percent computational and 10 percent experimental.
\nWe have already seen a glimpse of this with AlphaFold, which used AI to predict more than 20 million protein structures and to understand better how biology works. A cell takes it to a more complex, different level, because every cell has 20,000 protein-coding genes.
\nUnderstanding cells better means we can better understand our body, our health, our diseases, and our therapies. But when it involves running in 20,000 dimensions, that\u2019s beyond human brain capacity. That\u2019s something you can\u2019t do without AI.
\nBut that doesn\u2019t mean human judgment is stripped from the process. Instead, where the judgment is situated shifts. Arguably, the more AI can surface and test hypotheses for us\u2014more cheaply, quickly, and at scale\u2014the more important human judgment becomes. Because ultimately, humans must decide which hypotheses to pursue, which tasks and experiments to undertake, and how to apply results in ways that benefit humanity\u2014and weed out harmful applications, such as using that same technology for bioweapons.
\nBiology is a very clear example of how AI expands our perception that, as humans, we operate in three dimensions. When a task involves high-dimensional data, we need AI\u2019s help. But it makes our human judgment more important.
\nIf you move from science to art, it becomes controversial. In the creative world, many designers and creatives face this question: With AI on the rise, what is the value of creatives? To answer that question, I interviewed two masterminds of creativity and design: Brian Collins and Greg Hahn, both of whom run their own design advertising shops. They are at the top of their game, and no doubt both are using AI. The workflow is becoming like an hourglass. On the top, it\u2019s faster and cheaper to run variations and do research, and at the bottom, it is cheaper and faster to distribute it.
\nBut the narrow waist of the hourglass is our taste, our care for the audiences, and our judgment of what would work, and our understanding of cultures. These are the many things that make one design somehow click with you, and one advertising becomes viral, versus others simply that don\u2019t.
\nHuman judgment continues to be a scarce and valuable asset that we need to practice and own.
\nHow has your personal relationship with AI changed after writing this book?
\nMy personal relationship with AI is becoming intimate but unsentimental. I use AI every day by default, and I trust almost none of it. Writing this book didn\u2019t make me more or less cautious; it simply made me more specific. I started the process expecting to land somewhere on the optimist-to-pessimist spectrum, and I ended up rejecting the spectrum entirely.
\nI think binary trust is lazy trust. The question is, trust it to do what, under which conditions, and with what recourse? I don\u2019t really have a position on AI anymore. What\u2019s most important now is to have a practice, because trust is not practiced. It\u2019s not well-defined, but the most important skill is being able to protect yourself and your enterprise, and unlock AI\u2019s promise.
\nAs a parent, that\u2019s where trust matters most to me. My children talk to Alexa without hesitation. I\u2019m not trying to keep AI away from them. That ship sailed before they were born.
\nJudgment isn\u2019t taught to children in a lecture format. Children watch someone practice it, and then they learn it. I\u2019m trying to make sure my children see me make decisions and hear me say, \u201cLet\u2019s check it out.\u201d I\u2019m genuinely optimistic, but not because I think the systems will get safe on their own.
\nWe talked about accountability, the systematic thinking we need to have, and what agency enterprises and individuals need to take on. The variable that determines how this goes is not only the intelligence and capability of machines; it is also whether we, as humans, keep exercising that judgment in their presence.
\nUnfortunately, I see this less discussed and practiced, and that is the purpose of this book. I trust AI precisely, but not generally. I want my children to inherit the habit of deciding for themselves. That\u2019s really the trust code.
", "image": null, "banner_image": null, "date_published": "2026-10-22T16:00:00Z", "date_modified": "2026-10-22T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}, {"id": "https://www.mckinsey.com/insights/4c91caab-89d2-47b1-bd13-632fa0716e1a", "url": "https://www.mckinsey.com/featured-insights/future-of-asia/asias-energy-reckoning-securing-the-energy-system-to-drive-growth", "external_url": null, "title": "Asia\u2019s energy reckoning: Securing the energy system to drive growth", "content_text": "", "content_html": "Recent disruptions have shone a spotlight on the vulnerabilities of Asia\u2019s energy system. Countries could focus on five priority investment areas to support regional growth and resilience.", "summary": "Recent disruptions have shone a spotlight on the vulnerabilities of Asia\u2019s energy system. Countries could focus on five priority investment areas to support regional growth and resilience.", "image": null, "banner_image": null, "date_published": "2026-10-21T16:00:00Z", "date_modified": "2026-10-21T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}, {"id": "https://www.mckinsey.com/insights/fd32d73e-e7c3-45b9-a211-f46588c2ea31", "url": "https://www.mckinsey.com/uk/our-insights/the-uks-regional-talent-trap-and-how-it-restricts-regional-growth-uk-productivity-series-part-2", "external_url": null, "title": "The UK\u2019s regional talent trap and how it restricts regional growth: UK Productivity Series, part 2", "content_text": "", "content_html": "The United Kingdom\u2019s productivity gap is a management problem, and this management problem is a talent problem, and this talent problem is a geography problem.
\nThe UK\u2019s productivity gap is a well-documented and persistent economic challenge. The most recent OECD data shows the UK generates $86 of GDP per hour worked, a figure materially below that of Germany ($103) and the United States ($105). Recent revisions to growth rates have not changed the fact of the UK\u2019s lagging absolute productivity levels. This gap has significant consequences, reducing the fiscal headroom available to governments to implement policy and acting as a persistent drag on living standards.
\nOur previous piece demonstrated a connection between national GDP per hour worked and the adoption of management systems by a nation\u2019s businesses; it also showed that the United Kingdom is not implementing these management systems as well as comparable economies, even though the practices are well documented and widely understood. The same data exposes a split between London- and non-London-centric industries: London-centric industries are more likely to have implemented better management practices and so would likely benefit from the associated increased productivity.
\nAccording to research on behavioural change and our experience serving clients around the world, a group-level change in behaviour requires that at least 25 percent of a group advocate and behave differently. For a ten-person leadership team in an operational setting, that means a minimum of three people must have the drive, intellectual capability, knowledge, skills, and expertise to change the way an organisation runs day to day. People with this combination of capabilities are scarce. We observe that this top talent is created mainly through two routes: leaders cultivated from the shop floor up and executives with higher education who developed knowledge, skills, and expertise through varied early-career roles.
\nOperationally intensive firms outside London have limited access to the second pool of top talent. High-paying roles account for 40.8 percent of jobs in London, versus a national average of 23.2 percent, and their numbers are growing more than two and a half times faster in London than in the rest of the country. This concentration attracts a large share of the nation\u2019s high-performing graduates\u2014the top talent of tomorrow\u2014because of the real and perceived benefit to their careers. By their late 20s and early 30s, people are often part of a dual-high-earner household, each with high-value jobs. In many cases, only London can offer both partners a high-value job in the same place. Professional identity then makes a trailing partner reluctant to move out of the London economic area even where household economics favour it. As a result, the couple stays and remains accessible only to businesses there.
\nEver more of this type of talent is absorbed into London, while the regions are unable to attract enough of the kind of leaders they need to grow their business, compete globally, and generate wealth locally. Even a modest rebalancing could have a significant impact on British regions\u2019 prosperity while having minimal negative effect on the prosperity of London. But instead, firms outside London can get stuck in a self-perpetuating loop: Unable to recruit top talent, they do not experience the benefit and so struggle to justify the business case to invest in it. This is seen in a decline in operational pay. The median production manager sat at the 86th percentile of UK earners in 1970; today the equivalent role sits at the 67th.
\nGermany and the United States are less exposed on both counts. Their high-value economic activity is distributed; Berlin accounts for 5 percent of German GDP, versus London\u2019s 23 percent. And operational leadership carries higher pay and status, with premiums over median earnings of roughly 65 percent in Germany and 100 percent in the United States, versus 15 to 25 percent in the United Kingdom, once adjusted for measurement inconsistencies.
\nLondon is a high-performing economic region. It competes in industries across the world stage and generates tax revenue that funds public services across the whole country. The answer is not to restrict London\u2019s growth but to lift other areas up so they too can more easily compete on the world stage and return wealth to those regions. McKinsey Global Institute\u2019s Power of One research shows that national productivity growth comes from a small number of standout firms whose leaders make bold strategic moves through strong management practices which allow them to execute at scale. The United Kingdom needs such leaders to implement the management practices in businesses that would unlock growth across the whole nation.
\nWe believe that to increase the regions\u2019 access to top talent while maintaining London\u2019s prosperity, businesses and policymakers would have significant impact by pursuing five approaches:
\nRevalue operational leadership toward US and German pay premiums.
\nBuild prestigious graduate pathways into operations.
\nFormalise the route from shop floor to the top so that future leaders gain broad experience early in their careers.
\nHire for critical leadership mass, rather than individual stars.
\nStimulate economic activity across the country so that high-value roles are not almost exclusively limited to London.
\nThe implication of the talent trap described in this article is uncomfortable. London\u2019s success is a vital part of the national economy and contributes hugely, and this should be celebrated. But this has unforeseen consequences. As London absorbs a large share of the nation\u2019s highest-potential graduates, operationally intensive businesses outside of the London economic area have limited access to this talent pool. These businesses then struggle to reach a critical mass of leaders with the ability to install management systems that will spur productivity, compete more effectively on the world stage, and thus return more wealth to those regions.
\nThe United Kingdom\u2019s productivity problem is, in part, a management problem
\nThe United Kingdom\u2019s productivity gap is a well-documented and persistent challenge. The most recent OECD data shows the country generates $86 of GDP per hour worked, a figure materially below that of Germany ($103) and the United States ($105). This gap has significant consequences, notably reducing the fiscal headroom available to governments to implement policy and acting as a persistent drag on living standards. Recent adjustments by the Office for National Statistics (ONS) to the measurement of UK hours worked have improved the United Kingdom\u2019s position in the G7 productivity growth table, but the fundamental gap in output per hour with Germany and the United States still stands.
\nOur previous piece demonstrated a connection between GDP per hour worked and the adoption of management systems\u2014the mechanisms by which organisations allocate resources, track performance, and drive corrective intervention. That connection is supported both by McKinsey\u2019s recent survey research and by the World Management Survey, which draws on data from nearly 18,000 organisation manufacturing managers in 38 countries and finds that a one-point improvement in management score on its five-point scale is associated with the same output gain as a 40 percent expansion of the labour force or a 60 percent increase in capital.
\nImportantly, our research showed that the United Kingdom implements these management systems less than comparable economies, with a direct bearing on national productivity. Good management systems are not a secret; they are well documented in publicly available literature and deployed by thousands of leading companies worldwide. Something, therefore, is preventing UK businesses from implementing them at scale.
\nThe stakes are higher than the arithmetic of average firms suggests. The McKinsey Global Institute\u2019s Power of One study of 8,300 large companies in Germany, the United Kingdom, and the United States found that fewer than 100 standouts generated two-thirds of the productivity growth in its sample between 2011 and 2019, and that they did so through bold strategic moves executed at scale. Most were not start-ups or superstars but large incumbents that improved year after year. Productivity growth in the UK sample was close to zero, against 2.1 percent a year in the United States.
\nOperationally intensive firms need management systems most and are concentrated outside London
\nOperationally intensive businesses are those whose competitive advantage depends heavily on operational performance\u2014through lower costs, greater agility, or more effective use of assets. The defining characteristic is cost structure, not sector. A pharmaceutical company manufacturing generic drugs at a low price point is operationally intensive; one focused on drug discovery, for which manufacturing is simply a route to market, is not. Volume manufacturing, food and drink production, and logistics are typical examples, and they are more prevalent outside London. These are the businesses that stand to gain most from the management systems we know drive productivity.
\nIn firms where operations account for a large share of the cost base, marginal improvements in how work is planned, executed, and measured are competitively decisive. Our analysis shows that there is a \u201cregional talent trap\u201d that is preventing these operationally intensive businesses from tapping into the talent needed to implement strong management systems, which in turn will allow them to become more productive, better compete on the world stage, and return wealth to the region of the United Kingdom they operate in.
\nTwo types of leaders can transform an organisation\u2019s operations
\nModern management systems are both complex and specific to the organisation in which they are implemented. They require leaders who combine intellectual capability and drive with the knowledge, skills, and expertise of modern operational management. These individuals come from all walks of life, all career backgrounds, and every region of the country. Early in their career, they go through many roles that allow them to build the knowledge, skills, and expertise required to lead modern operational business.
\nIn our experience, organisations acquire such leaders through one of two main routes: by cultivating leaders from the shop floor up and by recruiting executives with higher education who have developed knowledge, skills, and expertise through varied early-career roles.
\nIn cultivation from the shop floor up, an individual joins an organisation early (for example through a technical apprenticeship) and builds deep operational knowledge, skills, and expertise needed to lead through successive roles\u2014either by promotion within the same business or by changing employers within the same industry. In operational businesses around the world, we consistently see that this route can generate very strong leaders able to drive real change in businesses and industries.
\nIn the graduate-professional route, individuals typically hold a degree, often in a subject related to their later work. They then accumulate knowledge, skills, and expertise through multiple employment experiences during their 20s and 30s, sometimes accelerated by a formal graduate scheme, often moving rapidly through roles within the same organisation or in different ones.
\nThe routes differ in their starting point, but the way this top talent is made is fundamentally the same: People with intellect and drive build capability through experience across a variety of roles. These are the people we see able to transform large parts of operational organisations by implementing management systems which then improve their organisation\u2019s productivity and therefore global competitiveness. Both routes are valuable and produce world-class leaders. In contrast, where organisations fail to cultivate this critical mass of talent, even capable leaders are reduced to caretaking\u2014holding the status quo rather than transforming it.
\nThe leadership critical mass needed to change an organisation
\nOperationally intensive businesses are typically organised into operational units. As an example, they might have 500 people working within each unit. The structure varies\u2014a manufacturing site, a cluster of retail stores, a distribution centre, a single call centre\u2014but the scale is similar. To transform the productivity of one of these units for the long term requires implementing a management system and making it stick, usually in the face of an organisation\u2019s natural inertia. Effective management systems are not secret; they have been well documented over the last 50-plus years. Despite this, businesses struggle to implement these approaches within their increasingly complex operational environments.
\nImplementing a management system is a collective act, requiring a critical mass of highly capable leadership. Putting one excellent leader into a business without a similarly capable team is unlikely to significantly improve how the organisation operates (see sidebar, \u201cCritical mass in operational leadership: A tale of two sites\u201d). This insight from our experience in organisations around the world appears consistent with a broad body of evidence from the social sciences. Fifty years of empirical research on organisational and community change has consistently found that a committed minority can shift group norms, but only once it reaches a threshold estimated at roughly 25 percent\u2014a figure corroborated by more recent experimental work, including research published in Science in 2018.
\nRegional businesses have not implemented management systems as well as London-based industries
\nA striking pattern in our own survey data suggests a disparity in management system implementation between London-centric industries and others. In McKinsey\u2019s 2025 Operational Excellence Survey, organisations in London-centric industries\u2014such as financial services and technology\u2014reported a score of 5.90 out of 7 on management system adoption, putting them at the 58th percentile; those in non-London-centric industries\u2014such as manufacturing and natural gas\u2014scored 5.09, placing them at the 19th percentile (Exhibit 1). Put plainly, against the German, US, and overall UK respondents, London-centric industries in the UK score better than more than half of all respondents, while non-London-centric industries score worse than most. The sample sizes are small, the scores are self-reported assessments, and the split is also by industry rather than location, so we infer the London link from where those industries concentrate.
\nWhile granular empirical evidence is relatively sparse, these findings are consistent with, for example, the Office for National Statistics regional data on UK management practice scores. This indicates that firms in the same sectors but located in the South East, London, and Scotland are much more likely to have structured management practices than those in, for example, Yorkshire and the Humber, the North West, or West Midlands.
\nThese differences in the adoption of management systems between UK industries are echoed in the productivity and global competitiveness differences between London and non-London sectors. London is a top global financial centre, for example, with UK financial services contributing more than \u00a3200 billion to the economy in 2025. In contrast, UK manufacturing slipped from 11th to 15th in the UN Industrial Development Organization\u2019s Competitive Industrial Performance Index during the first two decades of the 2000s. ONS data shows Germany to be approximately 16 percent more productive per hour worked in manufacturing, with UK output per hour roughly 10 percent below the G7 average.
\nThe Power of One study, cited earlier, asks which talent strategies \u201ccan nurture \u2026 managerial leaders who can make bold strategic moves\u201d\u2014the kinds of leaders who create standout firms. Our answer is that the United Kingdom already has such leaders, but the regional talent trap keeps them away from the operationally intensive businesses in the regions that could become the next standouts.
\nThe regional talent trap
\nOur analysis shows that the structure of the UK economy is incentivizing its graduate talent (which originates from the whole nation) to gravitate to and then remain in London and the surrounding economic area. For organisations outside London, this removes one of the two routes to generate top talent, making it harder to form the critical mass of leadership needed to implement the management systems they need to become more productive, compete better on the world stage, and provide these regions with better-paying jobs.
\nDissecting the regional talent trap
\nThe regional talent trap can be broken down into five interrelated parts:
\nHigh-value roles in the United Kingdom are heavily concentrated in London. High-value roles, defined by the Economics Observatory as highly skilled, well-paid roles, are concentrated in London and the South East. ONS data shows that 40.8 percent of jobs in London are high-paying\u2014nearly double the national average of 23.2 percent. And the gap is widening. Institute for Fiscal Studies analysis shows that employment in high-paying occupations grew 240 percent in Inner London between 1993 and 2022, versus 95 percent across the country as a whole.
\nHigh-value roles are a magnet for high-performing graduates. The concentration of these roles in London and the South East exerts a strong pull on academic high performers at the point of graduation. At the start of a career, it is a rare individual in this group who can transform a business, but some can later become the top talent described earlier who can. Research by the Centre for Cities found that 22 percent of all graduates who moved to a new city after university chose London, and the likelihood of choosing London increased along with the prestige of the graduate\u2019s university (Exhibit 2). Thirty-eight percent of graduates with a first-class or upper second-class degree from a Russell Group university who moved within six months of graduation chose London, alongside 52 percent of equivalent Oxbridge (Oxford or Cambridge) graduates.
\nThis is not a claim that selective universities make better leaders; it is a description of the natural graduate flow. Once graduates are in London, job concentration enables them to jump to different roles to develop professional experience early in their career with little personal disruption. Someone in London can have six different roles in the first decade of their career without having to move or undergo any geographic relocation. Outside London, this is much rarer, as fewer areas offer that density of transfer opportunities in a similar sector.
\nAssortative mating creates \u2018power couples\u2019 where both are in high-value roles. People build friendship groups and professional networks with others who have similar jobs, earning power, and educational backgrounds. Also, they often find long-term romantic partners within those same groups. As a result, many high-performing graduates (regardless of sexual orientation or gender) partner with other high-performing graduates, becoming a \u201cpower couple,\u201d a term coined by American sociologists, and both of them work in London in high-value roles during the early years of their careers.
\nThe United Kingdom\u2019s economic structure makes it hard for such a couple to maintain their careers outside London and its surroundings. Once someone in this group has a similar partner, the two of them become \u201ctrapped\u201d in the London economic zone. It is well documented that professional power couples gravitate to metropolitan areas economically large enough to support two high-value jobs.
\nMany may want to live outside London but are unwilling to move due to the perceived career impact. Ipsos polling from July 2026 captures this precisely: Asked whether they would rather live in London or Manchester, Britons more often selected Manchester (26 percent, versus 21 percent for London), due to cost of living, friendliness, and community. However, London decisively wins on job opportunities and as the place where \u201cmore successful people\u201d live and work.
\nFor this group, a career is more than a source of income. Professional identity is formed in one\u2019s 20s and 30s. For some, a career is not just what they do but who they are. Someone working at a prestigious law firm does not view themselves as \u201csomeone who works at a law firm\u201d; they are \u201ca lawyer.\u201d If they choose to leave London as a trailing partner, they are unlikely to find a high-value role in the same location as their partner and so would have to stop, pause, or move backward in their career. The effect on their sense of self can outweigh a net improvement in living standards for the couple. Put simply, a couple may have a better standard of living outside of London (the reduction in cost of living may be more than offsetting the reduction in income of the trailing partner), yet they still would choose not to move, due to the psychological impact of slowing their career.
\nA significant part of the nation\u2019s graduate top talent, therefore, choose not to leave London, even if they might have a better standard of living elsewhere. Large numbers of people move out of London in the 25-to-35 age bracket, but most move to adjacent regions and are still economically active within London itself (referred to in this article as the London economic area). The pull of London and its stickiness act as a brake on the mobility of talent into other parts of the country. That in turn could prevent operationally intensive firms from accessing the talent they need to implement effective management systems and increase their productivity.
\nFirms outside of London can face a talent challenge
\nIndustries outside of London feel this impact and often struggle to recruit top talent that can transform their business\u2019s productivity. Half of manufacturers say they cannot source the talent their business needs locally, for example, and 97.5 percent say hiring and retaining skilled people poses a significant challenge to their business. Note that although this talent issue affects many different sectors outside London, manufacturing has some of the cleanest data on the issue. The risk is clear: Organisations outside London then do not see the benefits of productivity improvements and therefore do not see the investment case to acquire top talent at higher levels of remuneration. This perpetuates lower adoption of the management practices, which undermines their ability to compete on the world stage.
\nThe pay and status of operational roles have fallen in relative terms over the last 50 years. In 1970, for example, the median production manager earned enough to sit at the 86th percentile of the national income distribution\u2014a salary equivalent to around \u00a368,000 a year in 2026. Today the median production manager has slipped to the 67th percentile, with an annual income of around \u00a348,000.
\nThe change is even starker at the top of the profession. In 1970, the best-paid 20 percent of production managers sat roughly at the 98th percentile of all earners, a position equivalent to about \u00a3133,000 a year in 2026. Today, the equivalent top-earning production managers sit at the 89th percentile, with an annual income of around \u00a374,000. The equivalent pay commanded by the top of the profession has fallen by almost half.
\nA decline of this kind is both a symptom and a cause of falling status. When a career\u2019s pay ceiling drops away from the top of society\u2019s income scale, the role loses its pull on people choosing a career, making it harder (but not impossible) for those roles to attract the ambitious future leaders who could drive the industry forward. Unlike in 1970, few people who command the highest salaries in the United Kingdom are in operational leadership roles.
\nSome firms break this pattern by offering prestigious formal operational leadership programs that capture talent at the point of graduation. However, these seem largely restricted to foreign-owned multinationals and a small number of global UK firms. This appears consistent with the World Management Survey finding that US-owned firms in the United Kingdom tend to import stronger operational management practices.
\nWithin the United Kingdom, it seems that the London-centric industries are better at developing formal graduate schemes. Our informal review of UK leadership graduate schemes shows a ratio of approximately six to one in favour of London-centric industries over non-London-centric operational ones.
\nWhy conditions in Germany and the United States seem more favourable
\nThe structural forces described here appear less pronounced in Germany and the United States, the comparison economies we are using in this series. This is significantly influenced by differences in the geographic concentration of economic activity and by the relative pay and status of operational roles.
\nGerman and US economies are less concentrated
\nIn Germany, economic activity is less geographically concentrated than in the United Kingdom. Berlin, the largest city, contributes only 5 percent of the country\u2019s GDP, compared with London\u2019s 23 percent share. With a distributed industrial base and thousands of hidden champion small and medium-size enterprises distributed across the country, a highly educated couple in Germany have more options to build dual careers outside the capital city. Although Germany\u2019s industrial economy has challenges of its own today, the distribution of its talent is not one of them\u2014and the productivity gap between the United Kingdom and Germany remains significant.
\nThe United States has a genuine multi-hub economy. A dual-career household can build two high-value careers not only in the coastal centres of Boston, New York, San Francisco, and Washington, DC, but also in Atlanta, Austin, Chicago, Denver, Houston, Seattle, and beyond. Furthermore, the interior metros such as Columbus, Nashville, and St. Louis have all grown their shares of graduates and knowledge workers in recent years. Collectively, US cities offer ambitious graduates a range of viable options that the United Kingdom cannot match.
\nOperational roles have higher status and pay
\nOperational roles in Germany and the United States often command higher pay and social status. In healthcare, for example, a UK healthcare general manager earns approximately \u00a358,000, while a medical and health services manager in the United States earns a median $117,960 (\u00a387,000). Pay is an imperfect proxy for status, but the gap is consistent with operational roles retaining a standing abroad that they have lost in the United Kingdom.
\nProduction managers are paid a premium\u2014the extra pay versus national median that a role commands for its responsibility and the scarcity of the skills needed to do it. Germany pays a premium of 65 percent, the United States 100 percent, and according to official statistics, the United Kingdom pays 50 percent. But the UK figure is inflated: The ONS category includes highly paid engineer and managing directors and other senior leaders whom the American and German statistics classify elsewhere. When we adjust the UK premium using advertised salaries for production managers, it falls to between 15 and 25 percent\u2014less than half of Germany\u2019s premium and a quarter of the US premium.
\nThe entry requirements for a production manager position tell a similar story. The ONS says of the UK role that \u201cthere are no pre-set entry standards\u201d and that entry is possible \u201cwith either a degree or equivalent qualification, and/or relevant experience.\u201d The US Bureau of Labor Statistics lists a bachelor\u2019s degree and five or more years of related experience as the typical route in. Germany classifies the role at its highest skill level, which normally means a university degree, usually in engineering. Britain treats as an occupation what the United States and Germany treat as a profession.
