Interview with Jim Breyer #venturecapital - Codice 11/07/2025

We're so excited to be at the very beginning of where AI, medicine, AI and healthcare are intersecting. It's really in the last 5 years that I've developed the thesis that AI is fundamental to almost everything that we'll see and do. It affects finance, transportation, but perhaps most importantly, it has profound and very positive potential to make a big difference for patients, for doctors, for drug discovery. So I've made about 15 investments in AI and health, AI and medicine and they span from diagnosis to preventive care to new therapeutics and drugs that can make a very big difference in cancer and cardiology. The model is very typical. I will find the best in the world, Nobel laureates or potential Nobel laureates in different areas of medicine and use their data and work with them to identify a promising new opportunity. My job is to go find the brilliant AI talent out of Google, Meta, Microsoft, people generally in their 30s who have been at these companies for 10 years who want to be on a mission to change medical outcomes for the better.

Well, the data is very interesting because so many people believe they have the best data and the fact of the matter is that much of the data that's most important for the world of medicine has not yet been invented or has not yet been captured. So one of the profound and positive implications I'll talk about one example of a wonderful, probably the single best children's heart transplant team in the world out of the University of Austin Texas and Chuck Fraser and his team specialize in doing heart surgery for infants. Much of the data they need to capture in the operating room to date has not been captured. So they will do operations, they will save lives, it'll make a profound difference, but if a child needs a cardiac operation, they'll have to go to Austin, Texas to get that operation. One of the great promises of what Chuck Fraser and his team are doing and others is you can capture the data and suddenly enable those kinds of surgeries to occur in other parts of the United States and other parts of the world. And as we capture more data, whether it's for second opinions or for other hospitals which don't have that specialization, we can very quickly determine where are the right places to go for that kind of surgery. And again, we've never captured the real time operating room data. And that's a new trend because there are some hospitals around the world which have very good specialized data, but they haven't scaled to provide that data in many cases to other hospitals and other systems and other countries. Through AI we're finding that opportunity to occur.

Well, I think it's largely a bubble. My personal view, but I do spend a lot of time in hospitals trying to understand what the very best doctors and leaders are thinking about. So one example at Harvard Medical School and Mass General Brigham in Boston, they are doing very deep research on psychedelics and what the impact might be. But the research is early and the preference of Breyer Capital has been to invest in many other areas of where medicine and AI intersect. That's not to say there aren't good opportunities down the road, but in many cases, I think it's a bubble. In many cases, I think the teams are far too promotional in terms of having hard, clear scientific evidence. I'm always open to learning more, but for now, it's not an area where we've spent a lot of time.

Again, if I had to characterize it, in nine cases out of 10 it's too promotional. However, there are cases where some of the science around biohacking is proving to be very important. And so there are two really fundamental parts of medicine that I think are important and where AI is playing a big difference. One is preventing disease and we are finding drug development companies that are making big breakthroughs through AI simulation that were not possible before. At the same time there is the longevity piece. I think many people would love to live an extra decade or two if they can live well and that's the importance of some of the biohacking capabilities. And so we are seeing some leading edge technologies where whether it's dementia, whether it's other elements of aging, where the fundamental science is proving to be very sound and being published in very important journals like the New England Journal of Medicine, Cell, Nature, and other important journals of medicine. I would say that by and large still like anything else there are a lot of bubbles that emerge when something is very fundamental. But there is some fundamental science which will unleash all sorts of very interesting and unique opportunities.

Well as a venture capitalist and investor I always worry. But of course, but now more than and perhaps a little more. But what I will also say with AI, I'm an optimist. And my friend Michael Dell, who founded Dell many years ago, will always tell me that to be a venture capitalist or an entrepreneur, we have to be optimists. I'm very optimistic about how AI will have a profound positive influence on venture capitalists and investors who think out of the box and creatively about what the implications are. It was actually last week where I was on the floor of the New York Stock Exchange and we took one of my investments public, Circle Internet. And again it was many hundred times the initial cost as we took it public last week but it was a 12-year journey and I think we can never forget that venture capital typically is a 7 to 10 year cycle. AI will not change that. But we have to be very receptive to understanding where does AI displace venture capitalists. Certainly I use AI five, six times a day for every new potential investment that I'm looking at. I could not do my job without doing deep research in AI.

