Venture Studio Model: Building the Future of AI Drug Discovery | Mati Gill (3/4)

Intro - 00:00:06: Welcome to The Biotech Startups Podcast by Excedr. Join us as we speak with first-time founders, serial entrepreneurs, and experienced investors about the challenges and triumphs of running a biotech startup from pre-seed to IPO with your host, Jon Chee. In our last episode, Mati shared the simultaneous juggle of law school and ministerial politics, the Teva internship that grew into a full career, and how he learned a 45,000-person organization from the ground up, turning a company-wide financial crisis into a mandate to build Teva's first external innovation program. If you missed it, check out part two. In part three, Mati talks about the three pillars of innovation strategy that he built at Teva: academic partnerships anchored at the Weizmann Institute, a talent fellowship program, and an early bet on AI in Israeli drug R&D. He traces how Israel's bio-convergence policy, Teva's ambition, and a German venture studio model called BiomedX all converged at the right moment, and how Mati threw his hat in the ring for the CEO role of what would become ION Labs.

Mati Gill - 00:01:36: Yeah. And I think there was a couple other things we had to learn on the industry side. We had to learn that we need to have a point of contact, be the point of contact for the different scientific institutions, and not replace that person continuously because scientific institutions, research organizations, and universities, they're there for tenure and they're there for a long time. And when you're continuously coming in with new people, new roles, new initiatives, new global blah blahs, and whatever they're coming in with, that kind of creates a lack of continuity and then you create confusion.

We had a lot of times where we would talk to innovators and they said, "Well, we already talked to Teva." But when I asked them who they talked with, they talked with someone who was in a local innovation role within the Teva Israel local market, and that's not Teva R&D. So we then had to get ourselves aligned to make sure we spoke with one voice with external partners, and that was very important.

And the second thing is that it had to be a scientist because I'll let you in on a little secret: strong leaders, both in academia as well as scientists, can be a little bit of snobs. And I kind of learned that as a non-scientist, that they want to talk to me about funding opportunities, investment opportunities, want to build companies, and have me in the room there, but they don't want to talk science with me. And if I start to appear or sit in scientific meetings and try to intervene there beyond just having a basic understanding, I lose their respect instead of trying to gain it. So, hire strong scientists that can then work with them instead of trying to do the opposite.

Jon Chee - 00:03:21: I love that. And, I mean, it's very tactical and it makes a lot of sense. I think this is outside of the R&D collaboration, but every time I think about the best business partnerships I have, usually it's just one sole contact that is continuous throughout. The worst experience is when you just get continuously handed off. And you're like, "I gotta explain this from the beginning again. Like, what are we even doing here?" Like we're recreating the wheel, and that's a big pain in the butt. And so the third aspect of the external innovation group that you started, applying AI and ML to this unit. Talk a little bit about that.

Mati Gill - 00:03:55: So, honestly, it really wasn't a group that I started. It was a program that I started, a strategy that I started, led by various people from throughout the organization that did this as part of their roles, but were able to then focus attention and resources into bringing in external innovation into Teva. So the third part, which was the last one, we started to approach in sequence after putting the other two in play—both sponsored research and academic relations, and then talent play in the talent forum. The third one was, how do we seed the best capabilities in Israel of artificial intelligence and machine learning and new drug modalities, cell therapy, stem therapies, etcetera, into Teva R&D in Israel?

And this was back in 2019, 2018 when these were very nascent technologies for our industry, but understanding that Israel was uniquely positioned to be a relevant source of innovation for AI-based drug discovery and development and the government. And I could talk a lot about things the government does wrong. The Israeli government at the time did something absolutely right, which was the Israeli Innovation Authority, a very unique model here. The IIA put together a strategy that they called bio-convergence, which was basically taking a look at the pharma and biotech industry and saying, "Okay, we have all these great scientists, great science, even drugs that have come out of academic labs in Israel, like Copaxone as a Parkinson's disease treatment and seven other blockbuster drugs that have some level of IP that have come out of Israeli labs and universities. But we haven't really developed an industry, and it hasn't forward-integrated beyond Teva Pharmaceuticals. It hasn't really forward-integrated into building up an industry and startups and building an Israeli version of Kendall Square." We haven't done it.