\nIn Germany this premium may reflect the high cultural status of operations management and engineering. Germany\u2019s leading engineering universities\u2014TU Munich (TUM) and RWTH Aachen among them\u2014have institutionalised relationships with the country\u2019s most prestigious industrial employers. TUM\u2019s formal Partners of Excellence include BMW, Bosch, Siemens, and Volkswagen, and RWTH Aachen alumni consistently fill some of Germany\u2019s most senior DAX executive positions.
\nA similar dynamic exists in the United States, where Boeing and the GE companies rank among the top dozen employers of Georgia Tech mechanical engineers. Caterpillar and John Deere recruit there, too.
\nIn the United Kingdom, by contrast, perceptions of operational roles lag those of other sectors. A Nestl\u00e9 study of 2,000 young people found that just 4 percent would consider a career in manufacturing.
\nGraduate schemes develop and attract potential future top talent at the point of graduation
\nBoth countries also have more established graduate pathways into operational leadership and treat these pathways as a credible route to the very top. In Germany, the chief executives of all three major carmakers started on their companies\u2019 graduate programs. Oliver Blume joined Audi\u2019s trainee program in 1994 and has led Volkswagen Group since 2022. Ola K\u00e4llenius joined Daimler-Benz as a trainee in 1993 and has led Mercedes-Benz Group since 2019. Oliver Zipse joined BMW as a trainee in 1991 and led the company from 2019 until May 2026, when he handed over to Milan Nedeljkovi\u0107, who joined BMW as a trainee in 1993.
\nIn the United States, Procter & Gamble hires at entry level and builds from within. The company reports that more than 99 percent of its senior leaders began their careers there. David Taylor, who served as the chairman, president, and CEO from 2015 to 2021, joined in 1980 as a production manager. GE Aerospace\u2019s two-year leadership programs follow the same logic: more than 25 percent of its top leaders began their careers on one.
\nWhat company leaders and policymakers could do
\nLondon is a high-performing economic region. Its industries compete across the world stage and generate tax revenue that funds public services across the whole country. London and the South East (broadly, the London economic area) are the only two regions that pay more into the Treasury than they receive in public spending. The answer is not to restrict London\u2019s economic activity but to lift other areas up so they too can better compete on the world stage and return wealth to those regions.
\nReversing the structural forces described in this article would entail deliberate actions by company leaders and policymakers.
\nCompany leaders
\nRevalue operational leadership. Raise the pay, profile, and prestige of operational leadership roles to shift their perceived status. A premium comparable to the 65 percent seen in Germany or the 100 percent in the United States could begin to change the career calculus for high-potential graduates.
\nBuild prestigious graduate pathways. Foster more industrial graduate programs\u2014structured, high-status routes into operational leadership that are explicitly positioned as a pathway to senior corporate leadership.
\nFormalise the route from the shop floor to the top. Make technical apprenticeships and progression from the front line into site and operations leadership an explicit pathway, so potential future leaders can gain a wide variety of experience early in their career. This can maximise the ability of operational businesses to generate top talent from within.
\nTarget critical leadership mass within organisations. Focus on building the collective operational capability of leadership teams, not just on individual hires, to create the conditions and momentum for management systems to take root.
\nPolicymakers
\nStimulate high-value economic activity. Growing high-value jobs in regional cities through sustained investment in those cities and their transport could increase the effective density of roles to encourage the top-talent couples to pursue careers away from London. The mechanism described in this article would indicate that the lack of implementation of strong management systems is caused by regional businesses\u2019 restricted access to national talent, which is in turn caused by the regional talent trap. Therefore, improving UK productivity will depend on not just creating high-value jobs outside London but also ensuring that the talent needed to embed these systems is distributed more evenly across the economy.
\nTo get access to this talent, regions outside London need to offer top jobs and attract top talent to fill them. Just moving the job will not fix the problem. These leaders are created over decades, building knowledge, skills, and expertise by working in a wide variety of roles within a specific industry. Transforming regions into attractive talent hubs will mean reversing a talent trap trend that has developed over 50 years and has been reinforced by economic, psychological, and social factors.
\nThe implication is an uncomfortable one. London and the South East are remarkable economic assets, competing on the world stage and contributing hugely to the UK\u2019s prosperity. But if leaders with the drive and capability to transform British industry come from two main routes\u2014the shop floor and the graduate route\u2014then one region is absorbing a very large share of the second group. Every year, graduates who are the possible leaders of tomorrow head to the capital and its surroundings. Then, for social and economic reasons, many never leave.
\nRebalancing this even slightly would give more operationally intensive businesses of the Midlands, the North, South West, East, Wales, Scotland, and Northern Ireland the opportunity to reach the critical mass of leaders they need to install the management systems that would make them more productive, more globally competitive, and more profitable. If just some of the talent London attracts could be deployed to lead businesses elsewhere, the overall effect on the country could be significant\u2014to the postcodes in which these businesses operate and to the United Kingdom as a whole.
", "summary": "The United Kingdom\u2019s productivity gap is a management problem, and this management problem is a talent problem, and this talent problem is a geography problem.
\nThe UK\u2019s productivity gap is a well-documented and persistent economic challenge. The most recent OECD data shows the UK generates $86 of GDP per hour worked, a figure materially below that of Germany ($103) and the United States ($105). Recent revisions to growth rates have not changed the fact of the UK\u2019s lagging absolute productivity levels. This gap has significant consequences, reducing the fiscal headroom available to governments to implement policy and acting as a persistent drag on living standards.
\nOur previous piece demonstrated a connection between national GDP per hour worked and the adoption of management systems by a nation\u2019s businesses; it also showed that the United Kingdom is not implementing these management systems as well as comparable economies, even though the practices are well documented and widely understood. The same data exposes a split between London- and non-London-centric industries: London-centric industries are more likely to have implemented better management practices and so would likely benefit from the associated increased productivity.
\nAccording to research on behavioural change and our experience serving clients around the world, a group-level change in behaviour requires that at least 25 percent of a group advocate and behave differently. For a ten-person leadership team in an operational setting, that means a minimum of three people must have the drive, intellectual capability, knowledge, skills, and expertise to change the way an organisation runs day to day. People with this combination of capabilities are scarce. We observe that this top talent is created mainly through two routes: leaders cultivated from the shop floor up and executives with higher education who developed knowledge, skills, and expertise through varied early-career roles.
\nOperationally intensive firms outside London have limited access to the second pool of top talent. High-paying roles account for 40.8 percent of jobs in London, versus a national average of 23.2 percent, and their numbers are growing more than two and a half times faster in London than in the rest of the country. This concentration attracts a large share of the nation\u2019s high-performing graduates\u2014the top talent of tomorrow\u2014because of the real and perceived benefit to their careers. By their late 20s and early 30s, people are often part of a dual-high-earner household, each with high-value jobs. In many cases, only London can offer both partners a high-value job in the same place. Professional identity then makes a trailing partner reluctant to move out of the London economic area even where household economics favour it. As a result, the couple stays and remains accessible only to businesses there.
\nEver more of this type of talent is absorbed into London, while the regions are unable to attract enough of the kind of leaders they need to grow their business, compete globally, and generate wealth locally. Even a modest rebalancing could have a significant impact on British regions\u2019 prosperity while having minimal negative effect on the prosperity of London. But instead, firms outside London can get stuck in a self-perpetuating loop: Unable to recruit top talent, they do not experience the benefit and so struggle to justify the business case to invest in it. This is seen in a decline in operational pay. The median production manager sat at the 86th percentile of UK earners in 1970; today the equivalent role sits at the 67th.
\nGermany and the United States are less exposed on both counts. Their high-value economic activity is distributed; Berlin accounts for 5 percent of German GDP, versus London\u2019s 23 percent. And operational leadership carries higher pay and status, with premiums over median earnings of roughly 65 percent in Germany and 100 percent in the United States, versus 15 to 25 percent in the United Kingdom, once adjusted for measurement inconsistencies.
\nLondon is a high-performing economic region. It competes in industries across the world stage and generates tax revenue that funds public services across the whole country. The answer is not to restrict London\u2019s growth but to lift other areas up so they too can more easily compete on the world stage and return wealth to those regions. McKinsey Global Institute\u2019s Power of One research shows that national productivity growth comes from a small number of standout firms whose leaders make bold strategic moves through strong management practices which allow them to execute at scale. The United Kingdom needs such leaders to implement the management practices in businesses that would unlock growth across the whole nation.
\nWe believe that to increase the regions\u2019 access to top talent while maintaining London\u2019s prosperity, businesses and policymakers would have significant impact by pursuing five approaches:
\nRevalue operational leadership toward US and German pay premiums.
\nBuild prestigious graduate pathways into operations.
\nFormalise the route from shop floor to the top so that future leaders gain broad experience early in their careers.
\nHire for critical leadership mass, rather than individual stars.
\nStimulate economic activity across the country so that high-value roles are not almost exclusively limited to London.
\nThe implication of the talent trap described in this article is uncomfortable. London\u2019s success is a vital part of the national economy and contributes hugely, and this should be celebrated. But this has unforeseen consequences. As London absorbs a large share of the nation\u2019s highest-potential graduates, operationally intensive businesses outside of the London economic area have limited access to this talent pool. These businesses then struggle to reach a critical mass of leaders with the ability to install management systems that will spur productivity, compete more effectively on the world stage, and thus return more wealth to those regions.
\nThe United Kingdom\u2019s productivity problem is, in part, a management problem
\nThe United Kingdom\u2019s productivity gap is a well-documented and persistent challenge. The most recent OECD data shows the country generates $86 of GDP per hour worked, a figure materially below that of Germany ($103) and the United States ($105). This gap has significant consequences, notably reducing the fiscal headroom available to governments to implement policy and acting as a persistent drag on living standards. Recent adjustments by the Office for National Statistics (ONS) to the measurement of UK hours worked have improved the United Kingdom\u2019s position in the G7 productivity growth table, but the fundamental gap in output per hour with Germany and the United States still stands.
\nOur previous piece demonstrated a connection between GDP per hour worked and the adoption of management systems\u2014the mechanisms by which organisations allocate resources, track performance, and drive corrective intervention. That connection is supported both by McKinsey\u2019s recent survey research and by the World Management Survey, which draws on data from nearly 18,000 organisation manufacturing managers in 38 countries and finds that a one-point improvement in management score on its five-point scale is associated with the same output gain as a 40 percent expansion of the labour force or a 60 percent increase in capital.
\nImportantly, our research showed that the United Kingdom implements these management systems less than comparable economies, with a direct bearing on national productivity. Good management systems are not a secret; they are well documented in publicly available literature and deployed by thousands of leading companies worldwide. Something, therefore, is preventing UK businesses from implementing them at scale.
\nThe stakes are higher than the arithmetic of average firms suggests. The McKinsey Global Institute\u2019s Power of One study of 8,300 large companies in Germany, the United Kingdom, and the United States found that fewer than 100 standouts generated two-thirds of the productivity growth in its sample between 2011 and 2019, and that they did so through bold strategic moves executed at scale. Most were not start-ups or superstars but large incumbents that improved year after year. Productivity growth in the UK sample was close to zero, against 2.1 percent a year in the United States.
\nOperationally intensive firms need management systems most and are concentrated outside London
\nOperationally intensive businesses are those whose competitive advantage depends heavily on operational performance\u2014through lower costs, greater agility, or more effective use of assets. The defining characteristic is cost structure, not sector. A pharmaceutical company manufacturing generic drugs at a low price point is operationally intensive; one focused on drug discovery, for which manufacturing is simply a route to market, is not. Volume manufacturing, food and drink production, and logistics are typical examples, and they are more prevalent outside London. These are the businesses that stand to gain most from the management systems we know drive productivity.
\nIn firms where operations account for a large share of the cost base, marginal improvements in how work is planned, executed, and measured are competitively decisive. Our analysis shows that there is a \u201cregional talent trap\u201d that is preventing these operationally intensive businesses from tapping into the talent needed to implement strong management systems, which in turn will allow them to become more productive, better compete on the world stage, and return wealth to the region of the United Kingdom they operate in.
\nTwo types of leaders can transform an organisation\u2019s operations
\nModern management systems are both complex and specific to the organisation in which they are implemented. They require leaders who combine intellectual capability and drive with the knowledge, skills, and expertise of modern operational management. These individuals come from all walks of life, all career backgrounds, and every region of the country. Early in their career, they go through many roles that allow them to build the knowledge, skills, and expertise required to lead modern operational business.
\nIn our experience, organisations acquire such leaders through one of two main routes: by cultivating leaders from the shop floor up and by recruiting executives with higher education who have developed knowledge, skills, and expertise through varied early-career roles.
\nIn cultivation from the shop floor up, an individual joins an organisation early (for example through a technical apprenticeship) and builds deep operational knowledge, skills, and expertise needed to lead through successive roles\u2014either by promotion within the same business or by changing employers within the same industry. In operational businesses around the world, we consistently see that this route can generate very strong leaders able to drive real change in businesses and industries.
\nIn the graduate-professional route, individuals typically hold a degree, often in a subject related to their later work. They then accumulate knowledge, skills, and expertise through multiple employment experiences during their 20s and 30s, sometimes accelerated by a formal graduate scheme, often moving rapidly through roles within the same organisation or in different ones.
\nThe routes differ in their starting point, but the way this top talent is made is fundamentally the same: People with intellect and drive build capability through experience across a variety of roles. These are the people we see able to transform large parts of operational organisations by implementing management systems which then improve their organisation\u2019s productivity and therefore global competitiveness. Both routes are valuable and produce world-class leaders. In contrast, where organisations fail to cultivate this critical mass of talent, even capable leaders are reduced to caretaking\u2014holding the status quo rather than transforming it.
\nThe leadership critical mass needed to change an organisation
\nOperationally intensive businesses are typically organised into operational units. As an example, they might have 500 people working within each unit. The structure varies\u2014a manufacturing site, a cluster of retail stores, a distribution centre, a single call centre\u2014but the scale is similar. To transform the productivity of one of these units for the long term requires implementing a management system and making it stick, usually in the face of an organisation\u2019s natural inertia. Effective management systems are not secret; they have been well documented over the last 50-plus years. Despite this, businesses struggle to implement these approaches within their increasingly complex operational environments.
\nImplementing a management system is a collective act, requiring a critical mass of highly capable leadership. Putting one excellent leader into a business without a similarly capable team is unlikely to significantly improve how the organisation operates (see sidebar, \u201cCritical mass in operational leadership: A tale of two sites\u201d). This insight from our experience in organisations around the world appears consistent with a broad body of evidence from the social sciences. Fifty years of empirical research on organisational and community change has consistently found that a committed minority can shift group norms, but only once it reaches a threshold estimated at roughly 25 percent\u2014a figure corroborated by more recent experimental work, including research published in Science in 2018.
\nRegional businesses have not implemented management systems as well as London-based industries
\nA striking pattern in our own survey data suggests a disparity in management system implementation between London-centric industries and others. In McKinsey\u2019s 2025 Operational Excellence Survey, organisations in London-centric industries\u2014such as financial services and technology\u2014reported a score of 5.90 out of 7 on management system adoption, putting them at the 58th percentile; those in non-London-centric industries\u2014such as manufacturing and natural gas\u2014scored 5.09, placing them at the 19th percentile (Exhibit 1). Put plainly, against the German, US, and overall UK respondents, London-centric industries in the UK score better than more than half of all respondents, while non-London-centric industries score worse than most. The sample sizes are small, the scores are self-reported assessments, and the split is also by industry rather than location, so we infer the London link from where those industries concentrate.
\nWhile granular empirical evidence is relatively sparse, these findings are consistent with, for example, the Office for National Statistics regional data on UK management practice scores. This indicates that firms in the same sectors but located in the South East, London, and Scotland are much more likely to have structured management practices than those in, for example, Yorkshire and the Humber, the North West, or West Midlands.
\nThese differences in the adoption of management systems between UK industries are echoed in the productivity and global competitiveness differences between London and non-London sectors. London is a top global financial centre, for example, with UK financial services contributing more than \u00a3200 billion to the economy in 2025. In contrast, UK manufacturing slipped from 11th to 15th in the UN Industrial Development Organization\u2019s Competitive Industrial Performance Index during the first two decades of the 2000s. ONS data shows Germany to be approximately 16 percent more productive per hour worked in manufacturing, with UK output per hour roughly 10 percent below the G7 average.
\nThe Power of One study, cited earlier, asks which talent strategies \u201ccan nurture \u2026 managerial leaders who can make bold strategic moves\u201d\u2014the kinds of leaders who create standout firms. Our answer is that the United Kingdom already has such leaders, but the regional talent trap keeps them away from the operationally intensive businesses in the regions that could become the next standouts.
\nThe regional talent trap
\nOur analysis shows that the structure of the UK economy is incentivizing its graduate talent (which originates from the whole nation) to gravitate to and then remain in London and the surrounding economic area. For organisations outside London, this removes one of the two routes to generate top talent, making it harder to form the critical mass of leadership needed to implement the management systems they need to become more productive, compete better on the world stage, and provide these regions with better-paying jobs.
\nDissecting the regional talent trap
\nThe regional talent trap can be broken down into five interrelated parts:
\nHigh-value roles in the United Kingdom are heavily concentrated in London. High-value roles, defined by the Economics Observatory as highly skilled, well-paid roles, are concentrated in London and the South East. ONS data shows that 40.8 percent of jobs in London are high-paying\u2014nearly double the national average of 23.2 percent. And the gap is widening. Institute for Fiscal Studies analysis shows that employment in high-paying occupations grew 240 percent in Inner London between 1993 and 2022, versus 95 percent across the country as a whole.
\nHigh-value roles are a magnet for high-performing graduates. The concentration of these roles in London and the South East exerts a strong pull on academic high performers at the point of graduation. At the start of a career, it is a rare individual in this group who can transform a business, but some can later become the top talent described earlier who can. Research by the Centre for Cities found that 22 percent of all graduates who moved to a new city after university chose London, and the likelihood of choosing London increased along with the prestige of the graduate\u2019s university (Exhibit 2). Thirty-eight percent of graduates with a first-class or upper second-class degree from a Russell Group university who moved within six months of graduation chose London, alongside 52 percent of equivalent Oxbridge (Oxford or Cambridge) graduates.
\nThis is not a claim that selective universities make better leaders; it is a description of the natural graduate flow. Once graduates are in London, job concentration enables them to jump to different roles to develop professional experience early in their career with little personal disruption. Someone in London can have six different roles in the first decade of their career without having to move or undergo any geographic relocation. Outside London, this is much rarer, as fewer areas offer that density of transfer opportunities in a similar sector.
\nAssortative mating creates \u2018power couples\u2019 where both are in high-value roles. People build friendship groups and professional networks with others who have similar jobs, earning power, and educational backgrounds. Also, they often find long-term romantic partners within those same groups. As a result, many high-performing graduates (regardless of sexual orientation or gender) partner with other high-performing graduates, becoming a \u201cpower couple,\u201d a term coined by American sociologists, and both of them work in London in high-value roles during the early years of their careers.
\nThe United Kingdom\u2019s economic structure makes it hard for such a couple to maintain their careers outside London and its surroundings. Once someone in this group has a similar partner, the two of them become \u201ctrapped\u201d in the London economic zone. It is well documented that professional power couples gravitate to metropolitan areas economically large enough to support two high-value jobs.
\nMany may want to live outside London but are unwilling to move due to the perceived career impact. Ipsos polling from July 2026 captures this precisely: Asked whether they would rather live in London or Manchester, Britons more often selected Manchester (26 percent, versus 21 percent for London), due to cost of living, friendliness, and community. However, London decisively wins on job opportunities and as the place where \u201cmore successful people\u201d live and work.
\nFor this group, a career is more than a source of income. Professional identity is formed in one\u2019s 20s and 30s. For some, a career is not just what they do but who they are. Someone working at a prestigious law firm does not view themselves as \u201csomeone who works at a law firm\u201d; they are \u201ca lawyer.\u201d If they choose to leave London as a trailing partner, they are unlikely to find a high-value role in the same location as their partner and so would have to stop, pause, or move backward in their career. The effect on their sense of self can outweigh a net improvement in living standards for the couple. Put simply, a couple may have a better standard of living outside of London (the reduction in cost of living may be more than offsetting the reduction in income of the trailing partner), yet they still would choose not to move, due to the psychological impact of slowing their career.
\nA significant part of the nation\u2019s graduate top talent, therefore, choose not to leave London, even if they might have a better standard of living elsewhere. Large numbers of people move out of London in the 25-to-35 age bracket, but most move to adjacent regions and are still economically active within London itself (referred to in this article as the London economic area). The pull of London and its stickiness act as a brake on the mobility of talent into other parts of the country. That in turn could prevent operationally intensive firms from accessing the talent they need to implement effective management systems and increase their productivity.
\nFirms outside of London can face a talent challenge
\nIndustries outside of London feel this impact and often struggle to recruit top talent that can transform their business\u2019s productivity. Half of manufacturers say they cannot source the talent their business needs locally, for example, and 97.5 percent say hiring and retaining skilled people poses a significant challenge to their business. Note that although this talent issue affects many different sectors outside London, manufacturing has some of the cleanest data on the issue. The risk is clear: Organisations outside London then do not see the benefits of productivity improvements and therefore do not see the investment case to acquire top talent at higher levels of remuneration. This perpetuates lower adoption of the management practices, which undermines their ability to compete on the world stage.
\nThe pay and status of operational roles have fallen in relative terms over the last 50 years. In 1970, for example, the median production manager earned enough to sit at the 86th percentile of the national income distribution\u2014a salary equivalent to around \u00a368,000 a year in 2026. Today the median production manager has slipped to the 67th percentile, with an annual income of around \u00a348,000.
\nThe change is even starker at the top of the profession. In 1970, the best-paid 20 percent of production managers sat roughly at the 98th percentile of all earners, a position equivalent to about \u00a3133,000 a year in 2026. Today, the equivalent top-earning production managers sit at the 89th percentile, with an annual income of around \u00a374,000. The equivalent pay commanded by the top of the profession has fallen by almost half.
\nA decline of this kind is both a symptom and a cause of falling status. When a career\u2019s pay ceiling drops away from the top of society\u2019s income scale, the role loses its pull on people choosing a career, making it harder (but not impossible) for those roles to attract the ambitious future leaders who could drive the industry forward. Unlike in 1970, few people who command the highest salaries in the United Kingdom are in operational leadership roles.
\nSome firms break this pattern by offering prestigious formal operational leadership programs that capture talent at the point of graduation. However, these seem largely restricted to foreign-owned multinationals and a small number of global UK firms. This appears consistent with the World Management Survey finding that US-owned firms in the United Kingdom tend to import stronger operational management practices.
\nWithin the United Kingdom, it seems that the London-centric industries are better at developing formal graduate schemes. Our informal review of UK leadership graduate schemes shows a ratio of approximately six to one in favour of London-centric industries over non-London-centric operational ones.
\nWhy conditions in Germany and the United States seem more favourable
\nThe structural forces described here appear less pronounced in Germany and the United States, the comparison economies we are using in this series. This is significantly influenced by differences in the geographic concentration of economic activity and by the relative pay and status of operational roles.
\nGerman and US economies are less concentrated
\nIn Germany, economic activity is less geographically concentrated than in the United Kingdom. Berlin, the largest city, contributes only 5 percent of the country\u2019s GDP, compared with London\u2019s 23 percent share. With a distributed industrial base and thousands of hidden champion small and medium-size enterprises distributed across the country, a highly educated couple in Germany have more options to build dual careers outside the capital city. Although Germany\u2019s industrial economy has challenges of its own today, the distribution of its talent is not one of them\u2014and the productivity gap between the United Kingdom and Germany remains significant.
\nThe United States has a genuine multi-hub economy. A dual-career household can build two high-value careers not only in the coastal centres of Boston, New York, San Francisco, and Washington, DC, but also in Atlanta, Austin, Chicago, Denver, Houston, Seattle, and beyond. Furthermore, the interior metros such as Columbus, Nashville, and St. Louis have all grown their shares of graduates and knowledge workers in recent years. Collectively, US cities offer ambitious graduates a range of viable options that the United Kingdom cannot match.
\nOperational roles have higher status and pay
\nOperational roles in Germany and the United States often command higher pay and social status. In healthcare, for example, a UK healthcare general manager earns approximately \u00a358,000, while a medical and health services manager in the United States earns a median $117,960 (\u00a387,000). Pay is an imperfect proxy for status, but the gap is consistent with operational roles retaining a standing abroad that they have lost in the United Kingdom.
\nProduction managers are paid a premium\u2014the extra pay versus national median that a role commands for its responsibility and the scarcity of the skills needed to do it. Germany pays a premium of 65 percent, the United States 100 percent, and according to official statistics, the United Kingdom pays 50 percent. But the UK figure is inflated: The ONS category includes highly paid engineer and managing directors and other senior leaders whom the American and German statistics classify elsewhere. When we adjust the UK premium using advertised salaries for production managers, it falls to between 15 and 25 percent\u2014less than half of Germany\u2019s premium and a quarter of the US premium.
\nThe entry requirements for a production manager position tell a similar story. The ONS says of the UK role that \u201cthere are no pre-set entry standards\u201d and that entry is possible \u201cwith either a degree or equivalent qualification, and/or relevant experience.\u201d The US Bureau of Labor Statistics lists a bachelor\u2019s degree and five or more years of related experience as the typical route in. Germany classifies the role at its highest skill level, which normally means a university degree, usually in engineering. Britain treats as an occupation what the United States and Germany treat as a profession.
