a16z Podcast - Novartis CEO Vasant Narasimhan on Transforming a 250-Year-Old Company

Episode Date: February 16, 2026

a16z general partner Jorge Conde talks with Vasant Narasimhan, CEO of Novartis International, about transforming a 250-year-old conglomerate into a pure play medicines company and unlocking $180 billion of value in the process. They cover Novartis's platform technologies: cell and gene therapies, RNA medicines, and radioligand therapies. They also discuss AI in drug discovery, the rise of China as a biotech competitor, and what Vasant looks for when evaluating startup partnerships, including his advice on the killer experiments and CMC work that can make or break a deal. Resources: Follow Vasant Narasimhan on X: https://twitter.com/VasNarasimhanFollow Jorge Conde on X: https://x.com/JorgeCondeBio  Stay Updated:Find a16z on YouTube: YouTubeFind a16z on XFind a16z on LinkedInListen to the a16z Show on SpotifyListen to the a16z Show on Apple PodcastsFollow our host: https://twitter.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

We unlocked almost $180 billion of value. Took this company that was a conglomerate, and we turned it into four companies. We had these other businesses, but it was never really the core of kind of who we are. We were misallocating capital. Spinning off companies sounds easy in retrospect. Actually, spitting off three public companies

is a massive, massive undertaking. You can rewind a disease or completely, in some cases, almost create a functional cure for patients. One of our sentinel patients was an individual in-bed, all the time receives our cell therapy. Six months later, no sign of disease. I mean, this is extraordinary. Cell therapy is an immunology alongside immune reset by specific, tri-specific.

That is going to be a whole renaissance for cell therapy. When Vassant Narasimhan became CEO of Novartis in 2018, he inherited a conglomerate, consumer health, animal health, vaccines, generics, devices. Over six years, he spun off three public companies and unlocked almost 100,000. $180 billion of value. Today, Novartis is focused on four disease areas and three platform technologies, cell and gene therapies, RNA medicines, and radioligan therapies. One of their gene therapies, Zolgensma, treats children with spinal muscular atrophy,

a disease that would otherwise be fatal by age two. It costs over $2 million, and it's reimbursed in 48 countries. Basin started his career doing public health work on tuberculosis in Peru with Paul Farmer and Jim Kim. That still shapes how he thinks about access to medicines today. A16Z general partner Jorge Condi talks with Vassant Narasim Han, CEO of Novartis International. So Voss, I want to welcome you back to the A6NZ podcast. Thank you very much for being here. It's great to be back.

Thank you. Really appreciate it. Appreciate you making time. So why don't we start with just a brief introduction? So you are the CEO of Novartis. Novartis is, of course, one of the top 10 global pharmaceutical companies. has had major impact across a range of diseases, has major impact on patients around the world.

But for the audience, maybe we start with that brief introduction. Can you give us what Novartis is in a nutshell and who Voss is in a nutshell? Why don't we start there? Yeah, absolutely. So Novartis, 250-year-old company that's really been a conglomerate for most of its history. But I think we'll talk about it over the last 10 years.

We really focused it down as a Pure Play Medicines company. we reach about 300 million patients around the world. We operate in over 100 countries, over roughly $55 billion in sales, nearing a 300 billion market cap. And we really focus on four core therapeutic areas, oncology, immunology, neuroscience, and cardiorenal. And we're really into platform technologies. I've really tried to pivot the company into leading platform technologies. Myself, I'm a physician scientist by training. I grew up mostly doing R&D work and public health work at the company and with some business roles,

became a CEO in 2018, and in that period of time, we've spun off a lot of businesses, focused down. I think there'll be a lot of what we talked about today. Yeah, one of the things I loved about your background is the work you did as a student. So if I read your bio correctly, you did your thesis in multi-drug resistance and tuberculosis patients in Peru. That is correct. Yeah, a lot of my time in medical school and shortly thereafter was focused on global public health. That's actually I think a lot of why even today I focus Novartis heavily into being a leader in neglected tropical diseases, access to medicines. But I worked with luminaries like Jim Kim and Paul Farmer, multi-drug resistant tuberculosis, access to antiretrovirals to patients in Africa.

That was a huge part of my early development, early part of my career, and I try to take that forward now. Yeah, and the reason why that stood out to me is my father's. Peruvian. So that half of me thanks you very much. Yeah, well, I spent a lot of time in Lima and it was amazing to work on that project. Absolutely. Wonderful. Okay, so let's jump into Novartis and the work you've done over the course of the last eight years or so. One of the things you said in your Novartis in a nutshell intro is that you really shifted the company into what you call the Pure Play Medicines Company. So can we unpack that a little bit? What does Pure Play mean in your mind? And what does it mean to call a pharmaceutical company a medicines company? Yeah, absolutely. So

So Novartis formed out of the merger of Sandoz and Sibaghi in 1996, and the focus at the time was becoming broadly diversified in health care, which was common, I think, for many of the main players. So Novartis at the time, we were in consumer health, animal health, vaccines, Gerber baby food and nutritionals, pharmaceuticals. And so we had this pretty broad range of businesses. Then we acquired Alcon in 2009 to get even further diversified. But I think there was a realization, at least as I started to oversee R&D at the company and then eventually became CEO, I really took the view that we would be better served to focus down as a pure play company. Focus only on biopharmaceuticals.

And even within biopharmaceuticals, try to focus on therapeutic areas we think would really matter. And so what we did over the last years, we did really pretty radical rethink of the company. We exited our consumer health business, a joint venture with GSK. now it's a standalone public company. We spun off Alcon, now standalone public company. We spun off Sandoz, standalone public company. We sold our Roche steak. And so we took this company that was a conglomerate,

and we turned it into four companies. And it's not, of course, perfect math. But when you look at it, I mean, we unlocked almost $180 billion of value by doing that. So I think the story worked out. And for you, like, when you have to make that decision, you go from conglomerate to Pure Play, you have to jettison things that by definition are working, right?