And why is that a market failure? Because for many reasons, Israel was not competitive in the classic sense of drug discovery and development and the way we used to develop drugs. But they identified—the Israel Innovation Authority, consulting with the industry, including with us—and identified this opportunity of the new wave of innovation that we're seeing emerge now over the last five-some years of what they call bio-convergence. Which is basically the convergence of engineering and binary scientific disciplines, computational disciplines of artificial intelligence, machine learning, and other engineering-based disciplines with life sciences of biology and chemistry. And that Israel, with great science, great technology, and the capability to build companies with great products, with great entrepreneurs out of the convergence of those two, could be a source of innovation for the future of the industry.

So our interest at Teva to build up AI capabilities for R&D really came at the same time that the government was starting to build up incentives and programs in the bio-convergence program. And they came out with a call for companies to come build an innovation lab that would build up companies or do projects in the AI for pharma space. And this was back in 2020 when we knew AI would impact our industry, but we still didn't really understand how.

And so they came out with this opportunity, and at Teva, Dana, myself, and another friend, Sherry, the three of us saw this as an opportunity. And we said this is an opportunity actually for us to start to tackle that third pillar of the strategy and to seed the capabilities, because they're going to be able to provide funding. We'll be able to build an innovation lab with partners that have more experience or more data and to do things at an industry-wide scale, learn from the best throughout the world.

And so we built up a model, brought in a model from Germany called BiomedX, from an institution BiomedX. And together with their founder, Christian Tidona, really built up a venture studio model based on building startups that address the biggest challenges in the pharmaceutical industry through computational technology, and finding the best talent to be able to build those companies with at a global scale. And then found great partners. We were able to then work with on building the innovation lab from the ground up: Pfizer, AstraZeneca, and Merck, the German Merck, all joined us as pharmaceutical partners. Four pharma partners: Teva, Pfizer, AstraZeneca, and Merck. We chose Amazon as our computational and cloud technology partner and brought in a venture capital firm called the Israel Biotech Fund, and have since also added aMoon, and, you know, currently in discussions with other funds to provide the funding and the expertise and how to build and run good startups, and together built this innovation lab.

And I just fell in love with the idea, the concept, the project, and the people mostly that were behind this. I said, "You know, I think this is my next career move," and decided to throw my hat in the ring and contend for actually the CEO role. And thankfully, I was chosen.

Jon Chee - 00:09:11: Very cool. That's like a perfect culmination of you talking about the timing of the government. Like, it's kind of coalescence, like stars aligning, like the perfect timing with the government.

Mati Gill - 00:09:21: It really was. It was perfect timing on three angles: what Teva wanted to achieve, the government strategy, and then on a personal level. I went and did an executive MBA, which is usually what you do when you want to make a career change because I was becoming curious and a little bit more aspirational to knock down a few other ceilings. And was really thinking together with the family about potentially relocating to the US for three years to gain some US market experience, even entered into a new role in the US and started traveling back and forth. But then COVID hit, and those plans kind of went out the window, and this whole opportunity emerged. So the stars really aligned both on a personal, company, and then national level to be able to say, "You know what? Let me try this out."

Jon Chee - 00:10:03: That's freaking awesome. And so you had this perfect constellation of the stars aligning, and now you're leading the charge over at ION. Talk a little bit about your guys' mission, your guys' values, and what is it that you guys seek to do?

Mati Gill - 00:10:17: Right. So ION Labs is a venture studio that was built up by this amazing group of partners that all coalesced around the same mission of how do we be part of fundamentally changing the way that we discover and develop new drugs using and unlocking the capabilities of artificial intelligence, machine learning, and computational technologies of the past, present, and future. And so basically, we have three missions that we all aligned on.

Number one is developing breakthrough technologies and computational technologies for the drug discovery and development space. Number two is to do that by building and nurturing and growing great startups in a venture studio model. And number three is helping to build up and support the Israeli ecosystem and the biotech space to be a strong, thriving ecosystem that all of our shareholders want to be part of and want to tap into together.

And I think one of the core values that is the way that we do things is through a co-development model. So basically bringing together all these partners to be able to support entrepreneurs in their journey and remove as many of the hurdles possible for them in an early stage so that they can build new technologies in the right space as defined by domain experts. So that they're innovating in the right space with the right expertise in the early stages that the pharma companies provide, with the best cloud technology out there in the pharmaceutical R&D space, great investors and seed investors at an early stage, with the support of the government that helps to de-risk everything, and the best talent that we can go out there and find, and removing all those hurdles so that we can then build startups together.