\nIn Germany this premium may reflect the high cultural status of operations management and engineering. Germany\u2019s leading engineering universities\u2014TU Munich (TUM) and RWTH Aachen among them\u2014have institutionalised relationships with the country\u2019s most prestigious industrial employers. TUM\u2019s formal Partners of Excellence include BMW, Bosch, Siemens, and Volkswagen, and RWTH Aachen alumni consistently fill some of Germany\u2019s most senior DAX executive positions.
\nA similar dynamic exists in the United States, where Boeing and the GE companies rank among the top dozen employers of Georgia Tech mechanical engineers. Caterpillar and John Deere recruit there, too.
\nIn the United Kingdom, by contrast, perceptions of operational roles lag those of other sectors. A Nestl\u00e9 study of 2,000 young people found that just 4 percent would consider a career in manufacturing.
\nGraduate schemes develop and attract potential future top talent at the point of graduation
\nBoth countries also have more established graduate pathways into operational leadership and treat these pathways as a credible route to the very top. In Germany, the chief executives of all three major carmakers started on their companies\u2019 graduate programs. Oliver Blume joined Audi\u2019s trainee program in 1994 and has led Volkswagen Group since 2022. Ola K\u00e4llenius joined Daimler-Benz as a trainee in 1993 and has led Mercedes-Benz Group since 2019. Oliver Zipse joined BMW as a trainee in 1991 and led the company from 2019 until May 2026, when he handed over to Milan Nedeljkovi\u0107, who joined BMW as a trainee in 1993.
\nIn the United States, Procter & Gamble hires at entry level and builds from within. The company reports that more than 99 percent of its senior leaders began their careers there. David Taylor, who served as the chairman, president, and CEO from 2015 to 2021, joined in 1980 as a production manager. GE Aerospace\u2019s two-year leadership programs follow the same logic: more than 25 percent of its top leaders began their careers on one.
\nWhat company leaders and policymakers could do
\nLondon is a high-performing economic region. Its industries compete across the world stage and generate tax revenue that funds public services across the whole country. London and the South East (broadly, the London economic area) are the only two regions that pay more into the Treasury than they receive in public spending. The answer is not to restrict London\u2019s economic activity but to lift other areas up so they too can better compete on the world stage and return wealth to those regions.
\nReversing the structural forces described in this article would entail deliberate actions by company leaders and policymakers.
\nCompany leaders
\nRevalue operational leadership. Raise the pay, profile, and prestige of operational leadership roles to shift their perceived status. A premium comparable to the 65 percent seen in Germany or the 100 percent in the United States could begin to change the career calculus for high-potential graduates.
\nBuild prestigious graduate pathways. Foster more industrial graduate programs\u2014structured, high-status routes into operational leadership that are explicitly positioned as a pathway to senior corporate leadership.
\nFormalise the route from the shop floor to the top. Make technical apprenticeships and progression from the front line into site and operations leadership an explicit pathway, so potential future leaders can gain a wide variety of experience early in their career. This can maximise the ability of operational businesses to generate top talent from within.
\nTarget critical leadership mass within organisations. Focus on building the collective operational capability of leadership teams, not just on individual hires, to create the conditions and momentum for management systems to take root.
\nPolicymakers
\nStimulate high-value economic activity. Growing high-value jobs in regional cities through sustained investment in those cities and their transport could increase the effective density of roles to encourage the top-talent couples to pursue careers away from London. The mechanism described in this article would indicate that the lack of implementation of strong management systems is caused by regional businesses\u2019 restricted access to national talent, which is in turn caused by the regional talent trap. Therefore, improving UK productivity will depend on not just creating high-value jobs outside London but also ensuring that the talent needed to embed these systems is distributed more evenly across the economy.
\nTo get access to this talent, regions outside London need to offer top jobs and attract top talent to fill them. Just moving the job will not fix the problem. These leaders are created over decades, building knowledge, skills, and expertise by working in a wide variety of roles within a specific industry. Transforming regions into attractive talent hubs will mean reversing a talent trap trend that has developed over 50 years and has been reinforced by economic, psychological, and social factors.
\nThe implication is an uncomfortable one. London and the South East are remarkable economic assets, competing on the world stage and contributing hugely to the UK\u2019s prosperity. But if leaders with the drive and capability to transform British industry come from two main routes\u2014the shop floor and the graduate route\u2014then one region is absorbing a very large share of the second group. Every year, graduates who are the possible leaders of tomorrow head to the capital and its surroundings. Then, for social and economic reasons, many never leave.
\nRebalancing this even slightly would give more operationally intensive businesses of the Midlands, the North, South West, East, Wales, Scotland, and Northern Ireland the opportunity to reach the critical mass of leaders they need to install the management systems that would make them more productive, more globally competitive, and more profitable. If just some of the talent London attracts could be deployed to lead businesses elsewhere, the overall effect on the country could be significant\u2014to the postcodes in which these businesses operate and to the United Kingdom as a whole.
", "image": null, "banner_image": null, "date_published": "2026-10-21T16:00:00Z", "date_modified": "2026-10-21T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}, {"id": "https://www.mckinsey.com/insights/7bb506d8-c846-431b-b6bc-242ae07478f7", "url": "https://www.mckinsey.com/industries/packaging-and-paper/our-insights/the-state-of-grocery-retail-europe-implications-for-packaging-players", "external_url": null, "title": "The State of Grocery Retail Europe: Implications for packaging players", "content_text": "", "content_html": "Continued growth pressure on European grocery retailers limits opportunities for packaging players. In this environment, cost competitiveness remains imperative.", "summary": "Continued growth pressure on European grocery retailers limits opportunities for packaging players. In this environment, cost competitiveness remains imperative.", "image": null, "banner_image": null, "date_published": "2026-10-21T16:00:00Z", "date_modified": "2026-10-21T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}, {"id": "https://www.mckinsey.com/insights/d13db981-c0db-454a-a400-5534353c31b2", "url": "https://www.mckinsey.com/industries/consumer-packaged-goods/our-insights/focus-scale-fund-the-ai-playbook-cpg-leaders-are-missing", "external_url": null, "title": "Focus, scale, fund: The AI playbook CPG leaders are missing", "content_text": "", "content_html": "Small brands are out-innovating giants, buying patterns are changing, and scale isn\u2019t the shield it used to be. Here\u2019s how AI could help fund the next wave of growth in food and beverage.", "summary": "Small brands are out-innovating giants, buying patterns are changing, and scale isn\u2019t the shield it used to be. Here\u2019s how AI could help fund the next wave of growth in food and beverage.", "image": null, "banner_image": null, "date_published": "2026-10-21T16:00:00Z", "date_modified": "2026-10-21T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}, {"id": "https://www.mckinsey.com/insights/fed97583-880b-4949-99ea-9ffb92b7022d", "url": "https://www.mckinsey.com/industries/public-sector/our-insights/federal-permittings-uneven-economic-impact-and-what-companies-can-do-about-it", "external_url": null, "title": "Federal permitting\u2019s uneven economic impact, and what companies can do about it", "content_text": "", "content_html": "Permitting delays ripple unevenly across states depending on their industry mix. Companies with projects in the queue can take steps to manage delays.", "summary": "Permitting delays ripple unevenly across states depending on their industry mix. Companies with projects in the queue can take steps to manage delays.", "image": null, "banner_image": null, "date_published": "2026-10-21T16:00:00Z", "date_modified": "2026-10-21T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}, {"id": "https://www.mckinsey.com/insights/8cbcfe1c-535f-4e44-8084-2d8bae412097", "url": "https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/three-horizons-of-quantum-a-business-road-map-to-the-future-computing-economy", "external_url": null, "title": "Three horizons of quantum: A business road map to the future computing economy", "content_text": "", "content_html": "Many business leaders now accept that quantum technology is coming sooner than they first expected, and they\u2019re taking steps today to capture a competitive advantage. We have found that over 300 organizations worldwide are already working with quantum technology providers, with many more experimenting. What many leaders lack is a sense of how much to invest in which technologies, and over which timeline.
\nFaced with quantum technology\u2019s rapid pace of development and the scale of its implications, leaders are now forced to reconsider their action plans. Across all three of its pillars\u2014quantum computing, quantum communication, and quantum sensing\u2014quantum technology is advancing rapidly. Quantum technology companies are hitting milestones more quickly than they initially predicted, and many have ambitious road maps that put the expected arrival of fault-tolerant quantum computing (FTQC) within the next few years. FTQC will enable reliable quantum computing at scale, opening new possibilities in everything from drug discovery to materials design to financial modeling. But FTQC is a double-edged sword; when fully scaled, it will also bring about Q-Day, the moment when quantum computers can break today\u2019s public-key cryptography. As a result, even companies that don\u2019t directly deploy quantum computing will need to update their security platforms to contend with it.
\nTo help business leaders understand the road ahead, we mapped three horizons of quantum technology development (Exhibit 1). This article will also highlight the roadblocks that could slow the quantum timeline\u2014and the emerging developments that could speed it up.
\nThe financial stakes justify immediate action. Our latest Quantum Technology Monitor finds that quantum computing alone could create $1.3 trillion to $2.7 trillion of economic value worldwide by 2035. To capture as much of this value as possible, business leaders need to understand how quantum technology is developing across the three horizons and map their companies\u2019 pathways through them. Waiting to invest is risky. Those that do could soon find that the partnerships, talent, and intellectual property needed to succeed with quantum technology were claimed years earlier by faster-moving competitors. They could also face severe security issues when Q-Day arrives.
\nHorizon 1: Early commercialization
\nThe clearest signal that quantum technology has crossed from research into commercial deployment is that the sector now has an exit market. In 2026, companies such as Horizon Quantum, Infleqtion, Pasqal, Quantinuum, and Xanadu went public. Meanwhile, IonQ acquired SkyWater Technology in a deal that pairs SkyWater\u2019s chip manufacturing with IonQ\u2019s quantum hardware and software. Governments are also supplying a stream of development capital, including, for example, recent commitments of $2 billion in the United States and more than \u20ac1 billion in France. Many of these investments are aimed at increasing manufacturing capacity rather than just research output, and they show that governments see economic opportunity in quantum technology.
\nThe demand side has changed just as much. Proof-of-concept projects are shifting to paid codevelopment. Our research finds that companies including Airbus, Boehringer Ingelheim, E.ON, JPMorgan Chase, and Liberty Mutual are actively developing applications in collaboration with quantum technology players. Cloud access and quantum-as-a-service models are also lowering the barrier to entry, turning what was once a large capital project into a strategic partnership decision. Companies are showing a strong interest in quantum. Our recent Technology Trends Outlook 2026 report shows that searches on quantum-related keywords and the number of research publications on quantum technology rose steadily between 2022 and 2025, while equity investment surged from $2.5 billion in 2024 to $13.3 billion in 2025.
\nNone of this means that quantum computers are ready today, but the market around them is clearly forming. Supplier relationships and intellectual property are being cemented, internal skills are being built, and companies are making investment decisions even with imperfect information about when reliable fault-tolerant hardware will arrive.
\nThe three quantum pillars are commercializing at different speeds. Quantum computing is currently the largest market of the three pillars. It draws its advantage from the quantum principles of superposition and entanglement, which allow a processor to evaluate an enormous number of possible states simultaneously. The unit doing that work is the qubit, the quantum equivalent of a classical bit. But while a classical bit must be either zero or one, a qubit can hold both values at once and behave as a single system.
\nOur research shows that quantum computing companies surpassed $1 billion in combined revenue in 2025\u2014and that revenue could grow to as much as $4.4 billion by 2028. Near-term deployment of quantum computing is overwhelmingly hybrid, with quantum processors attached to existing high-performance computing environments rather than replacing them. Several such systems are already operating. The US Department of Energy\u2019s Oak Ridge National Laboratory has connected IQM\u2019s Pathfinder system with its Frontier supercomputer; Europe\u2019s 54-qubit Euro-Q-Exa went live at the Leibniz Supercomputing Centre in early 2026; and Aramco and Pasqal operate a 200-qubit neutral-atom machine as the Middle East\u2019s first commercial quantum-computing-as-a-service platform.
\nQuantum communication is driven by the looming Q-Day deadline. To address upcoming security risks at the physical layer, quantum key distribution (QKD) uses quantum mechanics to make interception detectable in real time on protected links, and it\u2019s starting to move out of the laboratory. Meanwhile, postquantum cryptography (PQC)\u2014new algorithms designed to resist quantum attacks while running on existing hardware\u2014is here now and will be critical for companies migrating their cryptography systems. We expect the overall quantum communication market to reach $11 billion to $15 billion worldwide by 2035.
\nQuantum sensing is the most mature of the three pillars. It uses the same quantum properties that make qubits work but applies them to measurement rather than computation. Q-CTRL\u2019s quantum sensors, which navigate by reading Earth\u2019s magnetic field rather than satellite signals, have completed safety-of-flight testing, for example. Atomic clocks, gravimeters, and magnetometers that leverage quantum sensing are also finding commercial traction in sectors such as navigation, medical diagnostics, and mineral exploration. We expect this market to reach $7 billion to $10 billion worldwide by 2035.
\nFor business leaders, positioning their companies as front-runners in Horizon 1 is about building the foundations of success. This requires securing talent, intellectual property, and partnerships today. Those that wait risk entering the next phase dependent on technologies and ecosystems shaped by their faster-moving competitors.
\nHorizon 2: Establishing a competitive position
\nHardware road maps from leading quantum computing companies are converging on the possibility of delivering a fault-tolerant quantum computer within five years. IBM has published a potential path to large-scale fault tolerance by 2029, while IonQ is targeting two million physical qubits by 2030\u2014enough to yield up to 80,000 logical qubits, the error-corrected units of computation that fault tolerance depends on.
\nFault tolerance will bring a step change for quantum computing: machines that can sustain long computations reliably enough to be embedded into production workflows. If quantum technology companies achieve their development road maps, we expect the quantum computing market alone to capture between $43 billion and $71 billion of the worldwide \u201cinternal quantum market\u201d by 2035. (The internal quantum market represents the total global spending expected by companies on quantum hardware, software, and services, and we estimate this market will reach $60 billion to $100 billion in 2035.)
\nBut fault tolerance at scale is also a trigger for Q-Day, which could invoke a whole slew of new security challenges. For many large enterprises, getting ready for Q-Day is a multiyear, multibillion-dollar transformation\u2014but one that\u2019s necessary. Without PQC systems in place, companies\u2019 virtual private networks, public-key infrastructure, connected devices, and identity systems could all become exposed on Q-Day. Plus, sensitive data is already exposed to \u201charvest now, decrypt later\u201d attacks, in which encrypted traffic is collected by hackers today so it can be read by quantum computers on Q-Day. That means acting today is imperative to avoid severe security flaws in the near future.
\nHowever, the FTQC timeline faces several roadblocks. The case for fault tolerance by the early 2030s rests on road maps, and every technology developer\u2014and buyer\u2014knows that road maps often slip. Three key constraints deserve attention from leaders who are building their quantum investment plans around these dates.
\nThe first is that scaling is now largely a manufacturing problem. Quantum technology is advancing quickly, but building quantum machines with millions of physical qubits requires lasers, cryogenic infrastructure, control electronics, and manufacturing processes that don\u2019t yet exist in the capacity needed for true scaling.
\nThe second is that talent is scarce. The pool of researchers and engineers who can design and program quantum computers remains limited. To scale quantum computing, companies also need software developers and solution architects to translate quantum capabilities into business applications, as well as technicians to build manufacturing systems.
\nThe third is that there is no consensus on which quantum computing hardware modality\u2014superconducting, trapped ion, neutral atom, photonic, spin, topological, or an entirely new one\u2014will scale most effectively. A company choosing a single quantum technology provider is making a future-looking bet, not merely a procurement decision.
\nDespite these uncertainties, many companies are pushing ahead with plans for quantum technology adoption; they can\u2019t afford to be laggards. When we interviewed business leaders already working with quantum technology providers, one-third said their companies spent $10 million or more on quantum computing in 2025, with 7 percent of those companies spending $50 million or more (Exhibit 2). These are capability-building budgets rather than experimental ones, set by buyers that have determined that today\u2019s road maps, while not set in stone, are reliable enough to plan against.
\nHorizon 2 is when companies lock in their quantum strategies\u2014and real value starts to accrue. Fault-tolerant machines will be scarce and expensive when they first arrive, so the customers that worked with providers on earlier pilots and partnerships in Horizon 1 could get first dibs on this compute capacity. Thus, during Horizon 2, the most competitive companies will be those that continue to codevelop the quantum applications they started before the hardware even existed.
\nHorizon 3: Leading the quantum-enabled economy
\nIn Horizon 3, entire economic arenas could reorganize around the future capabilities that quantum technology will make possible. The McKinsey Global Institute (MGI) has identified 18 future arenas of competition that will shape the global economy over the next 15 years, collectively generating $29 trillion to $48 trillion in revenues by 2040. Quantum technology intersects with several of these arenas. This means that some of the most competitive companies in Horizon 3 will be those that redesign their operating models around quantum technology.
\nQuantum technology could create outsize value in four of the arenas:
\nCybersecurity. Quantum technology could soon upend the public-key cryptography that underpins most enterprise security systems. Business leaders need to act now to build the PQC systems that will ensure their data remains safe after Q-Day. This is a multiyear program, not a patch cycle. We expect PQC to emerge as one of the largest quantum communication verticals, representing a $2.4 billion to $3.4 billion global market by 2035.
\nBut quantum technology could also help companies design new, more robust security systems\u2014which is especially crucial in an era when AI has dramatically sped up attacks. QKD is a hardware-based security approach that uses physics to distribute keys and offers a level of protection that classical cryptography cannot. QKD is distance limited and cannot provide authentication or signatures on its own, so it remains a complement to PQC rather than a substitute in the near term. But we predict that by 2040, QKD could be a standard networking option. Financial exchanges, power grids, hospital systems, and government networks could at that point run over shared, QKD-protected corridors as the default. By 2035, we expect the QKD market to be worth $2 billion to $2.8 billion worldwide.
\nDrugs for obesity and related conditions. Quantum computing could lead to a new world of personalized medicine, especially for chronic conditions. The body\u2019s biological systems are far too complex for classical computers to model, so medical research is done on simplified subsets of data. In the near term, quantum computing could sharpen molecular simulation enough to design therapies for narrow patient subgroups rather than broad populations. For example, it could enable pharmaceutical companies to develop more effective therapeutics for obesity and its related chronic diseases. By 2035, we predict that quantum computing could deliver up to $400 billion in value to the pharmaceutical industry.
\nFurther out, quantum computing could support patient-level digital twins: computational replicas of an individual, detailed enough to simulate how a disease could progress and how a treatment would work. By 2040, diagnosis and treatment could be optimized continuously for one person at a time rather than for a population they resemble.
\nIndustrial and consumer biotechnology. Quantum technology could transform the discovery of biomaterials and biochemicals. We estimate that quantum computing could unlock up to $800 billion in value in the chemicals industry alone by 2035. A material\u2019s useful properties are a result of how its electrons interact. Those interactions are quantum mechanical, and the possible configurations multiply so fast that classical computers cannot always calculate them directly. Thus, researchers often fall back on trial and error: synthesize a candidate, test it, adjust, repeat.
\nQuantum processors don\u2019t rely on approximation, because they represent electron behavior using the same physics that governs it. In May 2026, Q-CTRL demonstrated a 3,000-fold speedup in modeling how materials behave, running on 120 qubits. That capability points toward designing materials that don\u2019t yet exist, such as superconductors that move electricity without loss and alloys that survive inside next-generation reactors.
\nRobotics. Our analysis shows that the global market for general-purpose robots could reach $370 billion by 2040. Quantum computing could have a big impact on the development of general-purpose robots, including humanoid, quadruped, and other form factors, because to operate effectively in the real world, these robots must continuously process information from cameras, microphones, sensors, and other inputs. They must assess their surroundings to move safely through unpredictable environments and coordinate with cloud infrastructure, enterprise software, and fleets of other robots. In other words, general-purpose robots require huge amounts of sensing, communication, and computational processing power\u2014all of which quantum technology could provide.
\nGeneral-purpose robots that harness quantum computing from near-bot and cloud-computing platforms will come first, with quantum networking and sensing later becoming on-bot quantum enhancements. We estimate that qubots, or robots that use quantum technology in at least some capacity, could represent up to a $67 billion global industry by 2040, with humanoid qubots alone representing up to $55 billion of that total.
\nBy 2040, quantum computing could become ordinary in many high-innovation sectors. Most organizations will likely access quantum computing through quantum-as-a-service providers. Quantum communication could feasibly scale from a handful of connections to a network layer. Quantum sensing could become embedded in common industry tools: magnetometers that read brain activity, gravimeters that map what lies underground for mineral exploration, and navigation systems that track positions without satellite signals.
\nHorizon 3 is not just about adopting quantum technology for its own sake. If quantum becomes as foundational as AI is today, the question facing business leaders will not be whether to use quantum technology at all, but how to do so in ways that materially change their operating models. Every leader also needs to consider how quantum technology and AI will intersect\u2014and the value that intersection could create for their businesses, the economy, and society.
\nMGI\u2019s latest research on the AI economy describes an interconnected system moving at mismatched speeds: Model capabilities are expanding exponentially, while the physical inputs they run on, such as chips, power, and data centers, are growing linearly and take years to build. Quantum computing could help unblock that bottleneck, dramatically expanding compute capacity to accelerate AI capabilities. And the AI\u2013quantum relationship is symbiotic. AI is speeding up quantum algorithm design and error correction in ways that could hasten the arrival of FTQC. As MGI notes, changes in even one part of the AI ecosystem can ripple through the other parts to cause change to the whole system. The combination of AI and quantum computing is an example of such an interconnected shift, and together they could accelerate the arrival of Horizon 3. This would leave leaders less time to prepare for its implications.
\nA way forward
\nOver the next 15 years, quantum computing will drive deep structural shifts in how companies do business. As fault-tolerant systems scale, first movers could gain access to computational capabilities that deliver hundreds of billions of dollars in value. Preparing for this future requires action today. To do so, leaders need to understand how quantum will develop across the three horizons.
\nTo ensure their organizations take leading positions during each horizon, leaders can focus on four strategic priorities: secure ecosystem partnerships and access to frontier systems; pilot applications ahead of commercial readiness; build organizational familiarity through talent and IP; and explore how quantum technology could ultimately alleviate compute and energy bottlenecks in AI scaling. Quantum technology will not wait for organizations to feel ready\u2014and the window to build a leading position will narrow with each horizon that passes.
", "summary": "Many business leaders now accept that quantum technology is coming sooner than they first expected, and they\u2019re taking steps today to capture a competitive advantage. We have found that over 300 organizations worldwide are already working with quantum technology providers, with many more experimenting. What many leaders lack is a sense of how much to invest in which technologies, and over which timeline.
\nFaced with quantum technology\u2019s rapid pace of development and the scale of its implications, leaders are now forced to reconsider their action plans. Across all three of its pillars\u2014quantum computing, quantum communication, and quantum sensing\u2014quantum technology is advancing rapidly. Quantum technology companies are hitting milestones more quickly than they initially predicted, and many have ambitious road maps that put the expected arrival of fault-tolerant quantum computing (FTQC) within the next few years. FTQC will enable reliable quantum computing at scale, opening new possibilities in everything from drug discovery to materials design to financial modeling. But FTQC is a double-edged sword; when fully scaled, it will also bring about Q-Day, the moment when quantum computers can break today\u2019s public-key cryptography. As a result, even companies that don\u2019t directly deploy quantum computing will need to update their security platforms to contend with it.
\nTo help business leaders understand the road ahead, we mapped three horizons of quantum technology development (Exhibit 1). This article will also highlight the roadblocks that could slow the quantum timeline\u2014and the emerging developments that could speed it up.
\nThe financial stakes justify immediate action. Our latest Quantum Technology Monitor finds that quantum computing alone could create $1.3 trillion to $2.7 trillion of economic value worldwide by 2035. To capture as much of this value as possible, business leaders need to understand how quantum technology is developing across the three horizons and map their companies\u2019 pathways through them. Waiting to invest is risky. Those that do could soon find that the partnerships, talent, and intellectual property needed to succeed with quantum technology were claimed years earlier by faster-moving competitors. They could also face severe security issues when Q-Day arrives.
\nHorizon 1: Early commercialization
\nThe clearest signal that quantum technology has crossed from research into commercial deployment is that the sector now has an exit market. In 2026, companies such as Horizon Quantum, Infleqtion, Pasqal, Quantinuum, and Xanadu went public. Meanwhile, IonQ acquired SkyWater Technology in a deal that pairs SkyWater\u2019s chip manufacturing with IonQ\u2019s quantum hardware and software. Governments are also supplying a stream of development capital, including, for example, recent commitments of $2 billion in the United States and more than \u20ac1 billion in France. Many of these investments are aimed at increasing manufacturing capacity rather than just research output, and they show that governments see economic opportunity in quantum technology.
\nThe demand side has changed just as much. Proof-of-concept projects are shifting to paid codevelopment. Our research finds that companies including Airbus, Boehringer Ingelheim, E.ON, JPMorgan Chase, and Liberty Mutual are actively developing applications in collaboration with quantum technology players. Cloud access and quantum-as-a-service models are also lowering the barrier to entry, turning what was once a large capital project into a strategic partnership decision. Companies are showing a strong interest in quantum. Our recent Technology Trends Outlook 2026 report shows that searches on quantum-related keywords and the number of research publications on quantum technology rose steadily between 2022 and 2025, while equity investment surged from $2.5 billion in 2024 to $13.3 billion in 2025.
\nNone of this means that quantum computers are ready today, but the market around them is clearly forming. Supplier relationships and intellectual property are being cemented, internal skills are being built, and companies are making investment decisions even with imperfect information about when reliable fault-tolerant hardware will arrive.