You just describe these or standalone companies are doing well. How do you sort of decide where to cut and where to keep? Because when things aren't working, that's pretty easy. When things are working, it must be tempting to say, we can keep these together and we are bigger. Yeah, and, you know, the underlying logic to being diversified as a pharmaceutical company is you have these patent expires and you say, if I can get in more stable businesses,

that those businesses can kind of create a base of growth that will help you stem those ups and down. That's the thesis, right? And I think that was one reason why Novartis did this. But in actual fact, what I observed was we were misallocating capital. When you look at it, the return on capital, of course, varies tremendously between a consumer health business, a generics business, a device business, and a pharmaceuticals business. And we were often forced to say we have to sub-optimized the pharmaceutical business to invest

in the other businesses, or I'd have to say no to very good decisions for the Sandus business. to support the pharmaceuticals business. So I think there was this kind of strategic misallocation of capital. But then the second thing is you ask, what is a company fundamentally great? I mean, of artists, I think, really good at discovering and developing novel medicines. We had these other businesses.

We were investing to build them, but it was never really the core of kind of who we are. When you go to Basel, you'll see we have chemists and technologists, and that's really where I think we're most successful. So it kind of felt like the right cultural fit, right from a capital allocation. standpoint. Of course, investors, as you know, prefer that they get to decide how to diversify. They don't want us to diversify. And so for all those reasons, it felt like the right thing to do. Now, it took six years, right? Spinning off companies sounds easy in retrospect.

Actually spinning off three public companies is a massive undertaking, I can say, from experience. Well, it's obviously remarkable in the evidence in terms of where things ended up. How do you think about the pure play move today in 2026 at a moment in time when there is a lot of sort of, I would say, macro or mega trend towards consumer health. Do you think at some point it makes sense to expand back out from the core for you? Or is this, we are a medicines company and that's where we're going to sort of focus? My interesting to say this is where we need to focus, I think we can look at consumer platforms to provide access to our innovation.

But the way I think about it is there's already, and you know this well from all your investments, already so much that you need to undertake to be diversified within, the world of biomedicines. So many new technologies. And anytime you take on a new whole business area, you're going to distract the energy of the management team, the capital that you have to deploy. And also you have to learn a whole new space

that you may not actually be very good at to really manage. So I think of it as how can we maybe use these consumer platforms or direct to patient channels to augment the core. But I don't see a reason at the moment, at least, to really diversify too far away from that core biomedical effort. Makes sense. And speaking of focus on that core, you highlighted four disease areas that the company is focused on.

And then you mentioned platform technologies that you're very interested in pursuing and obviously in the service of discovering new medicines. If we can make one of these annoying sort of consultant matrices where you have the disease areas on one axis and you have platform technologies on the other, we've labeled the disease area access. How would you label the platform technology axis? Yeah, the way I think about it is obviously small molecules and biologics. are the core of every large company of our scale. You have to have that. And, of course, we've had those technologies for either 100 years,

100 plus years, or in the case of biologics, 50 years. But we focused on three core technologies. We focused on cell and gene therapies first. We were the first company to have a licensed cell therapy, the first company to have, an effect of a widely used gene therapy with zoolgensma. Second, we made a relatively significant investment in RNA therapeutics.

We've actually now acquired multiple companies, and we can definitely get to that. And then radial ligand therapies. These are the three core technology platforms that we try to apply across those core therapeutic areas, hopefully to find better solutions for patients. And we're always asking ourselves, are those the right ones? Of course, you could get into DNA editing.

You could be much bigger on antibody drug conjugates, but at least for now, those are the focus areas. Okay, so let's take, if we could, let's take each one of those on, starting with the cell and gene therapies, complex modalities, remarkable milestones and sort of patient impact that has been achieved through these. If you had to weigh in, are cell and gene therapies punching above their weight or below their weight in terms of their potential? Because there have been setbacks, there have been challenges. So how do you sort of think about

the complex modalities and where we are on the journey of being able to really exploit them? You know, this has been a tough one. I mean, I think I would say right now, punching below their potential, but I think it's going to change. And I'll explain why. First, when we talk about cell therapies, of course, the focus after our initial discoveries and then other companies, Kai Juno, then later BMS in Gilead, Bluebird also was in there, successfully bringing forward cell therapies were primarily focused on B-cell cancers, right? And I think very successful.

But the business was not hugely successful, mostly because complexity, cost of goods, profitability challenges. Now I think what's really interesting is, and I could have never predicted it, but because of the work of Georgchett in Erlangen in Germany, we've suddenly learned that you can actually use these therapies to reset the immune system. And I think this is going to be a huge opportunity.

Just remarkable. When you look at this data sets, as you know, I mean, very rare that you see data like this, very rare that you see you can rewind a disease or completely, in some cases, almost create a functional cure for patients with end-stage immunological disease. Yeah, these are the autoimmune diseases

that have been sort of very life-impacting chronic conditions. Chronic, debilitating. Dabilitating with very sort of bad options. I mean, the story I would like to tell is one of our sentinel patients was an individual in bed all the time, not able to walk his children to school, receives our cell therapy. Strong results two months, six months later,

no sign of disease has gone from almost a near-death situation to normal life. I mean, this is extraordinary. It's like Lazarus-like things. And so, you know, we've now taken that forward into four pivotal studies. We have, you know, I think six more programs in early stage clinical, other companies as well. And I think cell therapies and immunology is going to be a huge area, alongside immune reset by specifics, tri-specifics. But that is going to be a whole renaissance for cell therapy and alongside that what also happened is we figured out how to make the manufacturing factors less costly and I think that's the other part of that story if we were just doing immunology but with the same you know cost structure super expensive bespoke manufacturing but now that we're all

working on rapid platforms you know that take two days in the factory eight days vein to vein time this is changing the game and so I think there now as as often happens there was the hype, then we went through the valley, and now we're coming back out the other end. I'm not sure any of us would have predicted that would have been in immunology and not cancer, but hey, so it would be it. I think gene therapies is a tough one at the moment, as you know, with the safety. Zolgenzma has been hugely successful. We think Zol.