Well, it's a wonderful question and years ago I came to Florence, Italy and studied the Renaissance, not just art, but the philosophy and many other elements of what the humanism was like in the Renaissance and put together a course on how that compared to modern-day Silicon Valley. And so I think there are very clear lessons for AGI and how we as humans will use AGI. In some cases, AGI will surpass so many of our different activities. But I'm also a believer that humanism, philosophy, empathy, these are skill sets where I don't see AGI matching human beings. And so we have to understand where the very advanced AGI technologies are going. But in terms of certain judgment in interdisciplinary areas where we have to make judgments that are not necessarily able to be modeled in AGI anytime soon, the brain has a way of helping us come to decisions where I don't see artificial intelligence changing that anytime soon. So the most powerful combination is where we are using AGI and artificial intelligence to improve a number of decisions. However, we must always remember I have not yet seen AGI or any artificial intelligence system that has the human ability to have empathy, to have the wisdom that is so important by studying philosophy and other elements of what goes into really important decision-making. So again, I'm very optimistic, but for sure AGI artificial intelligence is moving extraordinarily quickly and will only accelerate.

Well, let me dial back because it's almost exactly 20 years ago. It was April of 2005 and in Palo Alto I met a 20-year-old Mark Zuckerberg on a Monday and several people had referred me to Mark and so we sat down for a meeting on a Monday and there were several elements of what Mark represented at age 20, he was not yet 21, that really resounded with me which made me think after that first meeting I'd better move quickly. This is someone I'd really like to back. One, Mark had deep product passion. So the facebook.com website at the bottom said a Mark Zuckerberg production. A rule of thumb for me is the entrepreneur needs to own the product. And it was very clear Mark owned the product and stood by it. And I had looked at MySpace and several first generation social networking companies and I didn't feel the teams really cared deeply about the product. The other element that was so important at that first week of meetings, Mark told me one day I would like to connect three billion or more people. He was not yet 21, but he had that vision. And at that time, it was on several college campuses. There was no proven business model. But that sheer ambition and vision was something that I look for in all entrepreneurs. They're not all Mark Zuckerberg, but it caused me to move very quickly. And I knew Mark was someone I wanted to bet on very heavily. And so we negotiated that week and I invested $11 million and joined the board. And it's been a joy to have worked with Mark since April of 2005.

Well, sometimes I've worked with the entrepreneur once or twice before. So again, you ask me about some of the most promising, interesting AI opportunities. There's a third-time entrepreneur who I've backed now for the third time, Daniel Nadler, and a company called Open Evidence. And when Daniel came to me with his new idea two years ago, I immediately said, 'Count me in within two hours.' And his idea was very simple. There are thousands and thousands of important medical journals, but doctors and nurses do not have the time to keep up with the latest medical information. And so he thought and correctly so at Open Evidence that using AI would help doctors and nurses surface what information from these journals is most relevant to them either as cancer specialists, cardiologists, whatever the doctor, whatever the nurses are focused on. At the point of care they can pull up information, the latest information from all the journals and provide that to the patients and the doctors and the nurses. It's that insight which often to me really resonates. I will meet countless numbers of entrepreneurs. But when there's a potentially explosive large market with product passion, with judgment, with the ability to also hire extraordinary people like Daniel Nadler does, like Mark Zuckerberg does, and others. Those are all the skill sets where from a gut standpoint, I'm internally thinking this is someone I absolutely want to invest in.

I think that constant tension I'm usually early and I'm sometimes too early. So there are areas for example where I've invested in the area of quantum technologies and that was four years ago and now it appears indeed quantum algorithms and technologies are becoming very important by the end of the decade but it's still a seven or eight year horizon and often one of the challenges is to identify important fundamental technologies, but at the same time have a feel for how long it will take for those technologies to have an impact on customers, whether it's consumers or businesses. And a mistake I'm always wrestling with is how early is too early? Because we all know if it's far too early, we'll need to put in more and more capital and the level of financial return will be diminished. But you can't be late. The competition is better than ever. We have the single best set of competitors I've ever seen, whether it's Meta, Microsoft, Apple, Alphabet, I could go on and on. And so the room for error if one is not early enough is very small. So it's better to be early. But if one is far too early, then it becomes more of a research project and not a for-profit long-term successful investment.