\nThe three quantum pillars are commercializing at different speeds. Quantum computing is currently the largest market of the three pillars. It draws its advantage from the quantum principles of superposition and entanglement, which allow a processor to evaluate an enormous number of possible states simultaneously. The unit doing that work is the qubit, the quantum equivalent of a classical bit. But while a classical bit must be either zero or one, a qubit can hold both values at once and behave as a single system.
\nOur research shows that quantum computing companies surpassed $1 billion in combined revenue in 2025\u2014and that revenue could grow to as much as $4.4 billion by 2028. Near-term deployment of quantum computing is overwhelmingly hybrid, with quantum processors attached to existing high-performance computing environments rather than replacing them. Several such systems are already operating. The US Department of Energy\u2019s Oak Ridge National Laboratory has connected IQM\u2019s Pathfinder system with its Frontier supercomputer; Europe\u2019s 54-qubit Euro-Q-Exa went live at the Leibniz Supercomputing Centre in early 2026; and Aramco and Pasqal operate a 200-qubit neutral-atom machine as the Middle East\u2019s first commercial quantum-computing-as-a-service platform.
\nQuantum communication is driven by the looming Q-Day deadline. To address upcoming security risks at the physical layer, quantum key distribution (QKD) uses quantum mechanics to make interception detectable in real time on protected links, and it\u2019s starting to move out of the laboratory. Meanwhile, postquantum cryptography (PQC)\u2014new algorithms designed to resist quantum attacks while running on existing hardware\u2014is here now and will be critical for companies migrating their cryptography systems. We expect the overall quantum communication market to reach $11 billion to $15 billion worldwide by 2035.
\nQuantum sensing is the most mature of the three pillars. It uses the same quantum properties that make qubits work but applies them to measurement rather than computation. Q-CTRL\u2019s quantum sensors, which navigate by reading Earth\u2019s magnetic field rather than satellite signals, have completed safety-of-flight testing, for example. Atomic clocks, gravimeters, and magnetometers that leverage quantum sensing are also finding commercial traction in sectors such as navigation, medical diagnostics, and mineral exploration. We expect this market to reach $7 billion to $10 billion worldwide by 2035.
\nFor business leaders, positioning their companies as front-runners in Horizon 1 is about building the foundations of success. This requires securing talent, intellectual property, and partnerships today. Those that wait risk entering the next phase dependent on technologies and ecosystems shaped by their faster-moving competitors.
\nHorizon 2: Establishing a competitive position
\nHardware road maps from leading quantum computing companies are converging on the possibility of delivering a fault-tolerant quantum computer within five years. IBM has published a potential path to large-scale fault tolerance by 2029, while IonQ is targeting two million physical qubits by 2030\u2014enough to yield up to 80,000 logical qubits, the error-corrected units of computation that fault tolerance depends on.
\nFault tolerance will bring a step change for quantum computing: machines that can sustain long computations reliably enough to be embedded into production workflows. If quantum technology companies achieve their development road maps, we expect the quantum computing market alone to capture between $43 billion and $71 billion of the worldwide \u201cinternal quantum market\u201d by 2035. (The internal quantum market represents the total global spending expected by companies on quantum hardware, software, and services, and we estimate this market will reach $60 billion to $100 billion in 2035.)
\nBut fault tolerance at scale is also a trigger for Q-Day, which could invoke a whole slew of new security challenges. For many large enterprises, getting ready for Q-Day is a multiyear, multibillion-dollar transformation\u2014but one that\u2019s necessary. Without PQC systems in place, companies\u2019 virtual private networks, public-key infrastructure, connected devices, and identity systems could all become exposed on Q-Day. Plus, sensitive data is already exposed to \u201charvest now, decrypt later\u201d attacks, in which encrypted traffic is collected by hackers today so it can be read by quantum computers on Q-Day. That means acting today is imperative to avoid severe security flaws in the near future.
\nHowever, the FTQC timeline faces several roadblocks. The case for fault tolerance by the early 2030s rests on road maps, and every technology developer\u2014and buyer\u2014knows that road maps often slip. Three key constraints deserve attention from leaders who are building their quantum investment plans around these dates.
\nThe first is that scaling is now largely a manufacturing problem. Quantum technology is advancing quickly, but building quantum machines with millions of physical qubits requires lasers, cryogenic infrastructure, control electronics, and manufacturing processes that don\u2019t yet exist in the capacity needed for true scaling.
\nThe second is that talent is scarce. The pool of researchers and engineers who can design and program quantum computers remains limited. To scale quantum computing, companies also need software developers and solution architects to translate quantum capabilities into business applications, as well as technicians to build manufacturing systems.
\nThe third is that there is no consensus on which quantum computing hardware modality\u2014superconducting, trapped ion, neutral atom, photonic, spin, topological, or an entirely new one\u2014will scale most effectively. A company choosing a single quantum technology provider is making a future-looking bet, not merely a procurement decision.
\nDespite these uncertainties, many companies are pushing ahead with plans for quantum technology adoption; they can\u2019t afford to be laggards. When we interviewed business leaders already working with quantum technology providers, one-third said their companies spent $10 million or more on quantum computing in 2025, with 7 percent of those companies spending $50 million or more (Exhibit 2). These are capability-building budgets rather than experimental ones, set by buyers that have determined that today\u2019s road maps, while not set in stone, are reliable enough to plan against.
\nHorizon 2 is when companies lock in their quantum strategies\u2014and real value starts to accrue. Fault-tolerant machines will be scarce and expensive when they first arrive, so the customers that worked with providers on earlier pilots and partnerships in Horizon 1 could get first dibs on this compute capacity. Thus, during Horizon 2, the most competitive companies will be those that continue to codevelop the quantum applications they started before the hardware even existed.
\nHorizon 3: Leading the quantum-enabled economy
\nIn Horizon 3, entire economic arenas could reorganize around the future capabilities that quantum technology will make possible. The McKinsey Global Institute (MGI) has identified 18 future arenas of competition that will shape the global economy over the next 15 years, collectively generating $29 trillion to $48 trillion in revenues by 2040. Quantum technology intersects with several of these arenas. This means that some of the most competitive companies in Horizon 3 will be those that redesign their operating models around quantum technology.
\nQuantum technology could create outsize value in four of the arenas:
\nCybersecurity. Quantum technology could soon upend the public-key cryptography that underpins most enterprise security systems. Business leaders need to act now to build the PQC systems that will ensure their data remains safe after Q-Day. This is a multiyear program, not a patch cycle. We expect PQC to emerge as one of the largest quantum communication verticals, representing a $2.4 billion to $3.4 billion global market by 2035.
\nBut quantum technology could also help companies design new, more robust security systems\u2014which is especially crucial in an era when AI has dramatically sped up attacks. QKD is a hardware-based security approach that uses physics to distribute keys and offers a level of protection that classical cryptography cannot. QKD is distance limited and cannot provide authentication or signatures on its own, so it remains a complement to PQC rather than a substitute in the near term. But we predict that by 2040, QKD could be a standard networking option. Financial exchanges, power grids, hospital systems, and government networks could at that point run over shared, QKD-protected corridors as the default. By 2035, we expect the QKD market to be worth $2 billion to $2.8 billion worldwide.
\nDrugs for obesity and related conditions. Quantum computing could lead to a new world of personalized medicine, especially for chronic conditions. The body\u2019s biological systems are far too complex for classical computers to model, so medical research is done on simplified subsets of data. In the near term, quantum computing could sharpen molecular simulation enough to design therapies for narrow patient subgroups rather than broad populations. For example, it could enable pharmaceutical companies to develop more effective therapeutics for obesity and its related chronic diseases. By 2035, we predict that quantum computing could deliver up to $400 billion in value to the pharmaceutical industry.
\nFurther out, quantum computing could support patient-level digital twins: computational replicas of an individual, detailed enough to simulate how a disease could progress and how a treatment would work. By 2040, diagnosis and treatment could be optimized continuously for one person at a time rather than for a population they resemble.
\nIndustrial and consumer biotechnology. Quantum technology could transform the discovery of biomaterials and biochemicals. We estimate that quantum computing could unlock up to $800 billion in value in the chemicals industry alone by 2035. A material\u2019s useful properties are a result of how its electrons interact. Those interactions are quantum mechanical, and the possible configurations multiply so fast that classical computers cannot always calculate them directly. Thus, researchers often fall back on trial and error: synthesize a candidate, test it, adjust, repeat.
\nQuantum processors don\u2019t rely on approximation, because they represent electron behavior using the same physics that governs it. In May 2026, Q-CTRL demonstrated a 3,000-fold speedup in modeling how materials behave, running on 120 qubits. That capability points toward designing materials that don\u2019t yet exist, such as superconductors that move electricity without loss and alloys that survive inside next-generation reactors.
\nRobotics. Our analysis shows that the global market for general-purpose robots could reach $370 billion by 2040. Quantum computing could have a big impact on the development of general-purpose robots, including humanoid, quadruped, and other form factors, because to operate effectively in the real world, these robots must continuously process information from cameras, microphones, sensors, and other inputs. They must assess their surroundings to move safely through unpredictable environments and coordinate with cloud infrastructure, enterprise software, and fleets of other robots. In other words, general-purpose robots require huge amounts of sensing, communication, and computational processing power\u2014all of which quantum technology could provide.
\nGeneral-purpose robots that harness quantum computing from near-bot and cloud-computing platforms will come first, with quantum networking and sensing later becoming on-bot quantum enhancements. We estimate that qubots, or robots that use quantum technology in at least some capacity, could represent up to a $67 billion global industry by 2040, with humanoid qubots alone representing up to $55 billion of that total.
\nBy 2040, quantum computing could become ordinary in many high-innovation sectors. Most organizations will likely access quantum computing through quantum-as-a-service providers. Quantum communication could feasibly scale from a handful of connections to a network layer. Quantum sensing could become embedded in common industry tools: magnetometers that read brain activity, gravimeters that map what lies underground for mineral exploration, and navigation systems that track positions without satellite signals.
\nHorizon 3 is not just about adopting quantum technology for its own sake. If quantum becomes as foundational as AI is today, the question facing business leaders will not be whether to use quantum technology at all, but how to do so in ways that materially change their operating models. Every leader also needs to consider how quantum technology and AI will intersect\u2014and the value that intersection could create for their businesses, the economy, and society.
\nMGI\u2019s latest research on the AI economy describes an interconnected system moving at mismatched speeds: Model capabilities are expanding exponentially, while the physical inputs they run on, such as chips, power, and data centers, are growing linearly and take years to build. Quantum computing could help unblock that bottleneck, dramatically expanding compute capacity to accelerate AI capabilities. And the AI\u2013quantum relationship is symbiotic. AI is speeding up quantum algorithm design and error correction in ways that could hasten the arrival of FTQC. As MGI notes, changes in even one part of the AI ecosystem can ripple through the other parts to cause change to the whole system. The combination of AI and quantum computing is an example of such an interconnected shift, and together they could accelerate the arrival of Horizon 3. This would leave leaders less time to prepare for its implications.
\nA way forward
\nOver the next 15 years, quantum computing will drive deep structural shifts in how companies do business. As fault-tolerant systems scale, first movers could gain access to computational capabilities that deliver hundreds of billions of dollars in value. Preparing for this future requires action today. To do so, leaders need to understand how quantum will develop across the three horizons.
\nTo ensure their organizations take leading positions during each horizon, leaders can focus on four strategic priorities: secure ecosystem partnerships and access to frontier systems; pilot applications ahead of commercial readiness; build organizational familiarity through talent and IP; and explore how quantum technology could ultimately alleviate compute and energy bottlenecks in AI scaling. Quantum technology will not wait for organizations to feel ready\u2014and the window to build a leading position will narrow with each horizon that passes.
", "image": null, "banner_image": null, "date_published": "2026-10-19T16:00:00Z", "date_modified": "2026-10-19T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}, {"id": "https://www.mckinsey.com/insights/979e9880-467f-4021-80d4-4c3513f6d8cc", "url": "https://www.mckinsey.com/industries/automotive-and-assembly/our-insights/switch-tracks-rewiring-rolling-stock-for-a-new-era", "external_url": null, "title": "Switch tracks: Rewiring rolling stock for a new era", "content_text": "", "content_html": "Fast-growing markets in Asia are difficult for global rail OEMs to access. Value could shift from new builds toward servicing and renewals\u2014while OEMs prepare for future opportunities.", "summary": "Fast-growing markets in Asia are difficult for global rail OEMs to access. Value could shift from new builds toward servicing and renewals\u2014while OEMs prepare for future opportunities.", "image": null, "banner_image": null, "date_published": "2026-10-19T16:00:00Z", "date_modified": "2026-10-19T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}, {"id": "https://www.mckinsey.com/insights/4817c6ed-07b3-45d0-a446-64c040e1a7b7", "url": "https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/securing-software-in-the-age-of-ai-red-hat-cto-chris-wright", "external_url": null, "title": "Securing software in the age of AI: Red Hat CTO Chris Wright", "content_text": "", "content_html": "Over his roughly three decades in the open-source software industry, including the last two at Red Hat, Chris Wright has watched the sector evolve from a once niche technical field to a mainstream building block of enterprise infrastructure. As Red Hat\u2019s chief technology officer, Wright now oversees engineering across the company, from emerging-technology research to product development and ongoing support of products already in the market. In a recent conversation with McKinsey Senior Partner Martin Harrysson and Associate Partner Shilpa Khatana, Wright discussed how a recent wave of AI-driven discoveries of code vulnerabilities is reshaping trust in the software industry and why he believes the industry will emerge from this moment more secure.
\nThis interview has been edited for length and clarity.
\nMartin Harrysson: You\u2019ve been at Red Hat for more than two decades, and in the open-source community, including early years involved with Linux, for around three. What strikes you about where open source finds itself today?
\nChris Wright: It\u2019s a fascinating time for open source. Over 30 years, it went from a fringe, hobbyist activity to something pervasive across every industry\u2014consumer, enterprise, small business, data centers, devices. Open source is foundational to the global economy today. The vast majority of enterprise codebases\u2014well over 80 percent, typically over 90 percent\u2014contain open-source components. Every hyperscaler and every large consumer platform runs on open-source infrastructure. It\u2019s entrenched, and it\u2019s fundamental to our future.
\nNow it\u2019s back at the center of the conversation, this time because of cybersecurity concerns and AI\u2019s role in scanning code and even building models with open weights. There\u2019s a full-circle aspect to it: Many of the same debates we had decades ago about trusting open source are resurfacing in a new technology era.
\nOne of open source\u2019s greatest strengths is also what\u2019s driving today\u2019s concerns: It\u2019s open. Anyone can access the code, learn from it, and build with it; that\u2019s real democratization of technology. Now that AI models have been trained on the internet, including a huge amount of open-source code, they\u2019ve become remarkably capable at understanding, writing, documenting, and finding issues in code.
\nMartin Harrysson: Given those real concerns, how do you see AI reshaping open source\u2019s role and importance?
\nChris Wright: Starting around the middle of this year, a new class of cybersecurity-focused models\u2014beginning with Mythos and the Glasswing project, and followed by efforts like Daybreak\u2014began uncovering vulnerabilities that had been sitting latent in source code for years, in some cases upward of two decades. Some are individually exploitable; others are chains of smaller vulnerabilities that together give attackers a path to escalate privileges or break into systems. Because open-source code is visible and accessible, it has been the natural target for this kind of scanning.
\nThat\u2019s produced a narrative that open source is somehow uniquely insecure. I think that\u2019s the wrong conclusion. Open source has always made a virtue of transparency\u2014sunlight is a disinfectant. You can see exactly how a cryptographic algorithm is implemented rather than being told to just trust a vendor\u2019s word for it. And the community development model brings a much broader set of eyes to any given piece of code than a small proprietary team ever could. That means faster discovery and faster remediation\u2014there\u2019s no hiding behind security through obscurity.
\nMartin Harrysson: So you see this having a positive overall impact?
\nChris Wright: What\u2019s happening now is that AI has made it dramatically cheaper to scan all of that code at once, so we\u2019re seeing a concentrated burst of vulnerability discovery. It\u2019s a genuinely uncomfortable moment: Customers are worried, and rightly so, because open source is so pervasive that they often lack full visibility into what\u2019s actually deployed, especially when it arrives bundled inside third-party or commercial software. But I\u2019m optimistic about where this ends. We\u2019ll come out the other side having raised the bar for security across the board\u2014more scanning, faster patching, and security testing built directly into project pipelines. Efforts like Daybreak, which has talked about \u201cpatching the planet,\u201d are genuinely global in scope in a way that would be much harder to achieve if we were only securing smaller, more fragmented, proprietary codebases. And this is exactly what Red Hat and IBM are doing with our Lightwell initiative: making open source safer for enterprises.
\nMartin Harrysson: In the near term, what should a CIO, CTO, or CEO be doing about this?
\nChris Wright: I\u2019d point to three priorities\u2014two are basic hygiene, and the third is more about where this is all heading.
\nFirst: Know what software you have and where it\u2019s running. If you can\u2019t produce something like a software bill of materials or a dependency graph across your enterprise, you have no real way to assess your risk. That visibility also helps you prioritize\u2014not every system carries the same value or the same exposure. A database full of sensitive information behind several layers of firewalls is a very different risk profile than a public-facing service.
\nSecond: Build the automation to actually act on what you find. Imagine every vulnerability in your environment has already been discovered. Now you need to patch potentially hundreds of thousands of dependencies across your infrastructure. If your patching process still involves months of manual review and a spreadsheet, you will never catch up\u2014the backlog will always outrun your capacity. Mature CI/CD and DevSecOps practices, which the industry has been building toward for years but hasn\u2019t fully adopted everywhere, are what make addressing that kind of scale possible.
\nThird, and this is more forward-looking: The industry needs to keep building the tools such as scanning, exploit research, and mitigation techniques that let defenders keep pace with attackers. Right now, I think attackers have the upper hand. We\u2019re going to discover entirely new classes of vulnerabilities, not just new instances of existing ones, and each new class will trigger a fresh wave of activity. I think we\u2019re heading toward a world where both attack and defense increasingly happen at machine speed\u2014agents defending against agents, rather than relying on human-speed response.
\nShilpa Khatana: Who should own this? In most enterprises, we see engineering taking on more responsibility for security, but there\u2019s also a dedicated cybersecurity function, risk, governance, and now often an AI council. How should those groups divide the work?
\nChris Wright: It\u2019s a shared responsibility, but with real specialization by role.
\nThis builds on the shift we made with DevOps and then DevSecOps\u2014bringing security earlier into the development process, on the theory that a vulnerability caught early is far cheaper to fix than one caught in production. But developers aren\u2019t security specialists, so DevSecOps also means building automation and gates into the pipeline that give developers visibility into security issues without requiring them to become experts themselves.
\nThe application development team owns what\u2019s deployed and where\u2014they\u2019re best positioned to know what dependencies went into their software. One thing I don\u2019t think is well understood: When you write code, you take responsibility for it, but the moment you pull in a dependency from the internet, you\u2019ve taken on responsibility for that code as well, even a small, stable utility you\u2019ve never had reason to question.
\nAt the same time, other teams need to specialize in things individual developers shouldn\u2019t be expected to do alone: \u201cred-teaming\u201d at a systems level, setting compliance requirements as a CISO function, and translating policy into automated checks embedded in the pipeline. So you end up with a CISO setting requirements, teams turning that into policy as code, and development teams having responsibility for what they build and what they pull in. It\u2019s shared, but each group brings a different kind of expertise.
\nMartin Harrysson: The rise of agentic AI adds another layer to this\u2014agents are now writing code and can themselves introduce dependencies, sometimes through subagents. Doesn\u2019t that complicate accountability?
\nChris Wright: You\u2019re still ultimately accountable if you initiated the work\u2014that doesn\u2019t change. But I agree these layers of abstraction can obscure what\u2019s actually happening. We\u2019ll likely see more issues introduced by people who weren\u2019t being careless, just less aware of what an agent a few steps removed did on their behalf.
\nAt Red Hat, we see a real spectrum in how teams approach AI-assisted development\u2014similar to the broader industry. At one extreme, some teams genuinely don\u2019t look at the code an agent produces. At the other, everything is handcrafted and reviewed by humans. Most teams land somewhere in the middle.
\nWhat gives the most AI-forward teams confidence that they\u2019re not building something that will collapse under its own weight is that they\u2019ve built a serious harness around the entire development and deployment process, effectively a system of checks at every stage. You define what \u201cgood\u201d looks like for your codebase, simple principles like preferring simplicity over complexity or refactoring instead of letting the same rough code multiply. Then you build an adversarial reviewer (agent) that checks new code against those standards and sends it back with feedback until it passes. That point-counterpoint dynamic repeats at every stage of the pipeline: security scanning, penetration testing, active searches for exploitable code paths.
\nThat adversarial structure is, I think, how we avoid trading one set of risks for another. It\u2019s also why those leading teams aren\u2019t spending all their time on manual code review\u2014they\u2019re spending it building and refining the systems that do that review for them.
\nShilpa Khatana: Going back to your point about the risk of code quality degrading when teams don\u2019t actively refactor or simplify: Have you seen AI accelerate the pace at which vulnerabilities, or technical debt more broadly, get created, simply because code is now so much easier to generate?
\nChris Wright: I\u2019d argue technical debt and security vulnerabilities are often two symptoms of the same root cause: code that was written quickly, without enough thought given to how it fits into the broader system. Humans have always introduced both by taking shortcuts under time pressure, just as AI can introduce sloppy or buggy code that\u2019s harder to maintain and easier to exploit. We\u2019ve built institutional processes over time to identify where we\u2019re accumulating debt and to schedule work to pay it down. The task now is to take that same judgment, which used to live only in people\u2019s heads, and encode it into the specifications, skills, and context we give agents.
\nThere is one real difference worth flagging: Models are extremely good at optimizing for exactly the goal you give them, sometimes too literally. It\u2019s the same failure mode you see when an agent \u201ccheats\u201d on a test because its only instruction was to pass it, and nobody told it not to cheat. We have to be explicit about what good architecture looks like, what it means to secure codebase, what constitutes technical debt, and how to avoid introducing it in the first place. And that needs to be built into the guidance we give development agents\u2014alongside instructions to favor simplicity and avoid unnecessary layers of abstraction.
\nShilpa Khatana: How optimistic are you that that can be managed effectively?
\nChris Wright: What gives me confidence that this is manageable is that we\u2019ve looked at how much code gets rewritten over time on real projects. Over a three-year window during a high-growth phase, a project like Kubernetes saw roughly 70 to 80 percent of its code rewritten. Even a mature, three-decade-old project like the Linux kernel saw about 25 to 30 percent of its code rewritten over that same three-year period. Code gets revisited and improved continuously, with or without AI. While it looks different with agents in the loop, I don\u2019t see anything intrinsic to AI that makes technical debt a fundamentally bigger problem. It changes how precisely you need to specify what \u201cgood\u201d means, not whether the problem is more or less solvable.
\nMartin Harrysson: What do you think is underappreciated in this area right now? Is there something you wish more executives were focused on?
\nChris Wright: Some of it we\u2019ve already touched on\u2014understanding where you\u2019re introducing risk through the software you build and the dependencies you rely on. Once you ship software, you\u2019ve effectively become a vendor of your own product, with all the hygiene obligations that come with that\u2014knowing what you\u2019ve deployed, where, and how you\u2019ll update it. This isn\u2019t a new topic, but it tends to go in and out of fashion, and I think it deserves to stay in the foreground permanently.
\nI\u2019ve spent a lot of my career in security, and one thing I\u2019d urge leaders to do is get ahead of the narrative that open source is somehow inherently insecure. It\u2019s not; it\u2019s core infrastructure, and it\u2019s something we can improve collectively faster than most alternatives.
\nThe other area I\u2019d point to is open models. The code underlying most AI infrastructure\u2014agent frameworks, training libraries, harnesses\u2014is reasonably well understood and often open-source itself, through projects like PyTorch. But we haven\u2019t figured out the right collaborative model for open-weight models themselves: how we build them together and, more importantly, how we review a model\u2019s content for efficacy, safety, and security once it\u2019s built. We have decent guardrails, but the internals of how weights interact inside a large stochastic system aren\u2019t well understood industry-wide. I\u2019d like to see the same collaborative, transparent approach we\u2019ve applied to open-source software\u2014not just open weights, but open training data and training recipes, so the work is genuinely reproducible\u2014extended to models themselves. That\u2019s where I think the state of the art still needs to catch up.
", "summary": "Over his roughly three decades in the open-source software industry, including the last two at Red Hat, Chris Wright has watched the sector evolve from a once niche technical field to a mainstream building block of enterprise infrastructure. As Red Hat\u2019s chief technology officer, Wright now oversees engineering across the company, from emerging-technology research to product development and ongoing support of products already in the market. In a recent conversation with McKinsey Senior Partner Martin Harrysson and Associate Partner Shilpa Khatana, Wright discussed how a recent wave of AI-driven discoveries of code vulnerabilities is reshaping trust in the software industry and why he believes the industry will emerge from this moment more secure.
\nThis interview has been edited for length and clarity.
\nMartin Harrysson: You\u2019ve been at Red Hat for more than two decades, and in the open-source community, including early years involved with Linux, for around three. What strikes you about where open source finds itself today?