Remarkable drug. Amazing. I mean, the story, maybe for your listeners, this is for a terrible pediatric disease called spinal muscular atrophy. These children would face a certain death at two years of age, or if they have a less, reform. They live a life in wheelchairs and really debilitated. And so Zool Jonesmole, one-time gene therapy, if given early enough, can almost let children lead a normal life. So it's been a success story. Now, we had hoped we would have a whole series of those behind it.

It's a success story, but it's an end of one story so far. This is not platformized yet. You know, I think because a combination of factors, one, the tracking these gene therapies, to the plasmid to the right cell type, getting the expression, as you hope, safety profile, all of these things turned out to be much more of a puzzle, and there wasn't as plug-in-play as I think we initially expected. That said, I still think that for certain diseases,

if you could get a one-time therapy, be hugely beneficial, particularly in children, and I still think, you know, there's more opportunities here. We acquired some companies around this, and so let's see, but that one is still to be, to be proven, whereas I think cell therapy is really maturing now. Fantastic. So, okay, so let's shift to RNA medicines.

Why don't we give the audience a sort of a quick primer on why RNA medicines are interesting? And just to, just to like prompt you, I heard you in an interview, I think it was maybe one of the Davos interviews, where you said SIRNA, SIRNA is effectively a de-risk modality, which, by the way, for folks that have been in biotech for a long time, That's like music to the industry's ears, right? This incredible innovation has essentially been reduced to practice. And correct me if I've misquoted you. No, I think that is my view.

And I think it's an extraordinary story when you look at it. So the idea here is that you can use small sequences of RNA, a smaller chain of RNA to silence the mRNA inside the cells to prevent the protein from expressing. And that sounds like a simple thing, but actually the work goes back to the early 80s. decades of work. Al-Nylam and Ionis and other companies

worked on these technologies for a long time, struggled. And then there was an unlock around, I think, 2015, 2016, where they figured out how to traffic that RNA fragment in a relatively efficient way into the cell, have relatively high specificity, and that led to, you know,

really impressive, you know, data. So we went in primarily for cardiovascular disease at first, our first medicine is a PCSK9 SIRNA that allows you to take down PCSK9 for high cholesterol with a drug that you give every six months. And it works amazing. And right now actually is doing extremely well, particularly in China. Come back to that. But it's really amazing.

And now the opportunity is with that kind of technology, you can get to once-yearly dosing. People are working on more and more targets in cardiovascular disease. And so I personally think we're going to be. be at a place within the next five to seven years where you could get a single injection that covers HMG COA reductase, which is the statin, lipitor, statin pathway, pathway, PCSK-9 to knock down that other pathway from cholesterol, and probably, you know, if it works, L.P. L.A. another major cardiovascular factor. So one yearly injection, and think about that from a public health standpoint, it's unbelievable. But the SIRNA's story was all about liver

directed targets. So everything we were doing was targeting everything to deliver and we're seeing these impressive impressive results. And then I think what's happened now more recently is we've figured out how at least starting to see, we'll see how the data

plays out, can you traffic these SIRNAs into other parts of the body, into the muscle, into the brain, into the heart? And we recently made a proposed acquisition around this to try to see can we actually get those SIRNAs into the muscle. But this is opening up the field even more broadly, can you traffic them to the kidney. And so you can have all of these opportunities

to have less frequently administered high efficacy drugs. So I think this is a mature platform. So just to pull on that thread for a second, one of the fascinating things about all of the things you're described is that these molecules, as we'd learn to design them, eventually program them, can do pretty remarkable things. But first and foremost, you need to get them there. You need to get them to the right cells, to the right tissues. Yeah. There have been some breakthroughs on sort of what we'll call broadly delivery technology to get SIRNAs to other parts of the body. You mentioned the muscle. Things need to get into the CNS, the like and the like.

What's your view on where the next big breakthroughs in delivery technology might come from? And, you know, how do you all think about how to unlock, you know, new horizons, if you will, for where the medicines can be different? Yeah. And, you know, this is a topic not just for RNAs. also in gene therapy, as you know, as well. Yeah, I mean, that's a global comment. It's a general comment. You know, I think, so a couple of areas for us are interested. One is I think synthetic capsid design.

And this has been around, but I think we're getting better at can you synthetically create a kind of a capsule here. We'd be like a synthetic virus, you know, instead of using the AAB virus, which is often used and use that to traffic. Of course, there's been, you know, a lot of work done on certain receptors in the nervous system that can basically draw. into the central nervous system. And I think that's obviously the transferring receptor in particular, but there have been others as well. Antibodies conjugated to the S-R-N-A

or antibodies conjugated to an ASO. So then the antibody takes it to the target cell. So I think these are all in play. I think it's all maturing quickly. There's also more, I mean, all kinds of other technologies. You know, you would probably know better.

We can lipid conjugate. So we actually have, we actually acquired a company that also can lipid conjugate to get into certain cell types. So I think a lot of activity happening. And the key is, can you find the right address and can you hit that address with high specificity and not have the RNA go to all kinds of other places, right? Because depending on what you're delivering, if you don't want to hit the

heart and you only want to hit the kidney, but you still hit the heart at even small levels, you'll get the side effect profile that you don't want. That makes sense. All right, I don't want to give a short shrift to the last platform you mentioned, which is radio. So why don't you explain what radio is really quickly, and then I want to make sure we cover other platform technologies that may not be on the access right now. Yeah, absolutely. Absolutely. So I think the listeners likely know, radiation and cancer care has been around for decades.