No, I think quantum is going to work out. Now it will work out. But I was early and that's systematic of what I'm trying to balance most often. Usually that mistake you can recover but it may diminish the returns. I'm very optimistic about how quantum will play out whether it's quantum sensing for medicine whether it's for security applications but the biggest mistakes I rarely make the mistake of the wrong entrepreneur and I rarely make the mistake of investing in something a project where there is not a market eventually. The challenge is putting it all together: the market, the timing, in many cases regulatory risk and factoring that in. For instance for a lot of the medical AI companies, FDA and in Europe many of the medical regulatory agencies can play a very big role and perhaps I've underestimated how challenging it might be sometimes to navigate the regulatory agencies but I'll keep trying.

Generative and agentic. Well, we're watching it and so almost everything will be affected. So we can think about education and universities and how we teach that will be profoundly affected by AI and Gen AI. So today if students are not studying AI or AI is not incorporated into the curriculum there will be a real challenge. And so whether I'm in Italy or whether I'm in the United States, I encourage teachers, professors, our universities to be teaching our students how to appreciate the strengths and weaknesses of generative AI. Medicine is profoundly being changed, but at the same time, it's not just the generative AI piece. AI from a deep technology standpoint. Being able to make a big difference in simulations is not Gen AI but very deep modeling and simulation AI usually running on Nvidia chips and AMD but to do massive simulations whether it's molecular simulations whether it's areas around security for governments making sure our networks are secure making sure privacy is secure pure deep what I would call quantitative AI models will play a very fundamental role and that's not where Gen AI is necessarily the best solution.

Well, Eric is very smart. He's a very good friend and he knows much more about nuclear energy and has studied nuclear energy far more than I have. I think his fundamental premise is correct. We have to figure out whether it's electricity, power, how networks are built, how the utilities are serving the data centers. We have to figure out how to make this happen. And it's a big challenge. So for sure, these are big challenges. I don't know enough of the details about nuclear to have a strong conviction yet on the role of nuclear. I do know that data centers and a lot of these applications we've talked about will demand far more electricity than what is modeled currently and we have to come up with a variety of different solutions. So it is a bottleneck and one of the characteristics of the buildout of the data centers that I'd like to do over the next 12 months is understanding in far more detail the pros and cons of using nuclear energy for instance.

Well, I'm very optimistic about Silicon Valley remaining the center of breakthrough artificial intelligence. And again, the seven or eight leading companies in the world are largely based in Silicon Valley and Seattle, Washington, if we count Amazon and Microsoft. However, for startups, I also have the thesis which brings me to Europe often that now the really nimble smart startups, if we take Italy for instance, if we take Milan or Rome or Turin, these startups can now tap into the many technological breakthroughs that are occurring in Silicon Valley and localize and build solutions that were not possible 5 years ago. And so one of the boards I'm on, my lone public board is Blackstone. And Steve Schwarzman this week has announced that for Blackstone, we will invest $500 billion in Europe over the next 10 years. That is a true statement that I think is very encouraging for Europe in particular. I think that Europe has the intellectual talent from a university perspective whether it's Italy, France, the United Kingdom, universities are superb. Now with funding and capital becoming more and more available, that part of the equation helps get solved as well. What Italy in particular I think needs to do better and I've had these conversations with my venture capital and entrepreneurial leaders the last week in Milan in particular there needs to be local ecosystems like we have in Silicon Valley where the university professors are working very closely with the entrepreneurs and the students and the postdocs and so it needs to be more seamless and needs to move much more quickly. I think the opportunity that AI allows is for a re-imagining of what a startup here in Rome might look like or a startup in Milan might look like that taps into the great universities that are local and at the same time the many technological tools and layers that the best Silicon Valley companies are providing so that they develop very specific applications for their own markets where there's specialized local knowledge. That's why I'm very optimistic about investing. It's not too far behind for Europe. Sometimes we think it is, we Europeans think it is. The regulatory issues are challenging. I think that has held back a number of potentially wonderful startups. Even if you get an EU blessing, as we know, that doesn't mean that Italy or France will bless the opportunity. So, in some ways, the bureaucratic nightmare still exists. But again, from a regulatory standpoint, I'm encouraged by what I'm seeing as I meet with regulators on the understanding that if there isn't change, Europe again will play catchup. So the opportunity I think is really to understand regulatory has been at the heart of slowing down so much of the European startup ecosystems and I'm seeing optimistic signs which causes me to remain more and more bullish about opportunities in Italy very specifically.