Jon Chee - 00:12:08: Awesome. And I guess to set the table for anyone who's not familiar, how does a venture studio model differ from your traditional venture model?

Mati Gill - 00:12:19: Yeah. So, you know, usual venture model is you find a good technology with a team that you evaluate whether or not you want to invest in them based upon your investment thesis and the scope of where you want to invest in. You scout, you take a look at them, the opportunities flow to you, and then you evaluate the opportunity and decide to invest.

So what a venture studio model is, is systematically being a machine that can build new companies and build new deal flow, including investing in opportunities that come to us, but build new deal flow in spaces that we choose to innovate in. So the way our ION Labs model works is we have three core elements behind it. Number one is we define big problem statements and validate them that they are truly industry-grade, industry-wide problem statements that, if solved, will be the foundation of a great new startup. Number two is that we are able to locate the best team and technology out there to solve those problem statements together. And number three is that we're the right place to do that.

And I think that's the key differentiation because number one and number two, a lot of VCs do, but number three is that we have pharma partners that we ask them not only is this a good space and great team and technology, but are you willing to actually allocate the time of one of your R&D experts to work with that startup from day zero to get up and running, develop their technology. And within the first two years, validate it through a POC, through a proof of concept study that we've already agreed to do for a future technology from even before we build the startup. And if at least one of the four pharma partners that we work with says yes to that, then all three core elements align, and then we can make an investment and help build that new startup.

Jon Chee - 00:14:16: I love that. Because exactly what you said, it feels like you're getting ahead of it and solving backwards. Whereas I think a lot of the time, people just do it the opposite direction, which results in some harrowing surprises, where you're just like, "Oh, the commercial promise here is perhaps nonexistent," and you've already made the investment. And you're just like, "Oh, crap." So I really like how you've inverted that. Before even the company has started, let's talk to the people. Does this thing have legs? It's not just a conversation either. It's like, you gotta dedicate someone to give time on this, which is very differentiated.

Mati Gill - 00:14:59: Right. And imagine, we have a company called Denovi, which is in the de novo design of therapeutic antibody space, which we ideated four years ago, built the company a little over three years ago and launched it. And this company now has direct access to the heads of protein engineering of four pharmaceutical companies around the table IDing under very strict antitrust guidelines, thinking together what the future of protein design is going to look like using AI tools, and then helping that startup get up and running and doing proof of concept studies with all of those from day one.

And that's basically the model we decided to develop. And we could start with a problem statement. We can start with a technology, and we can even start with a team that has neither of the two, but we so much believe in that innovator that we can hire him as an EIR, as an entrepreneur in residence, and help him to come up with an idea and technology that then we bring to our scientific investment committee and decide to invest in. And we've done things through all three angles, starting with problem, team, or technology. But always, whether it's a problem statement looking for a technology or a technology looking for a problem to solve, or a great entrepreneur looking to do both, we always make sure that all three core elements always exist: big problem, technology and team, and we being the right place.

Jon Chee - 00:16:27: Very cool. And when you were first standing this model up, was all three pillars already just like, "This is our north star out the gate," or was this something that you had to kind of develop and iterate over time?

Mati Gill - 00:16:41: Yeah. So we were very fortunate. Dana introduced me to Christian Tidona from BiomedX, and he was at the time not building startups, but building research groups using the same core values and core professional values in the model. And knowing that model, when we learned it, we said, "Okay, how do we take these same elements and adapt them?" And we had to make changes, but how do we adapt them to be the same foundation for a startup venture studio model?

And together with Christian—and I give him a lot of credit, he's German, but he has some Italian ancestry and some Israeli cultural heritage, I guess, or person from Germany—we really worked with him to adapt and fine-tune his model together with all of our partners and the people from our partners to adapt that model to what could work as a venture studio model. And we're a learning organization, so we remain flexible and we learn from what works. We learn from what doesn't work, especially, and try to not make the same mistake twice and continuously adapt and improve on our model going forward.

Jon Chee - 00:17:55: Very cool. So just to double-click on each of the pillars, what works and doesn't work from the problem statement perspective?