But the idea that a company that we first acquired in 2019 called Advanced Accelerator applications is actually a spin-out from CERN, the particle accelerator outside of Geneva, the idea was could you actually bring a small radioactive particle right next to the cancer cells to basically have the cancer cells killed? Very elegant idea. It took many years again for them to actually mature, kind of a fringe idea when Novartis first acquired it, but it turns out to be very effective. And so the idea is you link basically a drug that is very specific to either the cancer cells you're trying to target

or the area around the cancer cells, you link it to a radioactive particle, and then the drug is bringing the radioactive particle near the cancer, relatively small doses, reasonably effective. And we have two drugs now. One of our drugs for prostate cancer, for a target called PSMA,

over $2 billion in sales, the other for neuroendocrine tumors approaching a billion dollars in sales, launched globally. Very challenging technology in that after you activate the drug, you have between four and five days from the factory to the patient. So the supply chain is very, very complex. It's a radioactive molecule, right? It's got a half-life and, you know, dedicated

supply chain. We've had to figure this out. It's taken us, I think, seven years, but we've now done it at scale. I mean, we have launched these products globally and are, you know, 99.9% on time on every dose, which is taking a huge effort. So once you build, once you actually build that infrastructure, the supply chain, the capability. I'm going to use the word incorrectly, but is it effectively trivial at this point to change what you target once with a radio isotope? Like in other words, is this essentially a platform technology that could go very broad, invalidated?

So for the given isotopes, so in the case of the one we've scaled lutecium, which is a beta emitter, now I think it is platformized. We just have to find, of course, and that turns out as always not as easy as it sounds was finding another target. The next target after PSMA is what we're looking for. But within our manufacturing network, yes. Now, it gets much more complicated if you want to switch to another isotope. So if you want to switch to actinium, alpha emitter, you know, there's terbium, there's kind of alpha beta mixed emitters.

And so those are much more complex because you would have to build a slightly or radically different supply chain. So those are not plug-a-play. But within Lutisham, and that's what we do now. We, I think, have eight or nine projects now in the clinic, all with lutecium tied to another target, targeting different cancer types to try to see if we can be successful. Got it. Okay. So we've covered the basic matrix right now, right? We talked about the disease areas you're interested, the platform technologies you have on hand. Two questions on the landscape as it exists today. The first one is, I assume that you obviously, you have this incredible tool to chest or war chest of modalities. And you pick the best tool for the job for any given target and any given disease indication.

There's a sort of a mix and match effect. I would imagine for radio, it's mostly cancer. But as you said, for self-therapy, it could be autoimmune disease now and the like. What have you and what has the industry learned about patient choice when it comes to modality? And how does that inform how you develop a therapeutic product profile? And just to put precision on the question, I guess,

is do we have a... a view as an industry on whether patients prefer a once-yearly injection versus a once-a-week pill. How do you think about those trade-offs and how do they factor into how you design a program? Yeah, so I think we do a lot more thinking now at Novartis, I'm sure, at other companies as well, on not just the target product profile, which is kind of what a clinician would say, who's maybe a science-based clinician, we think about the patient profile, but I'll come back to that. We even think about the kind of physician adoption profile because it's not always straightforward. What seems obvious to us when we're sitting in our company is when it actually goes into practice,

you know, is that something that a physician actually wants to deal with? And these considerations, I think these practical considerations are really, really important. I think the first lesson is there's no one-size-fits-all. You know, I think what we find is that there are segments of patients, let's say in the case of S-R-NAs who prefer a less frequent, administered medicine. Others want okay with a daily oral because they don't like an injection. These serenays are are injected. We even see geographical differences. I mentioned in Asia, there's huge interest in S RNA-based technologies for cardiovascular disease. In U.S., it's still building.

We're still working through it. I think the other thing I'd say on cancer is that, you know, we've moved away from an era where it was all about efficacy at all costs. You know, one of the interesting things about radioligin therapy and to juxtapose it to another technology antibody drug conjugates which we also are in but not as as biggest radioligin therapy the the safety profile is what often is very attractive to patients you know when they can have relatively high efficacy but a short course of radial ligin therapies are four to six usually doses that are given and then you stop and with a reasonable safety profile that's more attractive than chronically feeling sick over time. That's a shift. I mean, we used to, as you know, we used to just go out of the hammer.

We go as hard as we can, try to knock down the cancer and we don't worry about the safety proposal. That's another shift. So I think you have to think through one, there's going to be different segments for different people of different needs. And then I think you have to be, you know, comfortable. But, you know, it's interesting. With our radial ligand therapy portfolio, I'm often to ask, but there's all these orals coming. For us to be successful, if we just get 20% market share of every one of those cardiovascular categories, we have huge medicines. And because these medicines are actually in the U.S. context under the medical benefit, not the pharmacy benefit, actually it's quite economically attractive. So, you know, I think you have to be comfortable

with those ambiguities. Yeah. No, it clearly a lot of thought goes into this. As we talk about the things that aren't on the modality axis of that matrix, One of the things that's not on there, at least we haven't put it on there yet, is AI. So I'd love to take a few minutes to talk about what you see is happening in the world of AI and how you're thinking about that for Novartis. For me, it was really interesting. The industry goes every year to this J.P. Morgan Healthcare Conference in January, and this year felt like a tipping point year for AI.