\nChris Wright: It\u2019s a fascinating time for open source. Over 30 years, it went from a fringe, hobbyist activity to something pervasive across every industry\u2014consumer, enterprise, small business, data centers, devices. Open source is foundational to the global economy today. The vast majority of enterprise codebases\u2014well over 80 percent, typically over 90 percent\u2014contain open-source components. Every hyperscaler and every large consumer platform runs on open-source infrastructure. It\u2019s entrenched, and it\u2019s fundamental to our future.
\nNow it\u2019s back at the center of the conversation, this time because of cybersecurity concerns and AI\u2019s role in scanning code and even building models with open weights. There\u2019s a full-circle aspect to it: Many of the same debates we had decades ago about trusting open source are resurfacing in a new technology era.
\nOne of open source\u2019s greatest strengths is also what\u2019s driving today\u2019s concerns: It\u2019s open. Anyone can access the code, learn from it, and build with it; that\u2019s real democratization of technology. Now that AI models have been trained on the internet, including a huge amount of open-source code, they\u2019ve become remarkably capable at understanding, writing, documenting, and finding issues in code.
\nMartin Harrysson: Given those real concerns, how do you see AI reshaping open source\u2019s role and importance?
\nChris Wright: Starting around the middle of this year, a new class of cybersecurity-focused models\u2014beginning with Mythos and the Glasswing project, and followed by efforts like Daybreak\u2014began uncovering vulnerabilities that had been sitting latent in source code for years, in some cases upward of two decades. Some are individually exploitable; others are chains of smaller vulnerabilities that together give attackers a path to escalate privileges or break into systems. Because open-source code is visible and accessible, it has been the natural target for this kind of scanning.
\nThat\u2019s produced a narrative that open source is somehow uniquely insecure. I think that\u2019s the wrong conclusion. Open source has always made a virtue of transparency\u2014sunlight is a disinfectant. You can see exactly how a cryptographic algorithm is implemented rather than being told to just trust a vendor\u2019s word for it. And the community development model brings a much broader set of eyes to any given piece of code than a small proprietary team ever could. That means faster discovery and faster remediation\u2014there\u2019s no hiding behind security through obscurity.
\nMartin Harrysson: So you see this having a positive overall impact?
\nChris Wright: What\u2019s happening now is that AI has made it dramatically cheaper to scan all of that code at once, so we\u2019re seeing a concentrated burst of vulnerability discovery. It\u2019s a genuinely uncomfortable moment: Customers are worried, and rightly so, because open source is so pervasive that they often lack full visibility into what\u2019s actually deployed, especially when it arrives bundled inside third-party or commercial software. But I\u2019m optimistic about where this ends. We\u2019ll come out the other side having raised the bar for security across the board\u2014more scanning, faster patching, and security testing built directly into project pipelines. Efforts like Daybreak, which has talked about \u201cpatching the planet,\u201d are genuinely global in scope in a way that would be much harder to achieve if we were only securing smaller, more fragmented, proprietary codebases. And this is exactly what Red Hat and IBM are doing with our Lightwell initiative: making open source safer for enterprises.
\nMartin Harrysson: In the near term, what should a CIO, CTO, or CEO be doing about this?
\nChris Wright: I\u2019d point to three priorities\u2014two are basic hygiene, and the third is more about where this is all heading.
\nFirst: Know what software you have and where it\u2019s running. If you can\u2019t produce something like a software bill of materials or a dependency graph across your enterprise, you have no real way to assess your risk. That visibility also helps you prioritize\u2014not every system carries the same value or the same exposure. A database full of sensitive information behind several layers of firewalls is a very different risk profile than a public-facing service.
\nSecond: Build the automation to actually act on what you find. Imagine every vulnerability in your environment has already been discovered. Now you need to patch potentially hundreds of thousands of dependencies across your infrastructure. If your patching process still involves months of manual review and a spreadsheet, you will never catch up\u2014the backlog will always outrun your capacity. Mature CI/CD and DevSecOps practices, which the industry has been building toward for years but hasn\u2019t fully adopted everywhere, are what make addressing that kind of scale possible.
\nThird, and this is more forward-looking: The industry needs to keep building the tools such as scanning, exploit research, and mitigation techniques that let defenders keep pace with attackers. Right now, I think attackers have the upper hand. We\u2019re going to discover entirely new classes of vulnerabilities, not just new instances of existing ones, and each new class will trigger a fresh wave of activity. I think we\u2019re heading toward a world where both attack and defense increasingly happen at machine speed\u2014agents defending against agents, rather than relying on human-speed response.
\nShilpa Khatana: Who should own this? In most enterprises, we see engineering taking on more responsibility for security, but there\u2019s also a dedicated cybersecurity function, risk, governance, and now often an AI council. How should those groups divide the work?
\nChris Wright: It\u2019s a shared responsibility, but with real specialization by role.
\nThis builds on the shift we made with DevOps and then DevSecOps\u2014bringing security earlier into the development process, on the theory that a vulnerability caught early is far cheaper to fix than one caught in production. But developers aren\u2019t security specialists, so DevSecOps also means building automation and gates into the pipeline that give developers visibility into security issues without requiring them to become experts themselves.
\nThe application development team owns what\u2019s deployed and where\u2014they\u2019re best positioned to know what dependencies went into their software. One thing I don\u2019t think is well understood: When you write code, you take responsibility for it, but the moment you pull in a dependency from the internet, you\u2019ve taken on responsibility for that code as well, even a small, stable utility you\u2019ve never had reason to question.
\nAt the same time, other teams need to specialize in things individual developers shouldn\u2019t be expected to do alone: \u201cred-teaming\u201d at a systems level, setting compliance requirements as a CISO function, and translating policy into automated checks embedded in the pipeline. So you end up with a CISO setting requirements, teams turning that into policy as code, and development teams having responsibility for what they build and what they pull in. It\u2019s shared, but each group brings a different kind of expertise.
\nMartin Harrysson: The rise of agentic AI adds another layer to this\u2014agents are now writing code and can themselves introduce dependencies, sometimes through subagents. Doesn\u2019t that complicate accountability?
\nChris Wright: You\u2019re still ultimately accountable if you initiated the work\u2014that doesn\u2019t change. But I agree these layers of abstraction can obscure what\u2019s actually happening. We\u2019ll likely see more issues introduced by people who weren\u2019t being careless, just less aware of what an agent a few steps removed did on their behalf.
\nAt Red Hat, we see a real spectrum in how teams approach AI-assisted development\u2014similar to the broader industry. At one extreme, some teams genuinely don\u2019t look at the code an agent produces. At the other, everything is handcrafted and reviewed by humans. Most teams land somewhere in the middle.
\nWhat gives the most AI-forward teams confidence that they\u2019re not building something that will collapse under its own weight is that they\u2019ve built a serious harness around the entire development and deployment process, effectively a system of checks at every stage. You define what \u201cgood\u201d looks like for your codebase, simple principles like preferring simplicity over complexity or refactoring instead of letting the same rough code multiply. Then you build an adversarial reviewer (agent) that checks new code against those standards and sends it back with feedback until it passes. That point-counterpoint dynamic repeats at every stage of the pipeline: security scanning, penetration testing, active searches for exploitable code paths.
\nThat adversarial structure is, I think, how we avoid trading one set of risks for another. It\u2019s also why those leading teams aren\u2019t spending all their time on manual code review\u2014they\u2019re spending it building and refining the systems that do that review for them.
\nShilpa Khatana: Going back to your point about the risk of code quality degrading when teams don\u2019t actively refactor or simplify: Have you seen AI accelerate the pace at which vulnerabilities, or technical debt more broadly, get created, simply because code is now so much easier to generate?
\nChris Wright: I\u2019d argue technical debt and security vulnerabilities are often two symptoms of the same root cause: code that was written quickly, without enough thought given to how it fits into the broader system. Humans have always introduced both by taking shortcuts under time pressure, just as AI can introduce sloppy or buggy code that\u2019s harder to maintain and easier to exploit. We\u2019ve built institutional processes over time to identify where we\u2019re accumulating debt and to schedule work to pay it down. The task now is to take that same judgment, which used to live only in people\u2019s heads, and encode it into the specifications, skills, and context we give agents.
\nThere is one real difference worth flagging: Models are extremely good at optimizing for exactly the goal you give them, sometimes too literally. It\u2019s the same failure mode you see when an agent \u201ccheats\u201d on a test because its only instruction was to pass it, and nobody told it not to cheat. We have to be explicit about what good architecture looks like, what it means to secure codebase, what constitutes technical debt, and how to avoid introducing it in the first place. And that needs to be built into the guidance we give development agents\u2014alongside instructions to favor simplicity and avoid unnecessary layers of abstraction.
\nShilpa Khatana: How optimistic are you that that can be managed effectively?
\nChris Wright: What gives me confidence that this is manageable is that we\u2019ve looked at how much code gets rewritten over time on real projects. Over a three-year window during a high-growth phase, a project like Kubernetes saw roughly 70 to 80 percent of its code rewritten. Even a mature, three-decade-old project like the Linux kernel saw about 25 to 30 percent of its code rewritten over that same three-year period. Code gets revisited and improved continuously, with or without AI. While it looks different with agents in the loop, I don\u2019t see anything intrinsic to AI that makes technical debt a fundamentally bigger problem. It changes how precisely you need to specify what \u201cgood\u201d means, not whether the problem is more or less solvable.
\nMartin Harrysson: What do you think is underappreciated in this area right now? Is there something you wish more executives were focused on?
\nChris Wright: Some of it we\u2019ve already touched on\u2014understanding where you\u2019re introducing risk through the software you build and the dependencies you rely on. Once you ship software, you\u2019ve effectively become a vendor of your own product, with all the hygiene obligations that come with that\u2014knowing what you\u2019ve deployed, where, and how you\u2019ll update it. This isn\u2019t a new topic, but it tends to go in and out of fashion, and I think it deserves to stay in the foreground permanently.
\nI\u2019ve spent a lot of my career in security, and one thing I\u2019d urge leaders to do is get ahead of the narrative that open source is somehow inherently insecure. It\u2019s not; it\u2019s core infrastructure, and it\u2019s something we can improve collectively faster than most alternatives.
\nThe other area I\u2019d point to is open models. The code underlying most AI infrastructure\u2014agent frameworks, training libraries, harnesses\u2014is reasonably well understood and often open-source itself, through projects like PyTorch. But we haven\u2019t figured out the right collaborative model for open-weight models themselves: how we build them together and, more importantly, how we review a model\u2019s content for efficacy, safety, and security once it\u2019s built. We have decent guardrails, but the internals of how weights interact inside a large stochastic system aren\u2019t well understood industry-wide. I\u2019d like to see the same collaborative, transparent approach we\u2019ve applied to open-source software\u2014not just open weights, but open training data and training recipes, so the work is genuinely reproducible\u2014extended to models themselves. That\u2019s where I think the state of the art still needs to catch up.
", "image": null, "banner_image": null, "date_published": "2026-10-19T16:00:00Z", "date_modified": "2026-10-19T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}, {"id": "8cbcfe1c-535f-4e44-8084-2d8bae412097", "url": "https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/three-horizons-of-quantum-a-business-road-map-to-the-future-computing-economy", "external_url": null, "title": "Three horizons of quantum: A business road map to the future computing economy", "content_text": "", "content_html": "In the next decade, quantum technology could create trillions in value and break encryption, delivering opportunity but also risk. Leading companies are acting now to secure a competitive advantage.", "summary": "In the next decade, quantum technology could create trillions in value and break encryption, delivering opportunity but also risk. Leading companies are acting now to secure a competitive advantage.", "image": null, "banner_image": null, "date_published": "2026-10-19T16:00:00Z", "date_modified": "2026-10-19T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}, {"id": "4817c6ed-07b3-45d0-a446-64c040e1a7b7", "url": "https://www.mckinsey.com/capabilities/tech-and-ai/our-insights/securing-software-in-the-age-of-ai-red-hat-cto-chris-wright", "external_url": null, "title": "Securing software in the age of AI: Red Hat CTO Chris Wright", "content_text": "", "content_html": "Over his roughly three decades in the open-source software industry, including the last two at Red Hat, Chris Wright has watched the sector evolve from a once niche technical field to a mainstream building block of enterprise infrastructure. As Red Hat\u2019s chief technology officer, Wright now oversees engineering across the company, from emerging-technology research to product development and ongoing support of products already in the market. In a recent conversation with McKinsey Senior Partner Martin Harrysson and Associate Partner Shilpa Khatana, Wright discussed how a recent wave of AI-driven discoveries of code vulnerabilities is reshaping trust in the software industry and why he believes the industry will emerge from this moment more secure.
\nThis interview has been edited for length and clarity.
\nMartin Harrysson: You\u2019ve been at Red Hat for more than two decades, and in the open-source community, including early years involved with Linux, for around three. What strikes you about where open source finds itself today?
\nChris Wright: It\u2019s a fascinating time for open source. Over 30 years, it went from a fringe, hobbyist activity to something pervasive across every industry\u2014consumer, enterprise, small business, data centers, devices. Open source is foundational to the global economy today. The vast majority of enterprise codebases\u2014well over 80 percent, typically over 90 percent\u2014contain open-source components. Every hyperscaler and every large consumer platform runs on open-source infrastructure. It\u2019s entrenched, and it\u2019s fundamental to our future.
\nNow it\u2019s back at the center of the conversation, this time because of cybersecurity concerns and AI\u2019s role in scanning code and even building models with open weights. There\u2019s a full-circle aspect to it: Many of the same debates we had decades ago about trusting open source are resurfacing in a new technology era.
\nOne of open source\u2019s greatest strengths is also what\u2019s driving today\u2019s concerns: It\u2019s open. Anyone can access the code, learn from it, and build with it; that\u2019s real democratization of technology. Now that AI models have been trained on the internet, including a huge amount of open-source code, they\u2019ve become remarkably capable at understanding, writing, documenting, and finding issues in code.
\nMartin Harrysson: Given those real concerns, how do you see AI reshaping open source\u2019s role and importance?
\nChris Wright: Starting around the middle of this year, a new class of cybersecurity-focused models\u2014beginning with Mythos and the Glasswing project, and followed by efforts like Daybreak\u2014began uncovering vulnerabilities that had been sitting latent in source code for years, in some cases upward of two decades. Some are individually exploitable; others are chains of smaller vulnerabilities that together give attackers a path to escalate privileges or break into systems. Because open-source code is visible and accessible, it has been the natural target for this kind of scanning.
\nThat\u2019s produced a narrative that open source is somehow uniquely insecure. I think that\u2019s the wrong conclusion. Open source has always made a virtue of transparency\u2014sunlight is a disinfectant. You can see exactly how a cryptographic algorithm is implemented rather than being told to just trust a vendor\u2019s word for it. And the community development model brings a much broader set of eyes to any given piece of code than a small proprietary team ever could. That means faster discovery and faster remediation\u2014there\u2019s no hiding behind security through obscurity.
\nMartin Harrysson: So you see this having a positive overall impact?
\nChris Wright: What\u2019s happening now is that AI has made it dramatically cheaper to scan all of that code at once, so we\u2019re seeing a concentrated burst of vulnerability discovery. It\u2019s a genuinely uncomfortable moment: Customers are worried, and rightly so, because open source is so pervasive that they often lack full visibility into what\u2019s actually deployed, especially when it arrives bundled inside third-party or commercial software. But I\u2019m optimistic about where this ends. We\u2019ll come out the other side having raised the bar for security across the board\u2014more scanning, faster patching, and security testing built directly into project pipelines. Efforts like Daybreak, which has talked about \u201cpatching the planet,\u201d are genuinely global in scope in a way that would be much harder to achieve if we were only securing smaller, more fragmented, proprietary codebases. And this is exactly what Red Hat and IBM are doing with our Lightwell initiative: making open source safer for enterprises.
\nMartin Harrysson: In the near term, what should a CIO, CTO, or CEO be doing about this?
\nChris Wright: I\u2019d point to three priorities\u2014two are basic hygiene, and the third is more about where this is all heading.
\nFirst: Know what software you have and where it\u2019s running. If you can\u2019t produce something like a software bill of materials or a dependency graph across your enterprise, you have no real way to assess your risk. That visibility also helps you prioritize\u2014not every system carries the same value or the same exposure. A database full of sensitive information behind several layers of firewalls is a very different risk profile than a public-facing service.
\nSecond: Build the automation to actually act on what you find. Imagine every vulnerability in your environment has already been discovered. Now you need to patch potentially hundreds of thousands of dependencies across your infrastructure. If your patching process still involves months of manual review and a spreadsheet, you will never catch up\u2014the backlog will always outrun your capacity. Mature CI/CD and DevSecOps practices, which the industry has been building toward for years but hasn\u2019t fully adopted everywhere, are what make addressing that kind of scale possible.
\nThird, and this is more forward-looking: The industry needs to keep building the tools such as scanning, exploit research, and mitigation techniques that let defenders keep pace with attackers. Right now, I think attackers have the upper hand. We\u2019re going to discover entirely new classes of vulnerabilities, not just new instances of existing ones, and each new class will trigger a fresh wave of activity. I think we\u2019re heading toward a world where both attack and defense increasingly happen at machine speed\u2014agents defending against agents, rather than relying on human-speed response.
\nShilpa Khatana: Who should own this? In most enterprises, we see engineering taking on more responsibility for security, but there\u2019s also a dedicated cybersecurity function, risk, governance, and now often an AI council. How should those groups divide the work?
\nChris Wright: It\u2019s a shared responsibility, but with real specialization by role.
\nThis builds on the shift we made with DevOps and then DevSecOps\u2014bringing security earlier into the development process, on the theory that a vulnerability caught early is far cheaper to fix than one caught in production. But developers aren\u2019t security specialists, so DevSecOps also means building automation and gates into the pipeline that give developers visibility into security issues without requiring them to become experts themselves.
\nThe application development team owns what\u2019s deployed and where\u2014they\u2019re best positioned to know what dependencies went into their software. One thing I don\u2019t think is well understood: When you write code, you take responsibility for it, but the moment you pull in a dependency from the internet, you\u2019ve taken on responsibility for that code as well, even a small, stable utility you\u2019ve never had reason to question.
\nAt the same time, other teams need to specialize in things individual developers shouldn\u2019t be expected to do alone: \u201cred-teaming\u201d at a systems level, setting compliance requirements as a CISO function, and translating policy into automated checks embedded in the pipeline. So you end up with a CISO setting requirements, teams turning that into policy as code, and development teams having responsibility for what they build and what they pull in. It\u2019s shared, but each group brings a different kind of expertise.
\nMartin Harrysson: The rise of agentic AI adds another layer to this\u2014agents are now writing code and can themselves introduce dependencies, sometimes through subagents. Doesn\u2019t that complicate accountability?
\nChris Wright: You\u2019re still ultimately accountable if you initiated the work\u2014that doesn\u2019t change. But I agree these layers of abstraction can obscure what\u2019s actually happening. We\u2019ll likely see more issues introduced by people who weren\u2019t being careless, just less aware of what an agent a few steps removed did on their behalf.
\nAt Red Hat, we see a real spectrum in how teams approach AI-assisted development\u2014similar to the broader industry. At one extreme, some teams genuinely don\u2019t look at the code an agent produces. At the other, everything is handcrafted and reviewed by humans. Most teams land somewhere in the middle.
\nWhat gives the most AI-forward teams confidence that they\u2019re not building something that will collapse under its own weight is that they\u2019ve built a serious harness around the entire development and deployment process, effectively a system of checks at every stage. You define what \u201cgood\u201d looks like for your codebase, simple principles like preferring simplicity over complexity or refactoring instead of letting the same rough code multiply. Then you build an adversarial reviewer (agent) that checks new code against those standards and sends it back with feedback until it passes. That point-counterpoint dynamic repeats at every stage of the pipeline: security scanning, penetration testing, active searches for exploitable code paths.
\nThat adversarial structure is, I think, how we avoid trading one set of risks for another. It\u2019s also why those leading teams aren\u2019t spending all their time on manual code review\u2014they\u2019re spending it building and refining the systems that do that review for them.
\nShilpa Khatana: Going back to your point about the risk of code quality degrading when teams don\u2019t actively refactor or simplify: Have you seen AI accelerate the pace at which vulnerabilities, or technical debt more broadly, get created, simply because code is now so much easier to generate?
\nChris Wright: I\u2019d argue technical debt and security vulnerabilities are often two symptoms of the same root cause: code that was written quickly, without enough thought given to how it fits into the broader system. Humans have always introduced both by taking shortcuts under time pressure, just as AI can introduce sloppy or buggy code that\u2019s harder to maintain and easier to exploit. We\u2019ve built institutional processes over time to identify where we\u2019re accumulating debt and to schedule work to pay it down. The task now is to take that same judgment, which used to live only in people\u2019s heads, and encode it into the specifications, skills, and context we give agents.
\nThere is one real difference worth flagging: Models are extremely good at optimizing for exactly the goal you give them, sometimes too literally. It\u2019s the same failure mode you see when an agent \u201ccheats\u201d on a test because its only instruction was to pass it, and nobody told it not to cheat. We have to be explicit about what good architecture looks like, what it means to secure codebase, what constitutes technical debt, and how to avoid introducing it in the first place. And that needs to be built into the guidance we give development agents\u2014alongside instructions to favor simplicity and avoid unnecessary layers of abstraction.
\nShilpa Khatana: How optimistic are you that that can be managed effectively?
\nChris Wright: What gives me confidence that this is manageable is that we\u2019ve looked at how much code gets rewritten over time on real projects. Over a three-year window during a high-growth phase, a project like Kubernetes saw roughly 70 to 80 percent of its code rewritten. Even a mature, three-decade-old project like the Linux kernel saw about 25 to 30 percent of its code rewritten over that same three-year period. Code gets revisited and improved continuously, with or without AI. While it looks different with agents in the loop, I don\u2019t see anything intrinsic to AI that makes technical debt a fundamentally bigger problem. It changes how precisely you need to specify what \u201cgood\u201d means, not whether the problem is more or less solvable.
\nMartin Harrysson: What do you think is underappreciated in this area right now? Is there something you wish more executives were focused on?
\nChris Wright: Some of it we\u2019ve already touched on\u2014understanding where you\u2019re introducing risk through the software you build and the dependencies you rely on. Once you ship software, you\u2019ve effectively become a vendor of your own product, with all the hygiene obligations that come with that\u2014knowing what you\u2019ve deployed, where, and how you\u2019ll update it. This isn\u2019t a new topic, but it tends to go in and out of fashion, and I think it deserves to stay in the foreground permanently.
\nI\u2019ve spent a lot of my career in security, and one thing I\u2019d urge leaders to do is get ahead of the narrative that open source is somehow inherently insecure. It\u2019s not; it\u2019s core infrastructure, and it\u2019s something we can improve collectively faster than most alternatives.
\nThe other area I\u2019d point to is open models. The code underlying most AI infrastructure\u2014agent frameworks, training libraries, harnesses\u2014is reasonably well understood and often open-source itself, through projects like PyTorch. But we haven\u2019t figured out the right collaborative model for open-weight models themselves: how we build them together and, more importantly, how we review a model\u2019s content for efficacy, safety, and security once it\u2019s built. We have decent guardrails, but the internals of how weights interact inside a large stochastic system aren\u2019t well understood industry-wide. I\u2019d like to see the same collaborative, transparent approach we\u2019ve applied to open-source software\u2014not just open weights, but open training data and training recipes, so the work is genuinely reproducible\u2014extended to models themselves. That\u2019s where I think the state of the art still needs to catch up.
", "summary": "Over his roughly three decades in the open-source software industry, including the last two at Red Hat, Chris Wright has watched the sector evolve from a once niche technical field to a mainstream building block of enterprise infrastructure. As Red Hat\u2019s chief technology officer, Wright now oversees engineering across the company, from emerging-technology research to product development and ongoing support of products already in the market. In a recent conversation with McKinsey Senior Partner Martin Harrysson and Associate Partner Shilpa Khatana, Wright discussed how a recent wave of AI-driven discoveries of code vulnerabilities is reshaping trust in the software industry and why he believes the industry will emerge from this moment more secure.
\nThis interview has been edited for length and clarity.
\nMartin Harrysson: You\u2019ve been at Red Hat for more than two decades, and in the open-source community, including early years involved with Linux, for around three. What strikes you about where open source finds itself today?
\nChris Wright: It\u2019s a fascinating time for open source. Over 30 years, it went from a fringe, hobbyist activity to something pervasive across every industry\u2014consumer, enterprise, small business, data centers, devices. Open source is foundational to the global economy today. The vast majority of enterprise codebases\u2014well over 80 percent, typically over 90 percent\u2014contain open-source components. Every hyperscaler and every large consumer platform runs on open-source infrastructure. It\u2019s entrenched, and it\u2019s fundamental to our future.
\nNow it\u2019s back at the center of the conversation, this time because of cybersecurity concerns and AI\u2019s role in scanning code and even building models with open weights. There\u2019s a full-circle aspect to it: Many of the same debates we had decades ago about trusting open source are resurfacing in a new technology era.
\nOne of open source\u2019s greatest strengths is also what\u2019s driving today\u2019s concerns: It\u2019s open. Anyone can access the code, learn from it, and build with it; that\u2019s real democratization of technology. Now that AI models have been trained on the internet, including a huge amount of open-source code, they\u2019ve become remarkably capable at understanding, writing, documenting, and finding issues in code.
\nMartin Harrysson: Given those real concerns, how do you see AI reshaping open source\u2019s role and importance?
\nChris Wright: Starting around the middle of this year, a new class of cybersecurity-focused models\u2014beginning with Mythos and the Glasswing project, and followed by efforts like Daybreak\u2014began uncovering vulnerabilities that had been sitting latent in source code for years, in some cases upward of two decades. Some are individually exploitable; others are chains of smaller vulnerabilities that together give attackers a path to escalate privileges or break into systems. Because open-source code is visible and accessible, it has been the natural target for this kind of scanning.