Mati Gill - 00:18:03: So from the problem statement perspective, we always want to make sure that someone's going to pay for it, but primarily that it's a big enough problem that could be really the foundation for a future startup. And sometimes we come and get pitched with ideas for problem statements that might be big pain points, but they're too small within the process to actually be the foundation of a startup. It's a nice research project, or it can be part of a startup, but not in its own right. I think the size of the problem and the magnitude of the problem was one area that we had to learn through.

The other one was how do we test out the technology and get alignment on how to test the technology once it's built and to put that in through the problem statement stage. So not develop a problem statement, build a technology, and then think together how we're going to test it out, but actually make that one of the core questions we ask at the problem definition stage.

Jon Chee - 00:19:02: Question, if you're to attach a number, if you can, what is a big enough problem and what's a big enough market?

Mati Gill - 00:19:09: It has to be something that venture capital investors would be interested to solve, meaning that it's going to be able to potentially be billions of dollars of value in a company if built. It can't be a small company.

Jon Chee - 00:19:23: Yeah. And that's, I think, sometimes what I see when I talk to grad students, it's kind of just like a carryover from their grad studies. Not necessarily saying that whatever you study doesn't turn into these multibillion-dollar outcomes. But I think for anyone who's thinking about that, you gotta go big.

Mati Gill - 00:19:45: Right. So in the pharma world, if you're able to use the technology to develop a drug—we only deal at ION Labs because we have four competitors working together, four pharma companies working together, we never build a company that's going to do a single molecule. It's always platform technologies that will develop products. The product could be a pipeline asset, could be a pipeline itself and then drugs that come out of the pipeline, or the product could be a tool for how to navigate successfully the way that we conduct clinical trials to be able to reduce the attrition rate in clinical development from 90 percent failure rates to something much less. And that has billions of dollars of value for the industry across that timeline as well.

And drugs, of course, do as well when they're successful, especially when you're a platform technology that can spin out multiple drugs. So in all those cases, we're going to be able to develop a product, but we have to make sure that the problem is wide enough and we can envision what that future product is. Even if we don't know, because we're in a new industry, we don't necessarily always know what that business model would exactly look like because it's an evolving area. And there's great companies that don't bring new drugs to the market but still have tremendous value and are able to raise big rounds and strike good deals with pharma companies. But we always have to be able to envision that it's going to be big enough that if solved will be the basis of a company and not just a pet project of a startup or even a large company.

Another question we ask at the problem definition stage, we say, "If this is something as a pharma company you're going to try to do on your own, we're not the right place to actually build the technology for that. Only try to bring to us crazy problem statement ideas that you're not going to do internally." And then you can use us as a sandbox to test out new ideas that'll become relevant for five and ten years now.

So again, we go back to problem statements. Five years ago, when we started at ION Labs, AI was not everywhere in our industry, the hype was just starting. The idea was just starting. We didn't know exactly how it would impact our industry. It was very new. So we identified that and built ION Labs around that idea. But now as we enter into our second five-year period, we're starting to ask a question, "Okay, what's of the future?" So there's areas where AI will continuously innovate, but what's next? And that's really one of the questions we start asking ourselves now. It could be quantum theory and technology, could be new biological mechanisms, new chemistry mechanisms, etcetera, or preclinical studies and the way that we replace those, which is still very unsolved. But, you know, that's a question we're asking ourselves now.

Jon Chee - 00:22:28: Interesting. And you brought up a platform technology versus just having a singular asset. And I guess I kind of see the pendulum always kind of shift between going platform or getting a de-risked asset and then just taking it all the way and then flipping it. What is your philosophy? We kind of see it now where someone just licenses in one drug candidate that's as de-risked as possible and then flips it versus going for something bigger, whereas a platform play. How do you think about that?

Mati Gill - 00:22:58: So, you know, you have to build a hybrid platform technology that has products. So if it's in drug discovery, you're going to want a platform technology that could spin out multiple products and then be able to have several of those within your pipeline strategy advancing at various stages in clinical stages or preclinical stages in parallel. So the best companies at a platform level have some that they are spinning out at early stages and licensing out and then co-developing with pharma biotech companies, and some that they maintain internally to develop all the way, usually being able to do both with enough resources.

Jon Chee - 00:23:36: Absolutely. And you talked about the second pillar being the technology. What are you seeing work and not work at that stage?