And what I mean by that is there's the promise of it has been around. for a very, very long time. I think the conversation we had, you know, eight years ago, or it was seven years ago, we talked a lot about AI. So the promise was there for a long time. It felt like this year in 26 is finally when you started to see AI transact. And there were several deals announced where a large pharma company is partnering with, and I'll call it an AI first startup, which I think in and of itself is fascinating.

The other thing I thought was fascinating is that the sort of flavor of the, of deals is pretty varied, right? In some cases, it's a partnership, in some cases, it's a license, in some cases, I want to get access to a specific model. And so we're actually seeing not only technology innovation, but a little bit of business model innovation, where pharma is willing to transact on AI in a number of different ways.

So that to me was just like a fascinating thing to see in a lot of ways an exciting one, given where we sit and where we spend our time in terms of- And just a quick question. I mean, do you prefer in your investment portfolio, companies that build AI platform technologies for companies like us to apply on a research portfolios or ones that actually want to find their own drugs and try to turn into actual AI biotech. Yeah. So we actually have the full, we run the full gamut. We have the full spectrum in our portfolio.

We have, we just recently announced an investment in a company called Boltz. Bolt's bio came out of an open source project at MIT. Yeah, I know bolts well. Yeah, absolutely. Yeah. You used the model. So they just launched the company. And so they have maintained this commitment to open science. They're actually incorporated as a public benefit corporation to really, you know, make sure that they cement that point that they want to continue to release open source models. But on top of that, what they're looking to build is essentially a product, an application that companies like Novartis and others, they announced a partnership with Pfizer at their launch, can essentially use the Bolt's product to wire in the ability to have frontier AI accessible to the, to the, to

to all the scientists in their laboratories. So whether you're an industry-based lab or an academic lab, so in that case, that's an infrastructure investment. Pure platform, yeah. Pure platform investment. And, you know, our hope is that, you know, a lot of the pharmaceutical companies that are interested in integrating in AI will use that, those rails, if you will, to do that.

On the other end of the spectrum, we have several biotechnology companies that are AI first, and they're using their internal AI capabilities to develop their own pipeline of drugs. And a lot of times what that will look like is they'll partner with a company like Novartis on a program or on a discovery project. And that looks much more like traditional BD. And then you're starting to see sort of a couple of flavors in the middle. To do both. So they do a little bit of both. They're not an investment of ours, but a company called Noetic announced a partnership with GSK around JPMorgan,

where GSK is getting access to their software on a specific model to really, really be analyzed, I believe it's for lung cancer patients. So that's like a model access partnership, which is really interesting. And then you have, you know, other flavors in between. So, like, we're starting to really see a lot of this happening. And so we're trying, we believe that this is going to look like a spectrum. And so we want to have investments across that spectrum and back entrepreneurs that are pursuing different opportunities.

So that makes sense. But it's been really fun to watch. And that's why I say for me this year, 2026 was a kind of maturing, right? Yeah, exactly. So I'd be curious sort of how do you see AI as a platform? Well, let me start at the beginning. Do you see AI as a platform technology? Would you even put it on that axes?

And then my second question is, how does Voss use AI today? Yeah. So I see it as kind of an enabling technology across all of that matrix that you described. I'm not sure if I see it as its own technology platform yet. But I do see it as a powerful, powerful enabler. You know, when I think about when people always ask me, what can the impact be? I generally say in R&D, it'll take seven to 10 years because that's just how long it takes it to develop drugs. And that the things we have to be watching is how much does it shorten the overall time from when we have a target in the lab to when we get a drug license, which can be 12 to 14 years?

Can we bring that down by two to four years? And that already would have a huge impact. and can you move the probability of success number meaningfully from where we are today, you know, as you know from the research, it's 10 out of 1,000. When we get into the clinic, it's, you know, 10 out of 100. If you can move that meaningfully, even a little bit, frankly, the compounding effect is huge. So, and I think that should all be well within hand. I mean, the way we approach it with the Novartis is we break down the whole R&D value chain

and we say, what is the technologies we need? in target discovery, in candidate identification, candidate optimization, in preclinical safety, PKPD modeling, in, you know, phase one trial optimization, so on and so forth, all the way through, all the way through, you know, document, you know, work, protocol optimization, can you get automatically from the protocol to the CSR, all of, and basically try to find the right technology companies to support us across all of that. And now we're trying to integrate that much better and say, can we actually We have more of an integrated platform across.

And I think over time, I think we are now starting to see the benefit. I would agree we're at a point now where you could actually say that this will have an impact. I think the hard part is that whenever we set expectations high, people expected to happen within two years. And really, the first purely AI generated candidate, by definition, cannot come out the other end for eight to ten years. It's going to be a long time. We just don't know, right? But I think it's really interesting. I mean, we do work with, you know, isomorphic labs on what we call undruggable targets.

So we give them undruggable targets. We've not been able to drug ourselves. We think are interesting. They're working on drugging them using Alpha Fold 3. How many bolts also, you know, being a similar kind of technology that could be used for that? You have others that are using, you know, quantum technologies to say, can they better model the protein protein interactions to find these weak interactions and better drug design? And so I see all of this happening. And again, it's going to be one of those things where all of these little gains are

suddenly going to compound to a really big gain. And that's what we're going to see my expectation at the other end. Yeah. And I would agree with that. And I think that, you know, it's probably fair to say that the first AI generated drug candidate is somewhere in a pipeline, you know, for sure at the moment. So we will see it emerge or debut in a couple of years. And it sounds like from what you just said that the Novartis as an organization, has been or is in the process of adopting AI,

as being AI, there's AIification across the org. Is that a fair? Yeah, absolutely. I think it comes back to you. I also want to answer to how does, how do I use AI and I'll come back to that. But yes, we are. We have a very systematic approach,