I think that the privacy I would not necessarily change dramatically. However, I think the layers of bureaucracy need to be changed. It's very difficult to go try to strike agreement in Brussels with the EU and then strike individual deals for each country in Europe whether it's Italy, France, Germany to pick three. And I think that there needs to be more cohesion. Overregulation of AI, my personal view, would be a huge mistake because we don't yet know enough about AI to do detailed regulation of AI. Some high-level bars and grounds of AI regulation, I think, is very important. But not being too specific and not getting down into the weeds and trying to do 150 pages of documentation. That is my advice for the regulators.

Well, it's a race. There's no doubt. And I would think that of the thousand most important AI papers coming out of the world, the real breakthrough ideas, whether they're out of Stanford or MIT or elsewhere. A number of the breakthrough papers are coming out of China as well. The really important ideas. Again, one of the reasons I'm very optimistic, there's a lot of great research and a lot of great talent at the United Kingdom universities, the Italian universities. So, the thinking is very deep and world class. What needs to happen is it needs to be accelerated and it needs to be accelerated in a way where it can be commercialized and scaled and that's what's really held back Europe. So again, if I was advising European leaders, whether they're Italian leaders, French, United Kingdom, reduce the friction, allow startups to flourish. Encourage startups to flourish. Make it clear how if there are questions, here are the people startups and venture capitalists can work with because I think the talent here in Europe is extraordinary.

I'm fortunate at this point in my career. I care deeply about the impact and whether it has a positive impact. But before not, well over time, one never quite knows because for example at Facebook we didn't quite understand in those early years what the privacy and security issues might be seven or 10 years later. And so did we underinvest in privacy and security seven or 10 years after the founding? Yes, absolutely. So I've learned lessons along the way. But one of the reasons I'm so passionate about AI and healthcare, AI and medicine. It's an area where I think the returns can and will be outstanding, where the benefits to patients, to doctors, to nurses will be outstanding as well.

I would be investing in AI and medicine and I would have offices in Europe and the United States for sure, dual offices. I would be backing many of our great university professors and postdocs and I would be trying to attract the generation of AI expertise that's in their 20s and early 30s that are AI native. And so I'm not quite 25 but it's what I love to do and I would also be studying mathematics because mathematics is at the heart and soul of all of these opportunities. And so again, when I'm advising students, whether it's at Polytechnique, whether it's at Stanford, MIT, Harvard, Cambridge, Oxford, I encourage the students, be sure as part of your curriculum, take math and linear algebra because that is at the heart of all the opportunities we're talking about in AI and quantum.

I think it's artificial intelligence. It's not quantum yet. Well, AI and quantum are very related and Nvidia today and Jensen talked about how he's bringing in his timeline for quantum that in fact we're seeing more progress more quickly in quantum development and technologies than many assumed. So I think quantum will have a huge impact over the next 5 years. But it's still artificial intelligence. But it's both. I don't believe quantum computers are necessarily the be all and end all. There will be important quantum computers and IBM in particular has developed quantum computers that I think will by 2030 play a fundamental breakthrough role. But quantum algorithms running on AI chips from Nvidia, that's the future.

The most important set of skills is interdisciplinary sets of skills and that's very different from a decade ago. There was a time where I could invest in the very best computer scientists from Stanford or MIT and we could build a great company. Today it's not enough just to back the great physicists or computer scientists. There has to be domain knowledge. So if we go back to medicine, what's important in each of these teams? There needs to be medical expertise. There needs to be an understanding of biology, of chemistry in addition to the computer science. That team looks very different from what teams looked like where it might have been just 5 to 10 brilliant computer scientists a decade ago. So part of the challenge is identifying the brilliant doctors, chemists, biologists who appreciate that she or he needs to work with the brilliant AI developers in their 20s. That balance will be the recipe for huge outside success. Thank you.