Mati Gill - 00:23:43: So there, one of our key learnings was from the first company that we shut down. It was a company in the preclinical space, two great artificial intelligence AI entrepreneurs from the best AI company that Israel's built to date called Mobileye, which is in the autonomous driving space that these two guys developed some of the technologies that are in the cars that we all drive, including the Waymos all throughout San Francisco. And they were great people, which for me is always a core value, only working with good people, but great AI entrepreneurs and real visionaries in the area.

They had one thing missing. They wanted to build a company in the drug discovery and development space, specifically addressing a question of preclinical assessment of what drug candidate should go into the clinic, and they didn't have a biologist on the team. And they were very resistant even because they didn't understand the value because it came from outside the industry. And this is where I take responsibility. I said, "This is our role at ION Labs to actually make sure." And we tried and we encouraged them and we tried to convince them to hire a biologist as their first hire. And they said, "Yes, we will, but we'll get there in a year from now."

And what happened within that first year was they became so frustrated because they didn't really understand the language of the customers and the partners that they were working with because they were not biologists. And by the time they got around to hiring someone with a part-time employee that was a great biologist, we just sat down in the same room I'm speaking with you from and decided to shut down the company together because it wasn't advancing to a stage.

But then when we went to build a couple other companies of ours, like Protai and Renaisys that are within our portfolio, that again were built by great AI-based or computational-based and mathematical-based technologists, we said, "We're not signing the investment agreement until you either commit or identify a top employee that speaks the language of the customers that you're going to innovate for." So in the case of Protai in the small molecule space, we said we want a computational chemist, and they found one and she's their third co-founder. In the case of Renaisys in the RNA space, they hired a top biologist as their first key employee. And so we try to learn from those same mistakes on the team and, bottom line, it was ultimately becoming a multidisciplinary team. And we also wouldn't hire a team or invest in a team in the AI for pharma space that were only biologists. So again, it's a multidisciplinary team that can work across those two scientific disciplines and also raise capital and continue to innovate from.

Jon Chee - 00:26:27: That makes total sense. I see it sometimes having grown up in the Bay Area where you're almost solving the problem, but you're not intimately aware of that industry's cultural norms, the language that they're speaking. You gotta have someone who can be kind of that insider, and you gotta be able to have someone who can actually feel the pain, properly feel the pain firsthand, which gives you that insight.

Mati Gill - 00:26:52: And it's a different language, right? It's a different language and skill set. So, you know, I'm blessed to have great partners and board members on behalf of our partners that really accept making mistakes, learning from them, and moving on. So at a lot of board meetings of ours, we prepare slides just of lessons learned. What have we learned the last quarter and how are we going to build on these learnings in order to make new mistakes in the future, but not the same ones? And that's the only way you can really build and grow. And you see that, and that's not specific to what we do here, in any organization. You see that in politics, you see that in any organization that I've worked in or worked for and become expert in, even in sports. Right. We learn how to attack defenses in basketball in a different manner. Then it doesn't work. And that's why sometimes in seven-game series, you see the Timberwolves coming out just last week and beating San Antonio in game one, and then San Antonio obliterating them in game two. Because they're able to learn from how they're playing and then adapt it for game two.

Jon Chee - 00:27:57: Absolutely. And taking that to the third pillar for you guys, the pharma partnerships, what are your learnings there?

Mati Gill - 00:28:05: So there, it was a couple things I would say. Number one is making sure we have clear alignment with a key person within the organization that can sign off on doing a proof of concept study before we build the startup. The capital is the least important part of that. If the pharma partners are willing to put that in, the most expensive thing for them to allocate for the startup is the time of their R&D experts. And there, they really need to be excited about the science in order to allocate the time. And it needs to be strategic for the organization in that sense so that they can actually get resources. So therefore, if they then bring a target to a new hit design startup or new de novo design therapeutic antibody startup, then that company will be able to then come up with a new molecule and that will be the validation study. And they need to be able to sign off on a conceptual POC like that even before we build the startup. So that's learning number one and we get that sign off.

Learning number two, I would say there is making sure the contracts happen operationally very quickly. And it seems trivial, but if you're a startup with limited runway and you're going to build a startup and have to take a decision within a year and a half to two years on whether or not you're viable for continuous investment and you're reliant on a model that offers a validation for your technology through POCs with these pharma partners, then what we do is we get the startup signed off on master service agreements with pharma partners from day one. And have built up a cloud environment that is preapproved by the data offices of these pharma companies so that these heavily bureaucratic legal agreements and data offices to be able to approve the cloud environments are as simplified and streamlined as possible. So we can actually get them up and running to be able to conduct the POC operationally by the time the technology is ready for it. So they don't have to wait for it then. And we start them out on day one as they're building up the technology in parallel.