and we're all trying to scale it now across the whole company. I think we have one of the highest numbers of AI licenses for Microsoft co-pilot with all the different model options. that you have there. We're one of the lead partners for Palantir in terms of all our R&D data goes into a data like Powered by Foundry. And so, you know, we really try that. These are all things we did like six years ago, right? And we really, it's really not. And we weren't sure how it's going to pay off. And now I have, you know, 3,000 scientists on the Foundry platform, you know, picking the right models to actually just go into our data like and extract data. They used to take

six months and now it takes minutes. I mean, the power here is, is incredible. So we are Absolutely. Absolutely scaling it. How I'm using it, and it's part of our general push. One of the, I think, big opportunities in companies like us is not the high-end biotech side of things. It's just all the ways we could just automate very rudimentary things that we do. Regulated industry, so much document management, so much review, all of this with large language models can now be done super fast. So, you know, I can create a podcast to listen to my integrated financial report to prepare for my earnings call. I mean, this kind of stuff like who would have thought it's possible or kind of we can draft a press release pretty quickly or with a quote in my voice. I have an agent, right, and you can just basically say, please look at all of my historical comments and please draft it in my voice and it'll do it pretty well.

We have agents to look at all of our historical decisions in R&D. and then as a chair of a lot of these committees, I can use those agents to inform my next decision. Whereas I'm not just relying on my memory. I can actually go into the agent. Every single time we saw a situation like this, what were the key issues to be identified? And that makes you have a better next decision.

So all of that is now happening, even on my desk every day. That's fantastic. Speaking of sitting at your desk, you've been sort of in the chair now since 2018? January of 2018, yeah. Okay, so I want to sort of cover this in two parts if we can. One is the world has changed significantly since 2018, so let's start there. And then second, I want to talk about sort of, you know, how your views of,

how the view from your desk has changed since 2018. So let's talk about the world changing. So obviously, from the last time we spoke, which was shortly after, it was a year after you were in the seat, we've had a pandemic, and there's been a big change just in the world from a job. political standpoint, also just from industry dynamics and competition. So if we could unpack a couple of things, let me throw a couple of world events your way. So the first one is on the industry dynamic side. So we in the United States have long been a leader in sort of biotechnology as an industry.

Over the course of the last several years, China has really risen as a competitor in the industry. And I think there's two things I'd love to get your take on. The first one is from where we sit, we invest in early stage companies, early stage biotechnology companies, whose ambition is to, you know, obviously innovate and find ways to partner with folks like Novartis and other leading pharmaceutical companies. In a world where the U.S. biotechnology industry was the dominant game in town, there was a sort of natural equilibrium. I'm going to oversimplify it, but there was a natural equilibrium that existed where,

the small innovator could come up with something interesting, and if it was relevant to you, we could partner, get acquired by, or find other ways to work together. China Rising has shifted that equilibrium where now large global pharma players can go and in-license things directly from China. What do you think the impact will be on the U.S.-based biotechnology industry

in a world where they are competing with a really effective Chinese industry that is fast in some ways cheaper and just incredibly productive and has scaled. Yeah, I think this has to now sharpen the competitive edge that we have in US biotech because, you know, I think what we see in trying to very talented, you know, scientists who are now leveraging a lot of know-how

they've learned over many years, also in our labs and in our company, is to move extremely quickly. And I think it started off as more fast follower, working around biologic patents. And that continues, you know, at scale. But it's slowly morphed now into, when you look at as an example like bi specifics, I mean, a huge number of bi specifics now happening in China, cell therapies and related technologies. So I think last year, at least from my understanding of the data, I think China had more licensing

or exits than U.S. biotech or somewhere close to that. So it just gives you a sense of the speed and the scale. So I think a couple of things from U.S. biotech, I think clearly now speed is of the essence, and a part of that is also going to have to be regulatory frameworks. I mean, the NMPA simply moves incredibly fast for a first in human study. I mean, it's under a month now. And that, I think, enables those companies to generate human data incredibly, incredibly fast. The NMPA being China regulator. China regulator. You know, and I think that's one and one enabler. The second, and this is, as you know, in a long-term challenge, is just the speed of trial kind of bureaucracy in the United States.

Again, China is able to leverage these very large integrated centers, the China biotechs, and they generate that data extremely fast. So they get that de-risking phase one data, and they get that proof-of-concept data, and then they can go market themselves much more quickly. So I think we have to now tackle how can you get more scale and speed in U.S. clinical trials or even European clinical trial centers as well. And Europe has historically been a bit faster. Belgium, Australia, you can go very fast.

But even Belgium and Australia are not as fast now as the way China is moving. So that's going to be a second topic for sure. And then I think, as you know, we have to continue to support NIH and the things that make the U.S. bio eat. I was really heartened to see that the Senate Finance Committee is fully supported, you know, NIH funding. Because I think that's going to be really important in the long run because that is something unique,

I think that other countries, including China, haven't fully figured out how to replicate. That ecosystem that goes from universities to biotech is incredibly unique. But those other elements, we're going to have to tackle, I think.

And so, you know, to the FDA's credit, the current administration is, I think, demonstrated that there is a very high willingness to figure out ways to do, to have the necessary reform so we can accelerate

the path to, getting these medicines to patients. So there's a will, and I think the question is, what's the right way? If you had sort of your magic wand and say, hey, here are some ideas that I think would go a very long way to getting us there, what would be on your sort of, I don't know, top two, three wish list items for reform? Well, I think a lot of it actually is making sure we have talent and capabilities and really ensure that FDA is

enabled to hire the scientists they need at scale and can then review these IND applications really, really quickly. You know, I have two minds, right? It's a tricky one because do you want to reduce the

IND requirements? Because I do think in China you do have a leaner package, I think, to get to that first in human. And is that a good thing in the long run when you need to maintain trust with society to be able to do this work? But if we just want to just optimize for speed, you could look at, you know, the size of the preclinical package. Again, I think to be debated, I would say, is that the right thing or not in the long run there.