\nThat\u2019s produced a narrative that open source is somehow uniquely insecure. I think that\u2019s the wrong conclusion. Open source has always made a virtue of transparency\u2014sunlight is a disinfectant. You can see exactly how a cryptographic algorithm is implemented rather than being told to just trust a vendor\u2019s word for it. And the community development model brings a much broader set of eyes to any given piece of code than a small proprietary team ever could. That means faster discovery and faster remediation\u2014there\u2019s no hiding behind security through obscurity.
\nMartin Harrysson: So you see this having a positive overall impact?
\nChris Wright: What\u2019s happening now is that AI has made it dramatically cheaper to scan all of that code at once, so we\u2019re seeing a concentrated burst of vulnerability discovery. It\u2019s a genuinely uncomfortable moment: Customers are worried, and rightly so, because open source is so pervasive that they often lack full visibility into what\u2019s actually deployed, especially when it arrives bundled inside third-party or commercial software. But I\u2019m optimistic about where this ends. We\u2019ll come out the other side having raised the bar for security across the board\u2014more scanning, faster patching, and security testing built directly into project pipelines. Efforts like Daybreak, which has talked about \u201cpatching the planet,\u201d are genuinely global in scope in a way that would be much harder to achieve if we were only securing smaller, more fragmented, proprietary codebases. And this is exactly what Red Hat and IBM are doing with our Lightwell initiative: making open source safer for enterprises.
\nMartin Harrysson: In the near term, what should a CIO, CTO, or CEO be doing about this?
\nChris Wright: I\u2019d point to three priorities\u2014two are basic hygiene, and the third is more about where this is all heading.
\nFirst: Know what software you have and where it\u2019s running. If you can\u2019t produce something like a software bill of materials or a dependency graph across your enterprise, you have no real way to assess your risk. That visibility also helps you prioritize\u2014not every system carries the same value or the same exposure. A database full of sensitive information behind several layers of firewalls is a very different risk profile than a public-facing service.
\nSecond: Build the automation to actually act on what you find. Imagine every vulnerability in your environment has already been discovered. Now you need to patch potentially hundreds of thousands of dependencies across your infrastructure. If your patching process still involves months of manual review and a spreadsheet, you will never catch up\u2014the backlog will always outrun your capacity. Mature CI/CD and DevSecOps practices, which the industry has been building toward for years but hasn\u2019t fully adopted everywhere, are what make addressing that kind of scale possible.
\nThird, and this is more forward-looking: The industry needs to keep building the tools such as scanning, exploit research, and mitigation techniques that let defenders keep pace with attackers. Right now, I think attackers have the upper hand. We\u2019re going to discover entirely new classes of vulnerabilities, not just new instances of existing ones, and each new class will trigger a fresh wave of activity. I think we\u2019re heading toward a world where both attack and defense increasingly happen at machine speed\u2014agents defending against agents, rather than relying on human-speed response.
\nShilpa Khatana: Who should own this? In most enterprises, we see engineering taking on more responsibility for security, but there\u2019s also a dedicated cybersecurity function, risk, governance, and now often an AI council. How should those groups divide the work?
\nChris Wright: It\u2019s a shared responsibility, but with real specialization by role.
\nThis builds on the shift we made with DevOps and then DevSecOps\u2014bringing security earlier into the development process, on the theory that a vulnerability caught early is far cheaper to fix than one caught in production. But developers aren\u2019t security specialists, so DevSecOps also means building automation and gates into the pipeline that give developers visibility into security issues without requiring them to become experts themselves.
\nThe application development team owns what\u2019s deployed and where\u2014they\u2019re best positioned to know what dependencies went into their software. One thing I don\u2019t think is well understood: When you write code, you take responsibility for it, but the moment you pull in a dependency from the internet, you\u2019ve taken on responsibility for that code as well, even a small, stable utility you\u2019ve never had reason to question.
\nAt the same time, other teams need to specialize in things individual developers shouldn\u2019t be expected to do alone: \u201cred-teaming\u201d at a systems level, setting compliance requirements as a CISO function, and translating policy into automated checks embedded in the pipeline. So you end up with a CISO setting requirements, teams turning that into policy as code, and development teams having responsibility for what they build and what they pull in. It\u2019s shared, but each group brings a different kind of expertise.
\nMartin Harrysson: The rise of agentic AI adds another layer to this\u2014agents are now writing code and can themselves introduce dependencies, sometimes through subagents. Doesn\u2019t that complicate accountability?
\nChris Wright: You\u2019re still ultimately accountable if you initiated the work\u2014that doesn\u2019t change. But I agree these layers of abstraction can obscure what\u2019s actually happening. We\u2019ll likely see more issues introduced by people who weren\u2019t being careless, just less aware of what an agent a few steps removed did on their behalf.
\nAt Red Hat, we see a real spectrum in how teams approach AI-assisted development\u2014similar to the broader industry. At one extreme, some teams genuinely don\u2019t look at the code an agent produces. At the other, everything is handcrafted and reviewed by humans. Most teams land somewhere in the middle.
\nWhat gives the most AI-forward teams confidence that they\u2019re not building something that will collapse under its own weight is that they\u2019ve built a serious harness around the entire development and deployment process, effectively a system of checks at every stage. You define what \u201cgood\u201d looks like for your codebase, simple principles like preferring simplicity over complexity or refactoring instead of letting the same rough code multiply. Then you build an adversarial reviewer (agent) that checks new code against those standards and sends it back with feedback until it passes. That point-counterpoint dynamic repeats at every stage of the pipeline: security scanning, penetration testing, active searches for exploitable code paths.
\nThat adversarial structure is, I think, how we avoid trading one set of risks for another. It\u2019s also why those leading teams aren\u2019t spending all their time on manual code review\u2014they\u2019re spending it building and refining the systems that do that review for them.
\nShilpa Khatana: Going back to your point about the risk of code quality degrading when teams don\u2019t actively refactor or simplify: Have you seen AI accelerate the pace at which vulnerabilities, or technical debt more broadly, get created, simply because code is now so much easier to generate?
\nChris Wright: I\u2019d argue technical debt and security vulnerabilities are often two symptoms of the same root cause: code that was written quickly, without enough thought given to how it fits into the broader system. Humans have always introduced both by taking shortcuts under time pressure, just as AI can introduce sloppy or buggy code that\u2019s harder to maintain and easier to exploit. We\u2019ve built institutional processes over time to identify where we\u2019re accumulating debt and to schedule work to pay it down. The task now is to take that same judgment, which used to live only in people\u2019s heads, and encode it into the specifications, skills, and context we give agents.
\nThere is one real difference worth flagging: Models are extremely good at optimizing for exactly the goal you give them, sometimes too literally. It\u2019s the same failure mode you see when an agent \u201ccheats\u201d on a test because its only instruction was to pass it, and nobody told it not to cheat. We have to be explicit about what good architecture looks like, what it means to secure codebase, what constitutes technical debt, and how to avoid introducing it in the first place. And that needs to be built into the guidance we give development agents\u2014alongside instructions to favor simplicity and avoid unnecessary layers of abstraction.
\nShilpa Khatana: How optimistic are you that that can be managed effectively?
\nChris Wright: What gives me confidence that this is manageable is that we\u2019ve looked at how much code gets rewritten over time on real projects. Over a three-year window during a high-growth phase, a project like Kubernetes saw roughly 70 to 80 percent of its code rewritten. Even a mature, three-decade-old project like the Linux kernel saw about 25 to 30 percent of its code rewritten over that same three-year period. Code gets revisited and improved continuously, with or without AI. While it looks different with agents in the loop, I don\u2019t see anything intrinsic to AI that makes technical debt a fundamentally bigger problem. It changes how precisely you need to specify what \u201cgood\u201d means, not whether the problem is more or less solvable.
\nMartin Harrysson: What do you think is underappreciated in this area right now? Is there something you wish more executives were focused on?
\nChris Wright: Some of it we\u2019ve already touched on\u2014understanding where you\u2019re introducing risk through the software you build and the dependencies you rely on. Once you ship software, you\u2019ve effectively become a vendor of your own product, with all the hygiene obligations that come with that\u2014knowing what you\u2019ve deployed, where, and how you\u2019ll update it. This isn\u2019t a new topic, but it tends to go in and out of fashion, and I think it deserves to stay in the foreground permanently.
\nI\u2019ve spent a lot of my career in security, and one thing I\u2019d urge leaders to do is get ahead of the narrative that open source is somehow inherently insecure. It\u2019s not; it\u2019s core infrastructure, and it\u2019s something we can improve collectively faster than most alternatives.
\nThe other area I\u2019d point to is open models. The code underlying most AI infrastructure\u2014agent frameworks, training libraries, harnesses\u2014is reasonably well understood and often open-source itself, through projects like PyTorch. But we haven\u2019t figured out the right collaborative model for open-weight models themselves: how we build them together and, more importantly, how we review a model\u2019s content for efficacy, safety, and security once it\u2019s built. We have decent guardrails, but the internals of how weights interact inside a large stochastic system aren\u2019t well understood industry-wide. I\u2019d like to see the same collaborative, transparent approach we\u2019ve applied to open-source software\u2014not just open weights, but open training data and training recipes, so the work is genuinely reproducible\u2014extended to models themselves. That\u2019s where I think the state of the art still needs to catch up.
", "image": null, "banner_image": null, "date_published": "2026-10-19T16:00:00Z", "date_modified": "2026-10-19T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}, {"id": "979e9880-467f-4021-80d4-4c3513f6d8cc", "url": "https://www.mckinsey.com/industries/automotive-and-assembly/our-insights/switch-tracks-rewiring-rolling-stock-for-a-new-era", "external_url": null, "title": "Switch tracks: Rewiring rolling stock for a new era", "content_text": "", "content_html": "Fast-growing markets in Asia are difficult for global rail OEMs to access. Value could shift from new builds toward servicing and renewals\u2014while OEMs prepare for future opportunities.", "summary": "Fast-growing markets in Asia are difficult for global rail OEMs to access. Value could shift from new builds toward servicing and renewals\u2014while OEMs prepare for future opportunities.", "image": null, "banner_image": null, "date_published": "2026-10-19T16:00:00Z", "date_modified": "2026-10-19T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}, {"id": "https://www.mckinsey.com/insights/d13b7aac-f302-4786-969c-af53334ab9e1", "url": "https://www.mckinsey.com/institute-for-economic-mobility/our-insights/the-hidden-opportunity-in-the-next-2-trillion-of-us-latin-america-trade", "external_url": null, "title": "The hidden opportunity in the next $2 trillion of US\u2013Latin America trade", "content_text": "", "content_html": "Quirch Foods began in Miami in 1967 as a family-owned importer of Latino staple foods and today generates more than $4 billion in annual sales across 40 US states, the Caribbean, and Latin America. Latina-owned Nopalera built a Mexican-inspired beauty brand into hundreds of US retail stores before expanding into Mexico. Angel network Angeles Investors co-led a $32 million Series A investment in Canela Media, aiding its growth to more than 60 million active users across the United States and Latin America. The city of Laredo, with almost 275,000 residents and 36 percent of its nonfarm workforce in trade, has developed distinctive capabilities in logistics, warehousing, and customs brokerage, enabling it to handle $354 billion in international trade in 2025.
\nThese are just a few examples of the importers, exporters, investors, and specialized locations that fuel the US\u2013Latin America trade corridor. It is one of the world\u2019s biggest and most critical economic arteries, generating $1.5 trillion in trade in 2025 \u2014about 1.5 times the total trade between the United States and Canada and the equivalent of 85 percent of all trade within the European Union. Discussion of the corridor tends to center on large forces that shape it, such as trade agreements and tariffs, geopolitics, commodity cycles, inflation, or immigration. These forces help set the baseline of how the corridor may change in overall size. However, analysis of these factors does not fully explain how US companies continue to create enduring value within the corridor in new ways.
\nThe success of companies such as Quirch Foods, Nopalera, and Canela Media shows how US-based Latino consumers, businesses, and places can turn positional and proprietary advantages into value that flows back to communities.
\nLatino consumers in the United States, a population spanning distinct national-origin groups, generations, and cultures, can provide a large and growing base of demand and cultural influence able to extend to other consumer segments. Latino businesses can bring cross-border knowledge and networks that reduce friction. A fast-growing base of entrepreneurs can expand productive capacity and drive innovation. Latino communities often concentrate and connect these advantages\u2014from logistics and customs hubs in Laredo and El Paso to commercial and energy hubs in Miami and Houston\u2014and extend them digitally beyond any single geography.
\nThese factors converge to provide pathways for economic mobility in communities with sizable Latino populations in the United States. The number of Latino-owned businesses increased by 7 percent annually from 2017 to 2023, far outpacing the rest of the nation\u2019s near-flat growth. However, these businesses are underrepresented in some of the largest export sectors, accounting for 6 percent or less of manufacturing, mining, and IT firms. Instead of relying on organic opportunity creation as the only source of growth in trade, our analysis finds that a systematic approach to the development of the US\u2013Latin America trade corridor, with coordinated efforts across investors, businesses, trade ecosystem institutions, and governments, could increase trade volumes by 2.3 percent annually to $1.8 trillion by 2035, presenting a meaningful opportunity for US-based Latino consumers, businesses, and places.
\nYet while the advantages that US-based Latinos have through the trade corridor are real, they remain fragmented and underscaled, leaving limited pathways for stakeholders to holistically capture that value in each place. An integrated ecosystem and amplification engine for the US\u2013Latin America corridor could help create and capture this potential value by focusing on five key areas:
\nDesigning capital to enable scale. Expand access to fit-for-purpose capital, from working-capital loans, receivables financing, and trade finance to equity investment. Prior McKinsey analysis found that providing more business banking products and services for Latino business owners could mean closing a $200 billion business-lending gap. Extending capital from seed to scale-up will enable more firms to participate in trade, where longer cash cycles, currency risk, and high up-front investment are common challenges.
\nBuilding market access and distribution deliberately. Connect specialized hubs to Latino demand through multinational corporation (MNC) procurement, retailers, distributors, platforms, and import-export channels. Roughly 75 percent of US\u2013Latin America corridor trade is concentrated through a handful of major ports, while a long tail of more than 100 smaller ports handles the rest. But not all cities are equally connected to the opportunity this creates. Among markets with similarly sized Latino populations, cities connected to the most ports receive more than three times the consumer goods from Latin America relative to other cities, highlighting significant commercial opportunity for businesses involved in the corridor supply chain. These networks can also be the infrastructure through which new exporters emerge. A survey of small businesses from the US Export\u2013Import Bank (EXIM) found that while 53 percent of respondents would like to become exporters, 45 percent did not know where to start. As hubs become more connected, relationships between businesses can surface more knowledge about export opportunities and enable execution partnerships.
\nPrioritizing the coordination of institutions to lead change. Establish an accountable coordination layer (or corridor-wide institutions) that connects capital, distribution, and hub-level efforts. Support today could be scaled: The United States Hispanic Chamber of Commerce (USHCC) represents more than 260 chambers, while the National Minority Supplier Development Council (NMSDC) connects more than 15,000 minority-owned businesses through 23 regional councils. The Korea International Trade Association illustrates how this could work on a global scale by providing tailored support in target overseas markets for South Korean businesses in the form of advisory, market intelligence, customer matching, finance, training, and advocacy (see sidebar \u201cSouth Korea\u2019s example: Intentionally driving economic value\u201d). A US-based version of this could support businesses in the corridor to enter US Latino micromarkets with culturally distinct needs while also supporting Latino-owned businesses in the United States with export potential to enter international markets.
\nMaking commercialization an explicit objective. Build a repeatable pipeline that identifies promising firms, validates demand, develops products and talent, and prepares businesses for export and market entry. Latino-owned exporters show outsize productivity, yet only about 2 percent of Latino-owned employer firms export today.
\nProtecting ownership and reinvestment. Design mechanisms that support Latinos to participate more broadly in ownership as the corridor grows rather than taking income alone.
\nCountries around the world, including France, Japan, South Korea, and the United Kingdom, have connected these five elements to create broader economic value. This kind of integrated ecosystem and amplification engine could create similar economic opportunities for communities with significant US Latino populations by growing the number of successful corridor-related Latino businesses as well as creating significant ownership and employment opportunities.
\nUnderstanding a critical economic corridor
\nThe trade corridor between the United States and Latin America ranks among the six largest regional trade relationships globally. Almost one-fifth of total US trade activity flows through it, growing by an average of about 3 percent annually in real terms since 2015. While growth is projected to slow to 1.5 to 2.0 percent in the next decade, the corridor is still projected to be worth $1.7 trillion to $1.8 trillion by 2035. Additionally, potential headwinds and uncertainty exist around tariffs, inflation, and trade agreements, and scenarios with other challenges to trade could impact growth rates. But regardless of how these changes unfold, the US\u2013Latin America corridor is likely to continue to play a significant role in global economic activity, and other known structural shifts may alter where value comes from.
\nToday, around two-thirds of goods and services traded are concentrated in five categories: electronics, services, transport equipment, energy resources, and machinery (Exhibit 1). And while significant trade flows both ways, with around $779 billion heading north and $677 billion south, the linchpin is Mexico, which accounts for about 66 percent of US\u2013Latin America trade and nearly 80 percent of imported goods to the United States.
\nAn increasingly integrated economic relationship
\nThe production system between the United States and Latin America is also becoming increasingly integrated, with growing activity in higher-value sectors. While the corridor remains concentrated in the same major trading relationships, the roles different countries play within it are becoming more diversified. Between 2019 and 2025, for instance, Mexico added approximately $126 billion in electronics trade as well as growth in transport equipment, machinery, and medical and scientific instruments. That shift is continuing: Computer server exports from Mexico reached nearly $83 billion in the first half of 2026, with almost 94 percent of that destined for the United States.
\nThe corridor\u2019s integration is also deepening based on continued country-specific specialization. Costa Rica nearly doubled its bilateral trade with the United States from 2019 to 2025, driven in part by roughly $5 billion of additional medical and scientific instrument trade and $3 billion in electronics. Chile\u2019s and Peru\u2019s roles in metals, minerals, and energy value chains are expanding, as are Brazil\u2019s and Colombia\u2019s roles in higher-value services, including business, digital, financial, and intellectual property (IP)\u2013related services. Three factors are driving this:
\nLatin America\u2019s proximity, relatively lower vulnerability to transport disruption, time zone alignment, and supplier integration are powerful advantages as companies endeavor to secure supply chains and move production closer to customers in US markets to realize cost and operational efficiencies. Regional trade dynamics have recently reinforced the trend, with meaningful actions and announcements from leaders in Latin America, such as Mexico and Panama, focusing on the United States as a trade partner rather than other nations. This dynamic could create openings for US Latino-owned firms and workers across the value chain, from production and logistics to professional services and finance.
\nDigital connectivity is making more activity tradable across the corridor in both directions. Data costs fell by around 21 percent annually on major US\u2013Latin America routes from 2022 to 2025, fueling the expansion of cross-border delivery of software, fintech, media, professional services, data, and AI. Digitally deliverable services were worth about $85 billion in 2025, having increased by around 9 percent annually since 2023. Knowledge-based services alone could represent a $100 billion export opportunity. AI could accelerate this shift: A recent McKinsey analysis estimates that fully utilizing AI and automation could unlock up to $450 billion in annual economic value in Latin America by 2030 while helping the region gain efficiencies in the tradeable sectors where it leads and expand across the value chain toward digital and advanced manufactured goods, further strengthening and diversifying its economies.
\nCapital stock from foreign direct investment in the corridor reached $464 billion in 2025, with about 84 percent of that represented by investments into Latin America from the United States. However, US investment into Latin America has remained flat since 2020, while US companies building new businesses in Latin America, or greenfield investment, has fallen by about 7 percent annually (Exhibit 2). During the same period, global greenfield investment (excluding the United States) into Latin America nearly tripled from $45 billion to $133 billion, suggesting there could be further potential for the United States to position itself to capture the corridor\u2019s next wave of growth.
\nThese three shifts are creating new corridor opportunities, though who captures them will depend heavily on geography, sector, and existing market access. Latino-owned businesses and communities with distinctive positional advantages and large Latino populations (those in trade hubs, in high-growth sectors, or with established cross-border networks) are best placed to benefit. An emerging ecosystem is already turning those specific advantages into repeatable plays that scale value across markets.
\nThe importance of US Latinos in creating value from trade
\nFive distinctive advantages across consumers, businesses, and places highlight how Latinos today create outsize value in the corridor. First, Latino consumers provide a large and growing base of demand that companies are increasingly accounting for in their strategic decision-making. Second, Latinos wield a variety of unique cultural identities with influence that can extend beyond Latino markets. Third, Latino businesses can bring cross-border knowledge and networks that reduce friction and enable success in the corridor. Fourth, a fast-growing base of entrepreneurs and workers can expand productive capacity and unlock innovation. Finally, physical and digital integration in specific places, such as Latino-population hubs and ports of entry, concentrates and connects these advantages, enabling new business models and extending their reach across the corridor.
\nConsumers: Distinctive demand and cultural influence create a launchpad for growth
\nLarge and rapidly growing Latino-linked demand can provide a promising base market for businesses operating across the corridor. In broad terms, around 68 million US Latinos today represent $2.8 trillion in household consumption, which grew three times faster than non-Latino consumption from 2019 to 2024. Latino-linked demand already accounts for about 15 percent of Latin American final-goods imports, up from 12 percent a decade ago and above Latinos\u2019 share of US household income. As an example of how Latino-linked demand can enable mainstream adoption, Modelo Especial\u2019s US off-premises retail sales grew from around $2 billion in 2018 to more than $4 billion in 2024, while the share of the brand\u2019s business coming from non-Latino consumers rose from 30 percent to 45 percent. It is now the top-selling beer in the United States by dollar sales.
\nLatino cultural influence also extends beyond Latino consumers, creating a path to scale promising business models for products that are otherwise new to non-Latino consumers. For example, more than half of US music listeners engage with Latin music. Its share of US recorded music revenue has doubled since 2020 to roughly 9 percent of the market. Similar dynamics are seen across industries. Modelo and Taj\u00edn have become mainstream consumer brands, while beauty players such as Ceremonia and Nopalera are translating Latino heritage into differentiated propositions. In fashion, reggaeton superstar Bad Bunny\u2019s Adidas collaborations and designers such as Willy Chavarria have brought Latino cultural references to global audiences, while sports and travel\u2014from Inter Miami to Puerto Rico tourism around major cultural events\u2014show how cultural relevance can translate into broader spending and economic activity. These are some of the early examples of how business leaders are starting to realize the significant pool of value related to sports, entertainment, fashion, beauty, and other sectors within the Latino community.
\nBusinesses: Cross-border capabilities and a growing productive base serve and scale demand
\nKnowledge and networks can reduce cross-border friction, lowering the search and transaction costs of connecting customers, suppliers, partners, and institutions across markets. When paired with strong operations, they can turn cross-border familiarity and connectivity into commercial advantage. This emanates from deep interpersonal relationships that connect the United States and Latin American markets even outside the arena of trade. For instance, phone traffic between the United States and Mexico represents the world\u2019s largest bilateral route, accounting for about 8 percent of global international call minutes. Such interconnectivity has a tangible business impact: Latino-owned exporters generate around $325,000 in revenue per employee, about 3.5 times that of Latino-owned nonexporters (Exhibit 3).
\nLatino business and labor make up a prominent proportion of total productive capacity in the United States. Per the US Bureau of Labor Statistics, Latinos now make up roughly one in five US workers and have accounted for 70 percent of US workforce growth since 2020 (Exhibit 4). And according to the US Census Bureau American Community Survey 2024 release, Latinos are increasingly skilled: For example, the percentage of foreign-born Latinos who have been in the United States five years or less with a bachelor\u2019s degree or higher increased from 14 percent in 2010 to 26 percent in 2024, while US-born Latinos with a bachelor\u2019s degree or higher similarly increased from 17 percent to more than 26 percent during the same period.
\nThe number of Latino-owned firms with employees grew by about 6.8 percent annually from 2017 to 2023, compared with growth of less than 1.0 percent for non-Latino-owned firms. In addition, Latino-owned employer firms generated around 7 percent of net new jobs in the United States from 2017 to 2022 and, according to Stanford University, are more represented in emerging technologies such as blockchain and Web3 compared with White-owned businesses at 5 percent versus 1 percent, respectively (representation in AI sectors is comparable). Taken together, a larger, more skilled, and more entrepreneurial Latino base of workers and owners could translate to a latent resource for the corridor that increases both supply and demand for its commercial offerings.
\nPlaces: Physical concentration and digital integration multiply Latino advantages from trade
\nPhysical and digital integration concentrates the advantages of Latino businesses and consumers and connects them across markets. Trade capabilities are highly concentrated geographically, with eight leading US\u2013Latin America trade hubs accounting for roughly three-quarters of corridor imports. These hubs bring together specialized infrastructure, industries, networks, and demand, including logistics and customs capabilities in Laredo and El Paso, energy expertise in Houston, and commercial connections in Miami. Beyond ports of entry, Latino population hubs also create downstream opportunities born out of the corridor by concentrating demand and opening business opportunities around traded goods. However, direct connectivity to the corridor still matters for both consumers and businesses: Among places with similarly sized Latino populations, stronger connections to ports importing from Latin America are associated with substantially greater availability of Latin American consumer goods.