Jon Chee - 00:30:16: That's really clever. I can imagine it's overlooked because I was just talking to some colleagues of mine that are on that journey of trying to get the large pharma partnership, and it is taking way longer than I anticipated. And exactly what you said, you hit the nail on the head. There's finite runway. Time is of the essence, and I love again that it's started from the beginning. Like, let's not burn our dollars waiting for the bureaucratic machine to figure it out.

Mati Gill - 00:30:49: So our colleagues from AstraZeneca and Pfizer, they came to us with templates for sponsored research or master service agreements each. And this is where my legal background adds some value here again. And they say, "Okay, here's these big templates for these types of agreements." I tell the startup entrepreneurs, "You're not going to negotiate these. You can't change a letter within these agreements."

Jon Chee - 00:31:12: Completely.

Mati Gill - 00:31:14: Pre-approved FCPA clauses, data privacy clauses. Don't even try.

Jon Chee - 00:31:21: Yeah.

Mati Gill - 00:31:22: The one thing I did was make sure we signed off on the template, and there we actually made some changes, but once the template is locked down, the startups have no capability to change the template and they just have to trust that it's fair. But these large 50-page agreements sometimes say little to nothing on tangible work.

But what it does for a startup, the value it brings to a startup, it puts them in the system, or gets them into the system locked and loaded so that when they're ready with the technology to do a study or a collaboration or a partnership, or actually develop a pipeline asset, and it could be huge deals or it could be very small deals, the master service agreement is already signed. They then just need an appendix, which could take a lot less time to just commercially negotiate and bang out the terms and sign it. And you're already approved as a vendor in the system.

And it seems trivial, right? We all like to do scientific innovation and breakthrough things that can potentially even win Nobel prizes. But sometimes those get stuck when you don't have the legal agreement and the operational framework set up, and you can't accept the data that the pharma company wants to share with you because they don't have the confidence that it's going to be protected in accordance with all best practices for data privacy and federated in a manner that they all want to share. All our four pharma partners want to share their data potentially with the startup, but not with each other. They're not allowed to do that and they're competitors. So they can't do that.

So what Amazon did with us, what Amazon Web Services did with us and with our CTO specifically is actually build up this cloud environment, having direct proximity to the data offices of the pharma companies, understanding what the requirements are, and then building a data environment that they could then review one time for all startups. And then once we tell the Pfizer data office, as an example, or AstraZeneca data office, that they're sitting on the AWS 10X cloud environment of ION Labs, they don't need to start the review process every time from the beginning, and it shortens the timeline for the startup, and then they're able to get data.

Jon Chee - 00:33:33: I love that. And I geek over these kinds of things, honestly.

Mati Gill - 00:33:37: So do I, as you can see.

Jon Chee - 00:33:39: Yeah. I seriously geek over these innovations because they're true innovations. Exactly what you said. My colleagues at the bench, they're like, "I'm looking for the glory of the Nobel Prize," but there are things like this that will easily impede that, and it would be such a shame to have technology that has so much promise to just get hung up on things that you probably don't think are sexy. It's not sexy.

Mati Gill - 00:34:09: Yeah. There's scientific and technological innovation. There's business model innovation sometimes. Like, I always like to say Teva grew out of Israeli scientific innovation, specifically with Copaxone from the Weizmann Institute, but also business model innovation by pioneering the generic drug industry out of a place of need. And it probably wouldn't have been able to be pioneered by another company if it didn't have Israeli roots, because as an Israeli company, they had developed the capability to develop medicines for companies that for various reasons did not bring their drugs to Israel. So they developed those capabilities to then develop those and replicate those same innovative medicines for the Israeli market at a lower scale.

So when the Hatch-Waxman Act came out in the United States in 1984, back when the United States Congress actually did bipartisan legislation—seems like a hundred years ago—

Jon Chee - 00:35:06: Yeah.