And I think that's something we'll have to consider for sure. And then I think outside of the FDA's purview, I mean, I think we have to come up now with a more streamlined approach to IRB reviews, you know, common contracts, all of the stuff that can really slow down your ability to get a trial started up in, in the U.S. If we could get more platforms in place where you could just go right in in a particular disease without having to re-contract and redo everything,

I think you could generate the data a lot faster. And I think those kind of networks now need to get scaled if the U.S. wants to remain competitive. Yeah, I think that makes all the sense in the world. On the innovation piece, have you seen U.S. biotech

sort of respond to this competitive pressure in a meaningful way yet or in a measurable way where you're saying, okay, like folks are, you know, innovating to address, they're finding ways to innovate to address this essentially new global competitor that has emerged. Or is it too early to say?

I would say it's too early to say, and that's what say it might be happening. I may not just have visibility into it, right? But at least I have not seen it. I am sometimes concerned that the effort turns into how do we block China innovation from, you know, being accessed by global biotech companies or all of the efforts to do that.

Because I think in the long run, better to actually, you know, as you know, and the history would show us that we would want to have more competitive industries. We don't want to be able to say there's a beautiful discovery in Shanghai. Let's not use it because it was discovered in Shanghai. Right. I think it important, though, however, to make sure we have global clinical trials, clinical trials to reflect the ethnic diversity. Plainfields even, all of those things, for sure.

Yeah, for sure. Makes sense. Let's take one second to talk about access. So you're a global pharmaceutical company. You yourself care deeply about public health. One of the challenges that the industry has long had is when we innovate, we have to find ways to pay for it. And you have this remarkable example in your SMA gene therapy where you've created this incredible drug. The price tag, at least the sticker price, is north of $2 million.

That's right. And so where do you think there's opportunity to be more creative around how we ensure that we can get reimbursement for these kinds of innovations that are very expensive? And then on the other end of the spectrum, as we start to move towards a world where people are thinking about preventative medicine, things like that, where the system really isn't set up to figure out how to pay for that, have you know, have you used. seen any innovation really on how we think about reimbursement, how society pays for these innovations because the innovations are incredibly valuable? How do we essentially reflect that and how we pay for it? Yeah, I think first the baselining for the audience, I think some of the facts that get lost in this discussion, you know, about over 90% of scripts in the U.S. are generic. Yeah. Right? Generic drugs in the U.S. are the cheapest in the OECD. So, you know,

I think the system works insofar as innovation is rewarded during your IP. and then it's radically, you know, taken down in price. And it effectively becomes a public good. It effectively becomes a public good. And that's why actually the innovative medicine share of U.S. health care expenditures is actually going down, right? And net prices in the U.S. are only growing in the single digit level. We have certain population groups, now very U.S. focused at the moment,

that have a challenge in affording medicines. But actually for most people, you know, out-of-pocket costs are quite manageable. and, you know, many policies are in place to drive them, drive them down. I just wanted to say that as context. No, I think to your point, on these one-time therapies, they present a unique challenge. What's interesting is Zulgensma, even at that price point, is reimbursed in 48 countries around the world.

The reason for that is that the cost effectiveness is so high when any government around the world sees that the cost of caring for one of these patients is more like $10 million for their life. And if you actually intervene early and the child lives a normal life, to need all of that cost saving comes back to the to the system so I think that that has to be that story has to be told well now if you have many one-time therapies clearly then there's a phasing issue and then you would have to come up with new new approaches I have to say with zoolgens when we were not successful in moving to those models where you kind of phase payment out because the systems

just don't aren't ready for it and we've been talking about this forever this whole kind of idea value-based medicine will you get reimbursed if the drug doesn't work can you phase cost over time, the franchise model. Very difficult so far. Is it honest truth, right? Just very hard. It's just very hard because the systems just aren't. They're designed for annual payments and not tracking patients over time, right?

And that is reality at the moment. Now, would I get tipped if you had 10 of these? Maybe. But that's still the unknown. I think your question on the preventative therapy is the one long-term question we have to keep grappling with. from a number of different, I think, perspectives. One, these trials take a long time.

If you have a policy like the IRA in the United States that cap, you know, our ability to recoup investments at nine years, you can't run the long trial. Second, you need, I think, continued innovation on endpoints because if you're going to wait for outcomes and then you're going to do these preventative studies, you're never going to actually get to the outcome within your lap line. It'll take forever.

So we need to innovate on outcomes. And then I think you think about payment, models because now you're going to have to treat a lot of people, many of whom will never get the disease in order to prevent. So it's a number needed to treat problem. And so we're going to have to come up with new payment models. And then, of course, the system is designed to care for the sick, not to prevent. So I don't have the solution, but I do think this is something we want to look at more systematically, especially because our medicines are getting better and better at going early and preventing the cancer is from recurring. And I would,

worry that the system right now, if it doesn't start adapting, we just won't be able to systematically do that, you know, starting with the IRA, but also now, if you think about FDA regulations that require a overall survival benefit, doing that in an early stage cancer study, I mean, it's not possible, right? Absolutely. So I think there's a lot of thinking that needs to happen in each one of those areas to make preventative medicine and reality. I want to talk about your, sort of your, your tenure so far as the CEO of Novartis. And as I alluded to how the view has changed from your desk, I remember when you were relatively a fresh CEO,

there was this big focus on culture and really driving a culture of innovation, driving a culture. I remember of you talking about making sure that we don't sit in a world where we view things as not invented here in Rome. And you had a phrase. It was, I'm going to, I'm going to, unbossed. How has that, how, how, how, how, how, what is your view on the culture driven, uh, focused mandate played out? And, and how do you see it today?