\nDigital connectivity can extend the same advantage beyond a single geography. Recurring cross-border interactions include scalable services, proprietary data, platforms, and IP. Businesses are finding different ways to translate cultural knowledge, institutional access, and cross-border data into business models that serve customers across multiple countries, such as the following:
\nF\u00e9lix Pago, a fintech start-up, facilitates US\u2013Latin America remittances through a fully digital WhatsApp-native interface.
\nDesteia, which has raised seed funding of $11.5 million as of 2025, integrates with Mexico\u2019s customs filing portal to simplify compliance for MNCs.
\nSonoro, which raised $12.5 million in Series A funding, has built one of the world\u2019s largest podcast networks by creating culturally relevant audio content for audiences across the continents.
\nBy creating ways to serve long-standing needs of Latino customers across geographies, digital integration converts the advantages of Latino consumers and businesses into market-defining plays.
\nFrom distinctive advantages to proven plays
\nLatino consumer, business, and place-based advantages are already translating into action. Across the US\u2013Latin America corridor, companies, entrepreneurs, investors, governments, and ecosystem institutions are using them in a set of recurring plays that create and capture value, showing how Latino-linked demand and capabilities can successfully create opportunities across markets.
\nMNCs are importing and commercializing Latin American products in the United States, often using existing Latino demand as an initial market before scaling more broadly. Modelo illustrates the potential, scaling from a heavily Latino consumer base to become the number one beer by dollar sales in the United States. MNCs are also investing to expand across Latin America, bringing capital, technology, infrastructure, procurement, and operating capabilities. This trend is particularly pronounced in the technology sector: For example, Microsoft invested $2.7 billion into cloud and AI infrastructure in Brazil, and Amazon announced a $5 billion cloud-region investment in Quer\u00e9taro, Mexico.
\nUS Latino-owned businesses and entrepreneurs are pursuing a broad set of cross-border models. Across the models, each uses their unique knowledge of the Latino consumer and preexisting networks to carve out a competitive advantage. Some import and commercialize culturally connected goods, using their knowledge of Latino demand and sourcing networks to build scaled businesses. Others export brands and business models into Latin America, leveraging propositions first developed in the United States as well as cross-border market knowledge and owned assets. Nopalera, for example, which has an established footprint across more than 700 retail points in the United States, recently partnered with Ulta Beauty to launch in Mexico and leveraged a $4 million Series A funding to further fund its domestic and international expansion (see sidebar \u201cUS Latino-owned business spotlight: Nopalera\u201d). Others are building digital services that solve recurring cross-border frictions: F\u00e9lix Pago processed about $1 billion in remittances in 2024, Nuvocargo reports customs clearance times that are up to 34 percent faster, and Nowports reports shipment delays that are up to 40 percent lower through technology-enabled logistics and customs capabilities.
\nPlaces are developing plays of their own. State and local governments could build infrastructure and specialized capabilities to help make trade easier or focus on particular sectors rather than trying to replicate a full gateway. Laredo and El Paso illustrate deep logistics and customs ecosystems, for example, while Tampa shows how a secondary hub can build differentiated capabilities in cold-chain and food distribution. Specialized connectivity can materially expand access to corridor trade. Among cities with similarly sized Latino populations, stronger links to Latin American entry ports are associated with around three to four times greater availability of imported goods, and cities such as San Antonio illustrate how logistics and sector specialization can turn that connectivity into economic activity.
\nCapital providers and ecosystem institutions play a complementary role by helping promising corridor businesses access the capital, customers, and capabilities needed to scale. Capital providers offer fit-for-purpose capital, including growth equity, working capital, and trade finance. For example, L\u2019Attitude Ventures built a $100 million venture capital platform for early-stage investments in US Latino-owned businesses (including Nopalera), while fintech platform and marketplace Mercado Libre uses proprietary sales and payment data to rapidly assess creditworthiness and provides working-capital financing to merchants that are underserved by traditional lenders.
\nEcosystem institutions such as Latino chambers and supplier networks connect firms to customers, distributors, capital, specialized hubs, and country-specific expertise that would be difficult for any business to assemble alone. One example is ConnectAmericas, an Inter-American Development Bank (IDB)\u2013created platform enabling Latin American businesses to achieve 17 percent higher average export value by connecting them to a network of about 240,000 companies with capital opportunities, buyers, suppliers, and training.
\nTaken together, these initiatives show the corridor already contains many of the ingredients needed for broader value creation. The challenge is not the absence of successful models but rather that they remain fragmented and are not yet connected to an integrated system that can scale them consistently.
\nDriving greater US\u2013Latin America economic value in the corridor
\nBroad and largely external forces such as global policy, cross-border investment, and commodity trade cycles drive much of what forecasts anticipate for the corridor. Based on these trends, a McKinsey Global Institute analysis implies a change in growth rate from 3.0 percent in real terms in the past decade to around 1.5 to 2.0 percent per year through 2035, translating to $1.2 trillion to $1.7 trillion in cumulative added trade volume, with the current balance between northbound and southbound holding roughly stable as a share of corridor trade.
\nYet there is a potentially sizable additional opportunity centered around the distinctive capabilities of US Latinos. Despite these advantages across businesses, consumers, and places and proven value creation plays, the ecosystem needed to capture the value remains fragmented. Capital, market access, institutional support, commercialization pathways, and ownership mechanisms are frequently unavailable or too dispersed for the companies and places best positioned to use them. We estimate that systematically addressing these barriers that prevent proven models from scaling could add $170 billion to $295 billion in incremental trade through the corridor in the next ten years, expanding the corridor\u2019s growth by 18 percent. Taken together, the corridor\u2019s momentum and this added opportunity represent up to $2 trillion in cumulative added trade over the next decade. The benefits would extend beyond the businesses that gain new opportunities to import and export. There could also be sizable spillover effects into the broader economy that create jobs, lower prices, and generate more choice for consumers and more vibrant places.
\nBut the additional value is not guaranteed. Capturing it depends on addressing the five system constraints.
\nLimited access to capital
\nSurveys of Latino entrepreneurs have documented that more than 70 percent of Latino-business owners used their personal funds to start their businesses, and as few as 12 percent use banks or financial institutions ; Latino-owned firms are also less likely to receive the full financing they request compared with other groups. At the same time, firms based in Latin America that would otherwise be exporting or importing are unable to access finance to engage in trade. It is estimated that expanding access to trade finance (for example, through working-capital loans, receivables financing, letters of credit, and supplier finance) could increase trade volumes by 6 to 9 percent across Mexico, Guatemala, and Honduras, all of which count the United States among their largest trading partners. Improving access to different capital products at different stages of growth on both sides of the corridor could enable more businesses with promising business models to capture value from trade, especially in the critical transition when founder-funded firms seek institutional scale.
\nLatino-owned businesses have long shown their ability to effectively use capital to expand into international trade and business activity. By the late 2000s, Brightstar, founded in Miami by Marcelo Claure, became the largest Latino-owned business in the United States by building a scaled mobile phone exporting business that grew revenue from $140 million in 1999 to more than $1 billion by 2003. The key to its early expansion was using working-capital financing to build inventory that could be sent to Latin America. Today\u2019s businesses are using newer financial mechanisms to reach customers across markets. Miami-based tech start-up F\u00e9lix Pago, for example, has raised multiple rounds of venture funding to expand its remittance platform across multiple Latin American countries. Since its founding in 2020, it has served six million customers across 11 Latin American markets and processed $8 billion in transactions.
\nNarrow market and distribution connectivity
\nAbout 75 percent of trade in the corridor is concentrated in leading hubs that reach only around 16 percent of US Latinos. And further dispersion of traded goods from ports of entry to consumers in other Latino population hubs is uneven. Some of the largest Latino markets in the United States remain relatively disconnected from the trade infrastructure and distribution networks that would allow demand to translate more fully into corridor flows (Exhibit 5). This finding is consistent with McKinsey\u2019s The economic state of Latinos in America report, which documented that 41.5 million Latinos live in consumer deserts and are 1.3 times as likely as White populations to have low access to consumer goods.
\nThis highlights meaningful opportunities for greater consumer import penetration. For example, Los Angeles and Las Vegas have sizable Latino populations but a comparatively low level of Latin American imports, suggesting room to expand sourcing and distribution, particularly given their proximity to Mexico and cross-border logistics advantages. Companies may consider whether cost, lack of knowledge about Latin American imports, access, or other factors are barriers for consumers. Other markets show what stronger connectivity may unlock for cities that are not major ports of entry themselves: Tampa has built a differentiated position in food distribution, for instance, while Dallas functions as a major logistics hub that supports Latin American imports well beyond what local demographics alone would suggest.
\nMore broadly, among places with similarly sized Latino populations, stronger connections to Latin American entry ports are associated with materially greater availability of Latin American consumer goods. Latino businesses need consistent pathways to access MNC procurement, retailers, distributors, platforms, corporate customers, and export channels. Without them, commercially viable businesses tend to stay local or niche. Broader market access and distribution connectivity may unlock significant economic value by converting demand into repeat purchases, larger orders, and national or regional reach.
\nSource Logistics, founded by Latino entrepreneurs, highlights this opportunity. It has made growing coverage of the US Latino population its core growth strategy to meet rising demand for Latin American goods, especially food and beverage. Through M&A and new warehouse builds, it has expanded to 5.8 million square feet of warehousing nationwide, reaching 80 percent of the US Latino population and 75 percent of the total US population with same-day delivery; its longer-term goal is to achieve 95 percent total US population coverage. This growth helped it rank on the Inc. 5000 2025 list of America\u2019s fastest-growing private companies for the second consecutive year.
\nAs connectivity improves between trade hubs and other cities, opportunities will emerge for businesses with rights to win as exporters. Quirch Foods, which started as an importer of consumer goods from Latin America, built distribution networks across the United States from a central distribution node in Miami. While 80 to 90 percent of its business remains domestic, over time, its import business led to relationships with domestic meat and seafood producers that it combined with its cross-border networks and infrastructure assets in Miami to expand into exports. Now Quirch ships more than 100 containers a week globally, with a focus on Latin America. Other successful export businesses in the corridor highlight the value stronger networks can create. For example, SolvChem, a Latino-owned chemicals distributor based in Houston, created an international division in 2009 to export chemicals to Latin America. Having already built chemicals handling infrastructure in Houston and a regional network of suppliers, it used its proximity to a key corridor port to become a dominant player in chemicals exports, generating more than $100 million in revenue per year.
\nLow institutional coordination
\nDespite an extensive support ecosystem, Latino trade participation remains low. For example, the USHCC represents more than 260 chambers, and the NMSDC connects more than 15,000 minority-owned businesses through 23 regional councils, but only 2.1 percent of Latino-owned employer firms export, compared with 2.6 percent for all US employer firms. An accountable corridor-wide coordination layer could connect commercial development, capital, distribution, specialized hubs, country-specific regulation, market knowledge, and skills, surfacing promising opportunities that would otherwise be missed.
\nThe United States also lacks a coordinated and scaled set of efforts to amplify the commercial power of Latino cultural assets. Such vehicles have been deployed globally to great effect, such as Enterprise Ireland, which holds capital, capability building, and market access under one mandate: The same agency that funds a company also opens markets for it with its 42 overseas offices so the firm never has to assemble the system itself. In 2025, its client companies had record export sales of \u20ac38.86 billion and employed 232,425 people. The United States has already had major success promoting commercialization of other cultural assets at home and abroad, such as through its film industry, which has an estimated global annual trade surplus of $15 billion. A coordinated group of stakeholders in the United States, such as investors, trade associations, or nonprofits, could focus on identifying emerging trends with commercial potential from within the Latino community and support them to develop their business models, leveraging import and export opportunities.
\nBrazilian influencer Camila Coelho illustrates what\u2019s possible when entrepreneurs can scale products in the corridor. She leveraged her more than ten million social media followers to launch two companies generating more than $10 million in combined sales by 2021: a fashion label with Revolve and a beauty line built around ingredients native to Brazil reaching more than 400 Ulta Beauty stores and shipped to 17 countries. She has used the corridor by turning Brazilian ingredients into US-launched products sold back into Latin American markets. With an enabled ecosystem, entrepreneurs like Coelho could establish a new age of US cosmetics and fashion rooted in the US Latino community.
\nNascent commercialization pipeline
\nThe opportunity for Latino firms\u2019 broader participation in the corridor is not constrained by a lack of demand or business performance but by an uneven ability to commercialize and scale. The US\u2013Latin America corridor lacks a pipeline that identifies promising firms and assets, validates demand, develops products and talent, builds export or market entry readiness, protects IP, codifies playbooks, and replicates what works. Having repeatable capabilities, playbooks, and intermediaries that identify promising opportunities and help move businesses from proof of demand to scaled market participation could unlock substantial value for Latino-owned businesses. Other economies have built central agencies or intermediary institutions to perform this pipeline-building function. For example, Business Sweden, a collaboration between the Swedish government and Swedish Foreign Trade Association, institutionalized the concept of getting proven firms export-ready and has helped thousands of Swedish companies enter international markets, including prominent and fast-growing businesses, such as the electric and autonomous freight company Einride, which went public earlier this year.
\nThe missed commercialization opportunity is especially stark for Latino cultural assets, which are unlocked by trade but have potential for economic spillover effects in the US economy. Bad Bunny offers one example: The singer and rapper\u2019s rise was enabled in part by Puerto Rico\u2013based Rimas Entertainment, which secured an early YouTube monetization agreement that helped Bad Bunny reach audiences without relying on a traditional major-label model. That model has translated cultural demand into broader economic activity. During Bad Bunny\u2019s 2025 Puerto Rico residency, total spending in San Juan rose nearly 20 percent versus comparable weekends a year earlier, while international spending around the venue increased by more than 15 percent. The residency also connected visitors with local businesses and helped expand Puerto Rico\u2019s reach into new tourism markets, a tangible display of the economic power of culture that went global just months later with Bad Bunny\u2019s performance at the Super Bowl LX halftime show.
\nIf the United States could build a similar coordination and amplification apparatus to the same results that South Korea achieved from its Hallyu cultural engine, the economic opportunity in the United States could be worth more than $300 billion over ten years, with trade-related activity adding another $60 billion.
\nWeak ownership and reinvestment mechanisms
\nEconomic growth in the US\u2013Latin America corridor can generate income without creating durable, Latino-owned assets that let value compound over time, such as equity, brands, proprietary data, platforms, and distribution networks. Weak mechanisms for ownership and reinvestment mean founders, creators, and communities often don\u2019t retain a lasting stake in the value they help create, which is evident at both the household and the firm level.
\nAt the household level, Latino households that own a business hold roughly 2.2 times the net worth of those that do not, yet only 6 percent of Latino families receive an inheritance, part of a $380 billion wealth transfer gap that McKinsey documented.
\nAt the firm level, Latino-owned firms capture less of their industries\u2019 value than peers, generating less than half as much revenue per firm as other firms (excluding MNCs). This suggests a scale gap that compounds the ownership gap.
\nLatino firms in the corridor show how scaled ownership can create significant value. Goya Foods, for example, highlights the payoff of retained ownership and reinvestment. It has grown into a multibillion multinational food distributor while retaining family ownership despite multiple buyout attempts. Its former COO went on to found AUA Equity, which has an investment strategy built around its distinctive knowledge of scaling products nationally from the Latino business ecosystem.
\nCreating and capturing greater value from the US\u2013Latin America corridor
\nA more integrated US\u2013Latin America ecosystem holds the promise of generating significant economic value, especially for US Latinos\u2014although it will not materialize automatically. Creating and capturing this value requires investment to build an integrated ecosystem and amplification engine for the corridor that eases entry for new participants while helping existing players scale. Such a system would allow US Latinos to apply their distinctive capabilities to proven plays to capture more value while creating spillover effects that benefit communities. For stakeholders already engaged with the corridor, the ecosystem\u2019s amplification engine can help them succeed more often and scale faster by connecting them to system support they may be missing. As more plays succeed, assets are created along with knowledge, relationships, data, brands, and capital that strengthen the next cycle of growth in the corridor. New participants will also benefit by having better access to the existing ecosystem and access to the market entry playbooks that have been codified.
\nRealizing these opportunities in the corridor could have substantial impact on the economic mobility potential for Latinos in the United States. As Latino-owned businesses succeed and grow more frequently, the share of wealth that the Latino population claims through ownership increases. When stronger networks enable better exchanges between businesses about the proven plays, more businesses can enter adjacencies that have higher potential for export, a proven way to increase value creation. As businesses in the corridor grow, they create employment opportunities for other Latinos in their communities, which in turn enables skill development and opens career pathways for wealth creation. Finally, corridor participants reinvest resources into their communities through direct spending and through tax revenue generation, creating a financial base to improve quality of life through services and amenities.
\nInvestors, businesses, governments, and mission-driven organizations looking to participate in the corridor to access these opportunities don\u2019t need to build every capability themselves. Key stakeholders need to proactively collaborate to make this integrated ecosystem and amplification engine a reality. Each actor will be responsible for building distinct but complementary and coordinated levers that can work together to open the pipeline of opportunities. A summary of each stakeholder, how they benefit from the ecosystem, and steps they could take is in the table below.
\nThe trade corridor between the United States and Latin America is big, growing, and critical to the economic dynamism of the region\u2014but there are pockets of opportunity that have yet to be realized. The path is not without risk. Headwinds and uncertainty could impact how these opportunities can be realized. Nonetheless, commercial activity in the corridor will continue, and many of the ingredients for generating greater value are already in place. An integrated system that consistently scales distinctive Latino capabilities could turn this potential into greater economic momentum, providing another pathway for economic mobility for Latinos and the communities they live in nationwide.
", "summary": "Quirch Foods began in Miami in 1967 as a family-owned importer of Latino staple foods and today generates more than $4 billion in annual sales across 40 US states, the Caribbean, and Latin America. Latina-owned Nopalera built a Mexican-inspired beauty brand into hundreds of US retail stores before expanding into Mexico. Angel network Angeles Investors co-led a $32 million Series A investment in Canela Media, aiding its growth to more than 60 million active users across the United States and Latin America. The city of Laredo, with almost 275,000 residents and 36 percent of its nonfarm workforce in trade, has developed distinctive capabilities in logistics, warehousing, and customs brokerage, enabling it to handle $354 billion in international trade in 2025.
\nThese are just a few examples of the importers, exporters, investors, and specialized locations that fuel the US\u2013Latin America trade corridor. It is one of the world\u2019s biggest and most critical economic arteries, generating $1.5 trillion in trade in 2025 \u2014about 1.5 times the total trade between the United States and Canada and the equivalent of 85 percent of all trade within the European Union. Discussion of the corridor tends to center on large forces that shape it, such as trade agreements and tariffs, geopolitics, commodity cycles, inflation, or immigration. These forces help set the baseline of how the corridor may change in overall size. However, analysis of these factors does not fully explain how US companies continue to create enduring value within the corridor in new ways.
\nThe success of companies such as Quirch Foods, Nopalera, and Canela Media shows how US-based Latino consumers, businesses, and places can turn positional and proprietary advantages into value that flows back to communities.
\nLatino consumers in the United States, a population spanning distinct national-origin groups, generations, and cultures, can provide a large and growing base of demand and cultural influence able to extend to other consumer segments. Latino businesses can bring cross-border knowledge and networks that reduce friction. A fast-growing base of entrepreneurs can expand productive capacity and drive innovation. Latino communities often concentrate and connect these advantages\u2014from logistics and customs hubs in Laredo and El Paso to commercial and energy hubs in Miami and Houston\u2014and extend them digitally beyond any single geography.
\nThese factors converge to provide pathways for economic mobility in communities with sizable Latino populations in the United States. The number of Latino-owned businesses increased by 7 percent annually from 2017 to 2023, far outpacing the rest of the nation\u2019s near-flat growth. However, these businesses are underrepresented in some of the largest export sectors, accounting for 6 percent or less of manufacturing, mining, and IT firms. Instead of relying on organic opportunity creation as the only source of growth in trade, our analysis finds that a systematic approach to the development of the US\u2013Latin America trade corridor, with coordinated efforts across investors, businesses, trade ecosystem institutions, and governments, could increase trade volumes by 2.3 percent annually to $1.8 trillion by 2035, presenting a meaningful opportunity for US-based Latino consumers, businesses, and places.
\nYet while the advantages that US-based Latinos have through the trade corridor are real, they remain fragmented and underscaled, leaving limited pathways for stakeholders to holistically capture that value in each place. An integrated ecosystem and amplification engine for the US\u2013Latin America corridor could help create and capture this potential value by focusing on five key areas:
\nDesigning capital to enable scale. Expand access to fit-for-purpose capital, from working-capital loans, receivables financing, and trade finance to equity investment. Prior McKinsey analysis found that providing more business banking products and services for Latino business owners could mean closing a $200 billion business-lending gap. Extending capital from seed to scale-up will enable more firms to participate in trade, where longer cash cycles, currency risk, and high up-front investment are common challenges.
\nBuilding market access and distribution deliberately. Connect specialized hubs to Latino demand through multinational corporation (MNC) procurement, retailers, distributors, platforms, and import-export channels. Roughly 75 percent of US\u2013Latin America corridor trade is concentrated through a handful of major ports, while a long tail of more than 100 smaller ports handles the rest. But not all cities are equally connected to the opportunity this creates. Among markets with similarly sized Latino populations, cities connected to the most ports receive more than three times the consumer goods from Latin America relative to other cities, highlighting significant commercial opportunity for businesses involved in the corridor supply chain. These networks can also be the infrastructure through which new exporters emerge. A survey of small businesses from the US Export\u2013Import Bank (EXIM) found that while 53 percent of respondents would like to become exporters, 45 percent did not know where to start. As hubs become more connected, relationships between businesses can surface more knowledge about export opportunities and enable execution partnerships.
\nPrioritizing the coordination of institutions to lead change. Establish an accountable coordination layer (or corridor-wide institutions) that connects capital, distribution, and hub-level efforts. Support today could be scaled: The United States Hispanic Chamber of Commerce (USHCC) represents more than 260 chambers, while the National Minority Supplier Development Council (NMSDC) connects more than 15,000 minority-owned businesses through 23 regional councils. The Korea International Trade Association illustrates how this could work on a global scale by providing tailored support in target overseas markets for South Korean businesses in the form of advisory, market intelligence, customer matching, finance, training, and advocacy (see sidebar \u201cSouth Korea\u2019s example: Intentionally driving economic value\u201d). A US-based version of this could support businesses in the corridor to enter US Latino micromarkets with culturally distinct needs while also supporting Latino-owned businesses in the United States with export potential to enter international markets.
\nMaking commercialization an explicit objective. Build a repeatable pipeline that identifies promising firms, validates demand, develops products and talent, and prepares businesses for export and market entry. Latino-owned exporters show outsize productivity, yet only about 2 percent of Latino-owned employer firms export today.
\nProtecting ownership and reinvestment. Design mechanisms that support Latinos to participate more broadly in ownership as the corridor grows rather than taking income alone.
\nCountries around the world, including France, Japan, South Korea, and the United Kingdom, have connected these five elements to create broader economic value. This kind of integrated ecosystem and amplification engine could create similar economic opportunities for communities with significant US Latino populations by growing the number of successful corridor-related Latino businesses as well as creating significant ownership and employment opportunities.
\nUnderstanding a critical economic corridor
\nThe trade corridor between the United States and Latin America ranks among the six largest regional trade relationships globally. Almost one-fifth of total US trade activity flows through it, growing by an average of about 3 percent annually in real terms since 2015. While growth is projected to slow to 1.5 to 2.0 percent in the next decade, the corridor is still projected to be worth $1.7 trillion to $1.8 trillion by 2035. Additionally, potential headwinds and uncertainty exist around tariffs, inflation, and trade agreements, and scenarios with other challenges to trade could impact growth rates. But regardless of how these changes unfold, the US\u2013Latin America corridor is likely to continue to play a significant role in global economic activity, and other known structural shifts may alter where value comes from.
\nToday, around two-thirds of goods and services traded are concentrated in five categories: electronics, services, transport equipment, energy resources, and machinery (Exhibit 1). And while significant trade flows both ways, with around $779 billion heading north and $677 billion south, the linchpin is Mexico, which accounts for about 66 percent of US\u2013Latin America trade and nearly 80 percent of imported goods to the United States.
\nAn increasingly integrated economic relationship
\nThe production system between the United States and Latin America is also becoming increasingly integrated, with growing activity in higher-value sectors. While the corridor remains concentrated in the same major trading relationships, the roles different countries play within it are becoming more diversified. Between 2019 and 2025, for instance, Mexico added approximately $126 billion in electronics trade as well as growth in transport equipment, machinery, and medical and scientific instruments. That shift is continuing: Computer server exports from Mexico reached nearly $83 billion in the first half of 2026, with almost 94 percent of that destined for the United States.
\nThe corridor\u2019s integration is also deepening based on continued country-specific specialization. Costa Rica nearly doubled its bilateral trade with the United States from 2019 to 2025, driven in part by roughly $5 billion of additional medical and scientific instrument trade and $3 billion in electronics. Chile\u2019s and Peru\u2019s roles in metals, minerals, and energy value chains are expanding, as are Brazil\u2019s and Colombia\u2019s roles in higher-value services, including business, digital, financial, and intellectual property (IP)\u2013related services. Three factors are driving this:
\nLatin America\u2019s proximity, relatively lower vulnerability to transport disruption, time zone alignment, and supplier integration are powerful advantages as companies endeavor to secure supply chains and move production closer to customers in US markets to realize cost and operational efficiencies. Regional trade dynamics have recently reinforced the trend, with meaningful actions and announcements from leaders in Latin America, such as Mexico and Panama, focusing on the United States as a trade partner rather than other nations. This dynamic could create openings for US Latino-owned firms and workers across the value chain, from production and logistics to professional services and finance.