Mati Gill - 00:35:07: But when they actually did that with incentives, that's where Israeli risk-takers and innovators, specifically at Teva, both from the legal side and the IP side, including my former boss and mentor, Rich, and his replacement, David, they saw this as an opportunity to take legal risks, calculate it, and develop drugs that ultimately were generic drugs, but replicating the innovative drugs, knowing that we had developed these opportunities in Israel over a couple decades to develop drugs that other companies innovated in a very streamlined manner, an industrialized manner. So they pioneered an industry's business model innovation, as well as operational innovation and process innovation, like you said very articulately. So, yeah, both of those elements, and I geek out about this just as much as you. It's awesome. It drives value.

Jon Chee - 00:36:01: Absolutely. And I guess a parallel in my head, you could push back if it's a good parallel or not, it's kind of how Y Combinator popularized the SAFE. Because you're like, "Oh, we're just trying to get these companies started, but we don't wanna do a priced round," so let's just create the SAFE.

Mati Gill - 00:36:20: Yeah. It's like streamlining a process that you psychologically allow a startup and the investor to build up a new startup, but it's more of a psychological innovation than anything else. You're right. Great example.

Jon Chee - 00:36:33: Obviously, that's an innovation that you're honing in Israel. I'm like, "Damn. This could be awesome in America too." You know? And I know large pharma, they want to work with startup innovation, broadly speaking, but startups just really don't realize how big of a lift it is outside of the science. Like, it's freaking massive, but why are we recreating the wheel? Like you said, we have a template that is approved. It is good to go.

Mati Gill - 00:37:05: Yeah. It requires having a deep understanding in the domain and what really matters, and we learn from it. We didn't know everything, and we still don't know everything. Having a good team, so good legal counsel. In this case, for us, it's external counsel, having good finance people, having good scientists and understanding the processes, as well as good entrepreneurs that we choose that are able to articulate their needs, and especially having good partners and having access to all of their offices, to their legal counsel, their data offices, and as well as their obviously R&D groups that have opened up this access for us to enable us to instill these operational processes and a lot of hustling.

Jon Chee - 00:37:49: For sure. And I think what I get frustrated by when I think about the venture model, just in life sciences at large, is that it's like a capital furnace. Like, it really is. And I'm like, it doesn't have to be like this. Your kind of innovation on this side is a way to let's not just incinerate capital if—

Mati Gill - 00:38:12: That's small money. The big money in drug discovery and development is actually in the scientific experiments and the clinical development, and that's where the furnace actually—

Jon Chee - 00:38:22: Yeah. Yeah.

Mati Gill - 00:38:23: Burns are. So we're doing savings mostly in times that we can actually reach a point of validation of the technology and conviction in the technology as quickly as possible so that these entrepreneurs can then raise their next rounds, or if it's not working, shut it down.

So we built a company at ION Labs called Combinable. It was in a space of optimization of antibodies. At the time, we thought it was going to be a big idea, multi-objective optimization of antibodies. And when we started to build Combinable, they were probably one of the first companies in that space very quickly, because this is an area of optimization. You know, optimization is an area where machine learning and artificial intelligence is very automated. And people that come in with an engineering background, that's kind of their sixth sense. So how do we optimize something? So a lot of machine learning-based entrepreneurs went into the space of optimization of processes and antibody and large molecules, small molecules was a place for them to innovate in.

So we built this company back in 2022, 2023, when it started to ideate and then get off the ground running. It was a very open space, slowly became consolidated. But again, because they had such a great multidisciplinary scientific and technological team, four people working together, understanding the language of machine learning, including those that wrote books, literally, on artificial intelligence and machine learning. The CTO wrote a book on the basics and fundamentals of AI and how they could be applied to healthcare, and with a great structural biologist out of one of the best labs in the space from Germany. And they really understood both languages, and those four people with $1,000,000 under two years built the best technology in that space that was benchmarked in objective tests globally. And there's several companies out there that are great companies, all innovating. This one was the best, the best scientifically and technologically. And I say that not lightly because I saw the results from an objective benchmark from the leading company in that space that tested them all out.

And then they were acquired relatively very quickly when they were seeking a new home, because it wasn't a broad enough problem statement in its own right to build up a company. And they then either had to pivot and start developing new pipeline assets, etcetera, and going more the biotech route, or look for a new home as an acquisition. And InVitro Medicine that wanted to build up their own internal large molecule machine learning-based platform for large molecule development, tested out various potential partners, chose this one, this team, and bought them. Again, under two years, innovating very quickly because they were able to develop these technologies with the expertise, with access to the expertise of domain experts, both from our pharma partners, as well as from us, to be able to develop the technology that would be best in class very quickly.