I think we've come a long way. Now, there's always a long way to go when you think about culture and a huge organization like us. But, you know, when I go around the company now, this idea of inspired, curious, unbossed that inspired by purpose, curious about yourself and the outside world and empowered by the unboss philosophy. I think it's taken hold. Now, I think we keep, one of the things that's interesting about unboss I learned is it just always creates a conversation about culture. And if your organization is always talking about what is the culture we want to have, actually the culture kind of naturally improves because people are aware that they drive the culture, that this is something we all have to own. So I really see, I mean, we have all the survey data that

shows that, you know, engagement and all these things have gotten, I've gotten really to higher a place. But when I just look at the results, the innovation performance has been outstanding, you know, we're, I think, number two in TSR. We're the best non-obesity company in the sector, as I like to say, because we don't have an obesity portfolio. But there's a lot of diseases outside of obesity that also need to get treated. So I think we're doing pretty well on that front. So I think not only is, I think, do I see it, but I also see it in the performance. I see it in the R&D engine working well. So I'm pretty pleased with where we are.

Given where you sit and the view from your desk, I don't want to put you to work, but I'm going to ask you anyway, what is one company or idea that you think, I can't believe somebody hasn't started, hasn't created a startup to address it, whether it's a, you know, moonshot problem or a boring problem? Is there something where you say, wow, I'm surprised an entrepreneur hasn't created a startup to address this pain point? So there's a boring one, I think, that I always reflect on. And it's really, it's really kind of, it sounds so dull, but it affects every single trial in a company that does hundreds and hundreds of trials every year is every time we do a clinical trial, the study nurse has to record what is happening for the patient in a case report form that's, done usually on paper in the clinic.

Wow. And then has to input that into our source data. That's called the source data. Input it into our electronic case report form may have to also put information into an electronic health record and that all of this has to happen. And then we have people coming to check that the source data matches what's in the case report form. And I mean, we've been trying to innovate around this. And the fact that with all of the magic of AI,

and blockchain, and there's got to be a way to make this super streamlined for all of these clinical trial sites around the world. But with all of the EHR fragmentation, and then all the companies having different systems and all the hospitals having different, you know, documentation, it just hasn't happened. It is a pain point. And it's why we have, unfortunately, you know, a huge amount of effort to check all of that on an ongoing basis. On the bigger picture one, I mean,

I think we still have to crack, and I'm not sure it's a startup, but we still have to crack how can we use, you know, various early biomarkers for large-scale prevention? I mean, the fact that we still treat disease as late as we do when we know whether it's immunologic disease through atopic march or in certain cases or in cancer, that if you treat earlier, you're going to win. And the fact that we have not solved this problem at scale is, is the next great opportunity for us in healthcare. Well, I hope somebody listening just had a light bulb go off right over their head.

All right. So speaking of your vantage point, you're the CEO of a top 10 global biopharmaceutical company. What advice would you give to an entrepreneur to someone who's thinking about starting a company in this space, in the biotechnology space? Yeah, I think a couple things. One, I think that there's a huge value in intellectual honesty about, You know, where does your drug really fit into the treatment paradigm? Or your concept fit into the treatment program?

That would be one. And just really, I think, doing the hard work to understand that up front. I think second is to do the killer experiments. Because often when we're doing due diligence, we're looking for those killer experiments. That's in the preclinical setting to really prove that your target is doing what you say it does. Your drug is doing what you say is doing the preclinical safety work. It's tempting, right, to cut corners. on that early stuff, but often that's why we have to walk away just because we're looking for

that validation, particularly when we're going to enter in with relatively large upfronts on a drug. And my third, which is my biggest kind of pet peeve on all of these due diligence we do is all around technical development and manufacturing. The CMC work we consistently see is underinvested in. And what that means is tremendous. tremendous rework on our side, or that we just may walk away just because we can't be sure we can actually scale the production process. And is it because they didn't have the money to do the work well to do it right the first time, or they didn't know how to do it right the first time? So I suspect it's a mix of both, but I do wonder if it's a know-how. Like I've often thought, I've talked about to some of the VCs that I've met with over the years.

Should we just create like a CMC boot camp for biotech startups? So they at least know the right questions to ask. And because there's so many avoidable things that happen. And if you just did those things right, you might get an exit faster. We would certainly get the medicine to patients much more quickly than having to redo a huge amount of work to do the next trial and ultimately scale the process. So I think CMC, it seems like such a side topic that that thing you don't

want to deal with probably in your executive committee or in your board meetings, but it has such a huge impact on whether or not we can actually accelerate these innovations. And for the audience, CMC stands for Chemical Manufacturing's and Control. So this is basically the manufacturing controls and all of the work to make sure you can scale your production process from what is a lab scale process all the way up to a commercial scale process. If you set up that boot camp, you would sell it out in a week. We can do it as a joint venture. I would love that.

Thank you so much, Foss. Thank you for bringing here. Great to be here. Thank you for having me. Podcast. Thanks for listening to this episode of the A16Z podcast. If you like this episode, be sure to like, comment, subscribe, leave us a rating or review and share it with your friends and family. For more episodes, go to YouTube, Apple Podcast, and Spotify.

Follow us on X at A16Z and subscribe to our substack at A16Z.com. Thanks again for listening. and I'll see you in the next episode. As a reminder, the content here is for informational purposes only. It should not be taken as legal business, tax, or investment advice, or be used to evaluate any investment or security and is not directed at any investors or potential investors in any A16Z fund. Please note that A16Z and its affiliates may also maintain investments

in the companies discussed in this podcast. For more details, including a link to our investments, please see A16Z.com forward slash disclosures. Thank you.