\nDigital connectivity is making more activity tradable across the corridor in both directions. Data costs fell by around 21 percent annually on major US\u2013Latin America routes from 2022 to 2025, fueling the expansion of cross-border delivery of software, fintech, media, professional services, data, and AI. Digitally deliverable services were worth about $85 billion in 2025, having increased by around 9 percent annually since 2023. Knowledge-based services alone could represent a $100 billion export opportunity. AI could accelerate this shift: A recent McKinsey analysis estimates that fully utilizing AI and automation could unlock up to $450 billion in annual economic value in Latin America by 2030 while helping the region gain efficiencies in the tradeable sectors where it leads and expand across the value chain toward digital and advanced manufactured goods, further strengthening and diversifying its economies.
\nCapital stock from foreign direct investment in the corridor reached $464 billion in 2025, with about 84 percent of that represented by investments into Latin America from the United States. However, US investment into Latin America has remained flat since 2020, while US companies building new businesses in Latin America, or greenfield investment, has fallen by about 7 percent annually (Exhibit 2). During the same period, global greenfield investment (excluding the United States) into Latin America nearly tripled from $45 billion to $133 billion, suggesting there could be further potential for the United States to position itself to capture the corridor\u2019s next wave of growth.
\nThese three shifts are creating new corridor opportunities, though who captures them will depend heavily on geography, sector, and existing market access. Latino-owned businesses and communities with distinctive positional advantages and large Latino populations (those in trade hubs, in high-growth sectors, or with established cross-border networks) are best placed to benefit. An emerging ecosystem is already turning those specific advantages into repeatable plays that scale value across markets.
\nThe importance of US Latinos in creating value from trade
\nFive distinctive advantages across consumers, businesses, and places highlight how Latinos today create outsize value in the corridor. First, Latino consumers provide a large and growing base of demand that companies are increasingly accounting for in their strategic decision-making. Second, Latinos wield a variety of unique cultural identities with influence that can extend beyond Latino markets. Third, Latino businesses can bring cross-border knowledge and networks that reduce friction and enable success in the corridor. Fourth, a fast-growing base of entrepreneurs and workers can expand productive capacity and unlock innovation. Finally, physical and digital integration in specific places, such as Latino-population hubs and ports of entry, concentrates and connects these advantages, enabling new business models and extending their reach across the corridor.
\nConsumers: Distinctive demand and cultural influence create a launchpad for growth
\nLarge and rapidly growing Latino-linked demand can provide a promising base market for businesses operating across the corridor. In broad terms, around 68 million US Latinos today represent $2.8 trillion in household consumption, which grew three times faster than non-Latino consumption from 2019 to 2024. Latino-linked demand already accounts for about 15 percent of Latin American final-goods imports, up from 12 percent a decade ago and above Latinos\u2019 share of US household income. As an example of how Latino-linked demand can enable mainstream adoption, Modelo Especial\u2019s US off-premises retail sales grew from around $2 billion in 2018 to more than $4 billion in 2024, while the share of the brand\u2019s business coming from non-Latino consumers rose from 30 percent to 45 percent. It is now the top-selling beer in the United States by dollar sales.
\nLatino cultural influence also extends beyond Latino consumers, creating a path to scale promising business models for products that are otherwise new to non-Latino consumers. For example, more than half of US music listeners engage with Latin music. Its share of US recorded music revenue has doubled since 2020 to roughly 9 percent of the market. Similar dynamics are seen across industries. Modelo and Taj\u00edn have become mainstream consumer brands, while beauty players such as Ceremonia and Nopalera are translating Latino heritage into differentiated propositions. In fashion, reggaeton superstar Bad Bunny\u2019s Adidas collaborations and designers such as Willy Chavarria have brought Latino cultural references to global audiences, while sports and travel\u2014from Inter Miami to Puerto Rico tourism around major cultural events\u2014show how cultural relevance can translate into broader spending and economic activity. These are some of the early examples of how business leaders are starting to realize the significant pool of value related to sports, entertainment, fashion, beauty, and other sectors within the Latino community.
\nBusinesses: Cross-border capabilities and a growing productive base serve and scale demand
\nKnowledge and networks can reduce cross-border friction, lowering the search and transaction costs of connecting customers, suppliers, partners, and institutions across markets. When paired with strong operations, they can turn cross-border familiarity and connectivity into commercial advantage. This emanates from deep interpersonal relationships that connect the United States and Latin American markets even outside the arena of trade. For instance, phone traffic between the United States and Mexico represents the world\u2019s largest bilateral route, accounting for about 8 percent of global international call minutes. Such interconnectivity has a tangible business impact: Latino-owned exporters generate around $325,000 in revenue per employee, about 3.5 times that of Latino-owned nonexporters (Exhibit 3).
\nLatino business and labor make up a prominent proportion of total productive capacity in the United States. Per the US Bureau of Labor Statistics, Latinos now make up roughly one in five US workers and have accounted for 70 percent of US workforce growth since 2020 (Exhibit 4). And according to the US Census Bureau American Community Survey 2024 release, Latinos are increasingly skilled: For example, the percentage of foreign-born Latinos who have been in the United States five years or less with a bachelor\u2019s degree or higher increased from 14 percent in 2010 to 26 percent in 2024, while US-born Latinos with a bachelor\u2019s degree or higher similarly increased from 17 percent to more than 26 percent during the same period.
\nThe number of Latino-owned firms with employees grew by about 6.8 percent annually from 2017 to 2023, compared with growth of less than 1.0 percent for non-Latino-owned firms. In addition, Latino-owned employer firms generated around 7 percent of net new jobs in the United States from 2017 to 2022 and, according to Stanford University, are more represented in emerging technologies such as blockchain and Web3 compared with White-owned businesses at 5 percent versus 1 percent, respectively (representation in AI sectors is comparable). Taken together, a larger, more skilled, and more entrepreneurial Latino base of workers and owners could translate to a latent resource for the corridor that increases both supply and demand for its commercial offerings.
\nPlaces: Physical concentration and digital integration multiply Latino advantages from trade
\nPhysical and digital integration concentrates the advantages of Latino businesses and consumers and connects them across markets. Trade capabilities are highly concentrated geographically, with eight leading US\u2013Latin America trade hubs accounting for roughly three-quarters of corridor imports. These hubs bring together specialized infrastructure, industries, networks, and demand, including logistics and customs capabilities in Laredo and El Paso, energy expertise in Houston, and commercial connections in Miami. Beyond ports of entry, Latino population hubs also create downstream opportunities born out of the corridor by concentrating demand and opening business opportunities around traded goods. However, direct connectivity to the corridor still matters for both consumers and businesses: Among places with similarly sized Latino populations, stronger connections to ports importing from Latin America are associated with substantially greater availability of Latin American consumer goods.
\nDigital connectivity can extend the same advantage beyond a single geography. Recurring cross-border interactions include scalable services, proprietary data, platforms, and IP. Businesses are finding different ways to translate cultural knowledge, institutional access, and cross-border data into business models that serve customers across multiple countries, such as the following:
\nF\u00e9lix Pago, a fintech start-up, facilitates US\u2013Latin America remittances through a fully digital WhatsApp-native interface.
\nDesteia, which has raised seed funding of $11.5 million as of 2025, integrates with Mexico\u2019s customs filing portal to simplify compliance for MNCs.
\nSonoro, which raised $12.5 million in Series A funding, has built one of the world\u2019s largest podcast networks by creating culturally relevant audio content for audiences across the continents.
\nBy creating ways to serve long-standing needs of Latino customers across geographies, digital integration converts the advantages of Latino consumers and businesses into market-defining plays.
\nFrom distinctive advantages to proven plays
\nLatino consumer, business, and place-based advantages are already translating into action. Across the US\u2013Latin America corridor, companies, entrepreneurs, investors, governments, and ecosystem institutions are using them in a set of recurring plays that create and capture value, showing how Latino-linked demand and capabilities can successfully create opportunities across markets.
\nMNCs are importing and commercializing Latin American products in the United States, often using existing Latino demand as an initial market before scaling more broadly. Modelo illustrates the potential, scaling from a heavily Latino consumer base to become the number one beer by dollar sales in the United States. MNCs are also investing to expand across Latin America, bringing capital, technology, infrastructure, procurement, and operating capabilities. This trend is particularly pronounced in the technology sector: For example, Microsoft invested $2.7 billion into cloud and AI infrastructure in Brazil, and Amazon announced a $5 billion cloud-region investment in Quer\u00e9taro, Mexico.
\nUS Latino-owned businesses and entrepreneurs are pursuing a broad set of cross-border models. Across the models, each uses their unique knowledge of the Latino consumer and preexisting networks to carve out a competitive advantage. Some import and commercialize culturally connected goods, using their knowledge of Latino demand and sourcing networks to build scaled businesses. Others export brands and business models into Latin America, leveraging propositions first developed in the United States as well as cross-border market knowledge and owned assets. Nopalera, for example, which has an established footprint across more than 700 retail points in the United States, recently partnered with Ulta Beauty to launch in Mexico and leveraged a $4 million Series A funding to further fund its domestic and international expansion (see sidebar \u201cUS Latino-owned business spotlight: Nopalera\u201d). Others are building digital services that solve recurring cross-border frictions: F\u00e9lix Pago processed about $1 billion in remittances in 2024, Nuvocargo reports customs clearance times that are up to 34 percent faster, and Nowports reports shipment delays that are up to 40 percent lower through technology-enabled logistics and customs capabilities.
\nPlaces are developing plays of their own. State and local governments could build infrastructure and specialized capabilities to help make trade easier or focus on particular sectors rather than trying to replicate a full gateway. Laredo and El Paso illustrate deep logistics and customs ecosystems, for example, while Tampa shows how a secondary hub can build differentiated capabilities in cold-chain and food distribution. Specialized connectivity can materially expand access to corridor trade. Among cities with similarly sized Latino populations, stronger links to Latin American entry ports are associated with around three to four times greater availability of imported goods, and cities such as San Antonio illustrate how logistics and sector specialization can turn that connectivity into economic activity.
\nCapital providers and ecosystem institutions play a complementary role by helping promising corridor businesses access the capital, customers, and capabilities needed to scale. Capital providers offer fit-for-purpose capital, including growth equity, working capital, and trade finance. For example, L\u2019Attitude Ventures built a $100 million venture capital platform for early-stage investments in US Latino-owned businesses (including Nopalera), while fintech platform and marketplace Mercado Libre uses proprietary sales and payment data to rapidly assess creditworthiness and provides working-capital financing to merchants that are underserved by traditional lenders.
\nEcosystem institutions such as Latino chambers and supplier networks connect firms to customers, distributors, capital, specialized hubs, and country-specific expertise that would be difficult for any business to assemble alone. One example is ConnectAmericas, an Inter-American Development Bank (IDB)\u2013created platform enabling Latin American businesses to achieve 17 percent higher average export value by connecting them to a network of about 240,000 companies with capital opportunities, buyers, suppliers, and training.
\nTaken together, these initiatives show the corridor already contains many of the ingredients needed for broader value creation. The challenge is not the absence of successful models but rather that they remain fragmented and are not yet connected to an integrated system that can scale them consistently.
\nDriving greater US\u2013Latin America economic value in the corridor
\nBroad and largely external forces such as global policy, cross-border investment, and commodity trade cycles drive much of what forecasts anticipate for the corridor. Based on these trends, a McKinsey Global Institute analysis implies a change in growth rate from 3.0 percent in real terms in the past decade to around 1.5 to 2.0 percent per year through 2035, translating to $1.2 trillion to $1.7 trillion in cumulative added trade volume, with the current balance between northbound and southbound holding roughly stable as a share of corridor trade.
\nYet there is a potentially sizable additional opportunity centered around the distinctive capabilities of US Latinos. Despite these advantages across businesses, consumers, and places and proven value creation plays, the ecosystem needed to capture the value remains fragmented. Capital, market access, institutional support, commercialization pathways, and ownership mechanisms are frequently unavailable or too dispersed for the companies and places best positioned to use them. We estimate that systematically addressing these barriers that prevent proven models from scaling could add $170 billion to $295 billion in incremental trade through the corridor in the next ten years, expanding the corridor\u2019s growth by 18 percent. Taken together, the corridor\u2019s momentum and this added opportunity represent up to $2 trillion in cumulative added trade over the next decade. The benefits would extend beyond the businesses that gain new opportunities to import and export. There could also be sizable spillover effects into the broader economy that create jobs, lower prices, and generate more choice for consumers and more vibrant places.
\nBut the additional value is not guaranteed. Capturing it depends on addressing the five system constraints.
\nLimited access to capital
\nSurveys of Latino entrepreneurs have documented that more than 70 percent of Latino-business owners used their personal funds to start their businesses, and as few as 12 percent use banks or financial institutions ; Latino-owned firms are also less likely to receive the full financing they request compared with other groups. At the same time, firms based in Latin America that would otherwise be exporting or importing are unable to access finance to engage in trade. It is estimated that expanding access to trade finance (for example, through working-capital loans, receivables financing, letters of credit, and supplier finance) could increase trade volumes by 6 to 9 percent across Mexico, Guatemala, and Honduras, all of which count the United States among their largest trading partners. Improving access to different capital products at different stages of growth on both sides of the corridor could enable more businesses with promising business models to capture value from trade, especially in the critical transition when founder-funded firms seek institutional scale.
\nLatino-owned businesses have long shown their ability to effectively use capital to expand into international trade and business activity. By the late 2000s, Brightstar, founded in Miami by Marcelo Claure, became the largest Latino-owned business in the United States by building a scaled mobile phone exporting business that grew revenue from $140 million in 1999 to more than $1 billion by 2003. The key to its early expansion was using working-capital financing to build inventory that could be sent to Latin America. Today\u2019s businesses are using newer financial mechanisms to reach customers across markets. Miami-based tech start-up F\u00e9lix Pago, for example, has raised multiple rounds of venture funding to expand its remittance platform across multiple Latin American countries. Since its founding in 2020, it has served six million customers across 11 Latin American markets and processed $8 billion in transactions.
\nNarrow market and distribution connectivity
\nAbout 75 percent of trade in the corridor is concentrated in leading hubs that reach only around 16 percent of US Latinos. And further dispersion of traded goods from ports of entry to consumers in other Latino population hubs is uneven. Some of the largest Latino markets in the United States remain relatively disconnected from the trade infrastructure and distribution networks that would allow demand to translate more fully into corridor flows (Exhibit 5). This finding is consistent with McKinsey\u2019s The economic state of Latinos in America report, which documented that 41.5 million Latinos live in consumer deserts and are 1.3 times as likely as White populations to have low access to consumer goods.
\nThis highlights meaningful opportunities for greater consumer import penetration. For example, Los Angeles and Las Vegas have sizable Latino populations but a comparatively low level of Latin American imports, suggesting room to expand sourcing and distribution, particularly given their proximity to Mexico and cross-border logistics advantages. Companies may consider whether cost, lack of knowledge about Latin American imports, access, or other factors are barriers for consumers. Other markets show what stronger connectivity may unlock for cities that are not major ports of entry themselves: Tampa has built a differentiated position in food distribution, for instance, while Dallas functions as a major logistics hub that supports Latin American imports well beyond what local demographics alone would suggest.
\nMore broadly, among places with similarly sized Latino populations, stronger connections to Latin American entry ports are associated with materially greater availability of Latin American consumer goods. Latino businesses need consistent pathways to access MNC procurement, retailers, distributors, platforms, corporate customers, and export channels. Without them, commercially viable businesses tend to stay local or niche. Broader market access and distribution connectivity may unlock significant economic value by converting demand into repeat purchases, larger orders, and national or regional reach.
\nSource Logistics, founded by Latino entrepreneurs, highlights this opportunity. It has made growing coverage of the US Latino population its core growth strategy to meet rising demand for Latin American goods, especially food and beverage. Through M&A and new warehouse builds, it has expanded to 5.8 million square feet of warehousing nationwide, reaching 80 percent of the US Latino population and 75 percent of the total US population with same-day delivery; its longer-term goal is to achieve 95 percent total US population coverage. This growth helped it rank on the Inc. 5000 2025 list of America\u2019s fastest-growing private companies for the second consecutive year.
\nAs connectivity improves between trade hubs and other cities, opportunities will emerge for businesses with rights to win as exporters. Quirch Foods, which started as an importer of consumer goods from Latin America, built distribution networks across the United States from a central distribution node in Miami. While 80 to 90 percent of its business remains domestic, over time, its import business led to relationships with domestic meat and seafood producers that it combined with its cross-border networks and infrastructure assets in Miami to expand into exports. Now Quirch ships more than 100 containers a week globally, with a focus on Latin America. Other successful export businesses in the corridor highlight the value stronger networks can create. For example, SolvChem, a Latino-owned chemicals distributor based in Houston, created an international division in 2009 to export chemicals to Latin America. Having already built chemicals handling infrastructure in Houston and a regional network of suppliers, it used its proximity to a key corridor port to become a dominant player in chemicals exports, generating more than $100 million in revenue per year.
\nLow institutional coordination
\nDespite an extensive support ecosystem, Latino trade participation remains low. For example, the USHCC represents more than 260 chambers, and the NMSDC connects more than 15,000 minority-owned businesses through 23 regional councils, but only 2.1 percent of Latino-owned employer firms export, compared with 2.6 percent for all US employer firms. An accountable corridor-wide coordination layer could connect commercial development, capital, distribution, specialized hubs, country-specific regulation, market knowledge, and skills, surfacing promising opportunities that would otherwise be missed.
\nThe United States also lacks a coordinated and scaled set of efforts to amplify the commercial power of Latino cultural assets. Such vehicles have been deployed globally to great effect, such as Enterprise Ireland, which holds capital, capability building, and market access under one mandate: The same agency that funds a company also opens markets for it with its 42 overseas offices so the firm never has to assemble the system itself. In 2025, its client companies had record export sales of \u20ac38.86 billion and employed 232,425 people. The United States has already had major success promoting commercialization of other cultural assets at home and abroad, such as through its film industry, which has an estimated global annual trade surplus of $15 billion. A coordinated group of stakeholders in the United States, such as investors, trade associations, or nonprofits, could focus on identifying emerging trends with commercial potential from within the Latino community and support them to develop their business models, leveraging import and export opportunities.
\nBrazilian influencer Camila Coelho illustrates what\u2019s possible when entrepreneurs can scale products in the corridor. She leveraged her more than ten million social media followers to launch two companies generating more than $10 million in combined sales by 2021: a fashion label with Revolve and a beauty line built around ingredients native to Brazil reaching more than 400 Ulta Beauty stores and shipped to 17 countries. She has used the corridor by turning Brazilian ingredients into US-launched products sold back into Latin American markets. With an enabled ecosystem, entrepreneurs like Coelho could establish a new age of US cosmetics and fashion rooted in the US Latino community.
\nNascent commercialization pipeline
\nThe opportunity for Latino firms\u2019 broader participation in the corridor is not constrained by a lack of demand or business performance but by an uneven ability to commercialize and scale. The US\u2013Latin America corridor lacks a pipeline that identifies promising firms and assets, validates demand, develops products and talent, builds export or market entry readiness, protects IP, codifies playbooks, and replicates what works. Having repeatable capabilities, playbooks, and intermediaries that identify promising opportunities and help move businesses from proof of demand to scaled market participation could unlock substantial value for Latino-owned businesses. Other economies have built central agencies or intermediary institutions to perform this pipeline-building function. For example, Business Sweden, a collaboration between the Swedish government and Swedish Foreign Trade Association, institutionalized the concept of getting proven firms export-ready and has helped thousands of Swedish companies enter international markets, including prominent and fast-growing businesses, such as the electric and autonomous freight company Einride, which went public earlier this year.
\nThe missed commercialization opportunity is especially stark for Latino cultural assets, which are unlocked by trade but have potential for economic spillover effects in the US economy. Bad Bunny offers one example: The singer and rapper\u2019s rise was enabled in part by Puerto Rico\u2013based Rimas Entertainment, which secured an early YouTube monetization agreement that helped Bad Bunny reach audiences without relying on a traditional major-label model. That model has translated cultural demand into broader economic activity. During Bad Bunny\u2019s 2025 Puerto Rico residency, total spending in San Juan rose nearly 20 percent versus comparable weekends a year earlier, while international spending around the venue increased by more than 15 percent. The residency also connected visitors with local businesses and helped expand Puerto Rico\u2019s reach into new tourism markets, a tangible display of the economic power of culture that went global just months later with Bad Bunny\u2019s performance at the Super Bowl LX halftime show.
\nIf the United States could build a similar coordination and amplification apparatus to the same results that South Korea achieved from its Hallyu cultural engine, the economic opportunity in the United States could be worth more than $300 billion over ten years, with trade-related activity adding another $60 billion.
\nWeak ownership and reinvestment mechanisms
\nEconomic growth in the US\u2013Latin America corridor can generate income without creating durable, Latino-owned assets that let value compound over time, such as equity, brands, proprietary data, platforms, and distribution networks. Weak mechanisms for ownership and reinvestment mean founders, creators, and communities often don\u2019t retain a lasting stake in the value they help create, which is evident at both the household and the firm level.
\nAt the household level, Latino households that own a business hold roughly 2.2 times the net worth of those that do not, yet only 6 percent of Latino families receive an inheritance, part of a $380 billion wealth transfer gap that McKinsey documented.
\nAt the firm level, Latino-owned firms capture less of their industries\u2019 value than peers, generating less than half as much revenue per firm as other firms (excluding MNCs). This suggests a scale gap that compounds the ownership gap.
\nLatino firms in the corridor show how scaled ownership can create significant value. Goya Foods, for example, highlights the payoff of retained ownership and reinvestment. It has grown into a multibillion multinational food distributor while retaining family ownership despite multiple buyout attempts. Its former COO went on to found AUA Equity, which has an investment strategy built around its distinctive knowledge of scaling products nationally from the Latino business ecosystem.
\nCreating and capturing greater value from the US\u2013Latin America corridor
\nA more integrated US\u2013Latin America ecosystem holds the promise of generating significant economic value, especially for US Latinos\u2014although it will not materialize automatically. Creating and capturing this value requires investment to build an integrated ecosystem and amplification engine for the corridor that eases entry for new participants while helping existing players scale. Such a system would allow US Latinos to apply their distinctive capabilities to proven plays to capture more value while creating spillover effects that benefit communities. For stakeholders already engaged with the corridor, the ecosystem\u2019s amplification engine can help them succeed more often and scale faster by connecting them to system support they may be missing. As more plays succeed, assets are created along with knowledge, relationships, data, brands, and capital that strengthen the next cycle of growth in the corridor. New participants will also benefit by having better access to the existing ecosystem and access to the market entry playbooks that have been codified.
\nRealizing these opportunities in the corridor could have substantial impact on the economic mobility potential for Latinos in the United States. As Latino-owned businesses succeed and grow more frequently, the share of wealth that the Latino population claims through ownership increases. When stronger networks enable better exchanges between businesses about the proven plays, more businesses can enter adjacencies that have higher potential for export, a proven way to increase value creation. As businesses in the corridor grow, they create employment opportunities for other Latinos in their communities, which in turn enables skill development and opens career pathways for wealth creation. Finally, corridor participants reinvest resources into their communities through direct spending and through tax revenue generation, creating a financial base to improve quality of life through services and amenities.
\nInvestors, businesses, governments, and mission-driven organizations looking to participate in the corridor to access these opportunities don\u2019t need to build every capability themselves. Key stakeholders need to proactively collaborate to make this integrated ecosystem and amplification engine a reality. Each actor will be responsible for building distinct but complementary and coordinated levers that can work together to open the pipeline of opportunities. A summary of each stakeholder, how they benefit from the ecosystem, and steps they could take is in the table below.
\nThe trade corridor between the United States and Latin America is big, growing, and critical to the economic dynamism of the region\u2014but there are pockets of opportunity that have yet to be realized. The path is not without risk. Headwinds and uncertainty could impact how these opportunities can be realized. Nonetheless, commercial activity in the corridor will continue, and many of the ingredients for generating greater value are already in place. An integrated system that consistently scales distinctive Latino capabilities could turn this potential into greater economic momentum, providing another pathway for economic mobility for Latinos and the communities they live in nationwide.
", "image": null, "banner_image": null, "date_published": "2026-10-18T16:00:00Z", "date_modified": "2026-10-18T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}, {"id": "https://www.mckinsey.com/insights/b5702bdf-5280-4b4a-a64e-265aaf395035", "url": "https://www.mckinsey.com/mhi/our-insights/closing-the-womens-health-gap-switzerlands-chf-7-billion-opportunity", "external_url": null, "title": "Closing the women\u2019s health gap: Switzerland\u2019s CHF 7 billion opportunity", "content_text": "", "content_html": "Addressing the women\u2019s health gap in Switzerland could deliver eight additional healthy days per year for every woman and potentially contribute more than 7 billion Swiss francs in annual GDP by 2040.", "summary": "Addressing the women\u2019s health gap in Switzerland could deliver eight additional healthy days per year for every woman and potentially contribute more than 7 billion Swiss francs in annual GDP by 2040.", "image": null, "banner_image": null, "date_published": "2026-10-18T16:00:00Z", "date_modified": "2026-10-18T16:00:00Z", "authors": ["Unknown"], "tags": [], "language": null, "attachments": ["\n {\"url\": string, \n 'mime_type': string, \n 'title': strinrg,\n 'size_in_bytes': int,\n 'duration_in_seconds': int\n "]}]}