Jon Chee - 00:41:30: That's wild. And I guess from the venture model, do you think there's gonna be a shift where things are moving really quickly? I can't even keep up anymore. Do you think there's just gonna be a wave of quickly innovating, getting the technology validated, and then just finding a home, and then just kind of like a platform almost, where you're just spinning these things out and they find permanent homes? Or what are you seeing at the pace of innovation?

Mati Gill - 00:41:58: Yeah. So the pace of artificial intelligence innovation for pharma is very quick. It's very fast. Five years ago, it was very innovative. Now it's all integrated, AI into drug discovery and development. So what we're seeing right now is either companies that are going to be able to use and unlock value in pipeline development through platform technologies, and then to be able to do it at a rapid pace, a little bit faster than in classical methods. And then in the future, a lot faster and hopefully to be able to—and this is most important, I would say—to develop drugs for previously undruggable targets.

So the same in central medicine, they're trying to tackle ALS, the holy grail of our industry. I mean, that's God's work. Hopefully, they'll be able to do that. We all know people that have passed away from ALS. So I have great admiration for their mission that they're taking on in one of the drug programs. It's not the only one. They're not putting all their eggs in that basket, but I mean, Godspeed. Hopefully, they'll solve this and hopefully artificial intelligence will be able to unlock the secrets behind Lou Gehrig's disease that we named over a hundred years ago, when the famous Lou Gehrig passed away from it and had to sit out his first baseball game.

So we're seeing the capabilities of AI be applied to scientific innovation in a platform for product development stage for pipeline asset stage and drugs, to be able to bring out and churn out new drugs, but also to be able to drug the previously undruggable inshallah, hopefully that'll work.

But we're also seeing that process innovation, to go to our previous discussion, that it's not just about scientific and technological asset development. It's also about process development, and there we're seeing great companies. Just to give you an example, let's say PhaseV. It's a great company that is helping to navigate the way that we develop drugs throughout clinical trials. And it's built by a company that says there's a 90% attrition rate. AI and machine learning is supposed to help us predict, and it's supposed to help us to navigate. And the CTO there was on a company called Via that does navigation systems for public transportation. And he was working with a neuroscientist as a CEO, and they have actually developed a technological platform to be able to help lower the attrition rate for pharma companies throughout clinical development.

And there's huge hundreds of millions of dollars per drug in unlocked value that we can then capture as an industry. If we're able to build companies that can lower the time and speed and costs, or lower the time, increase the speed, lower the costs and improve the efficiency rates, to be able to bring new drugs to the market. And there's huge value just in these technological capabilities for processes.

And so we're going to see entrepreneurs coming in in both angles, both to be able to bring drugs to the market by using AI, but also to be able to bring a new technological process innovation that can cut things in half, and it'll probably be even lower than that. And that's where you're going to see innovation come in. Some of these startups will be bought and acquired and exit. Some will then sell off certain assets in licensing deals, and some will go public and become mid- to large-sized companies in their own rights, and some will fail and shut down.

Outro - 00:45:32:

That's all for this episode of The Biotech Startups Podcast featuring Mati Gill. Join us next time for part four where Mati breaks down ION Labs' venture studio model, defining industry-grade problem statements, building multidisciplinary teams, and securing pharma partner POC commitments before a startup is even built, along with the hard lessons from companies shut down and the portfolio company benchmarked as best-in-class globally and acquired by AstraZeneca in under two years. If you enjoy the show, subscribe, leave a review, or share it with a friend. Thanks for listening, and see you next time.

The Biotech Startups Podcast is produced by Excedr. Don't want to miss an episode? Search for The Biotech Startups Podcast wherever you get your podcasts, and click subscribe. Excedr provides research labs with equipment leases on founder-friendly terms to support paths to exceptional outcomes. To learn more, visit our website, www.excedr.com. On behalf of the team here at Excedr, thanks for listening.

The Biotech Startups Podcast provides general insights into the life science sector through the experiences of its guests. The use of information on this podcast or materials linked from the podcast is at the user's own risk. The views expressed by the participants are their own and are not the views of Excedr or sponsors. No reference to any product, service, or company in the podcast is an endorsement by Excedr or its guests.