BioSpace: My name is Jennifer Smith Parker, Director of Insights at Biospace, and you're listening to Denatured. In this episode, I'll be speaking with Vishwa Kholaru, CEO and founder of Invada, and Akshay Ray, Principal Healthcare and Biotech Investments at Premji Invest. We discuss how artificial intelligence is reshaping the search for new medicines out of what investors actually look for when evaluating AI biotech platforms. Vishwa and Akshay, thank you so much for joining this episode of Denatured. If you can please introduce yourselves to our audience, that would be great. Vishwa, can we start with you? Sure. Vishwa Kolaru, founder and CEO of Enveda. We are a clinical stage private biotech company that's discovering new chemistry from living systems and turning them into first-in-class medicines. Perfect. Akshay, please go ahead. Shreeray, Lead Healthcare and Last Night's Investments at TransJain West. We are a medium ball of fun. focused on technology and healthcare investments in the energy growth space right from early stage. Okay. Great. Thank you so much. So for this topic, I mean, talking here about AI platforms and differentiation, I think let's start here with your reflection. And if you can talk a bit about how would you describe Premji Invest's current investment thesis in biotech and healthcare, and where does AI native drug discovery fit within that? So investment thesis is largely built on backing durable long-term companies building generational models. AI native discovery is very compelling to us because we've seen a lot of evidence that it does compress early cycles for discovery. But at the same time, to create durable value, what we need to see is that cycle time operation also translates into actual biological insights, which lead to real drugs. As Vishwa keeps reminding me, eventually biotechs are judged by how many young people produce. And I think that part is why we are still looking for more evidence, but it's a very competitive space. Yeah, very much a competitive space. So let me ask you, Vishwa, about that. You know, it's one of the companies that are the recipients from funding from JetPrivate. Tell me a bit about your company, what you're doing. what the future is. I think that's a lot rolled into one, so I'll try my best. It's a couple. Very simply, know, Envida was born out of the realization that it looked like the entirety of our solution space that we were searching to try and solve the translation problem in drug discovery, which, know, in simple words is things that work in the lab don't work in people. was occupied by biological ideas and biological technologies. And I couldn't square this circle in my head with the fact that we had many of our medicines be discovered or invented depending on the exact medicine before we had biology labs or even molecular biology as a formal field of investigation. And it felt to me that if biology and chemistry are two sides of the same coin or two sides of the same drug, if you will, that it really begs the question, should we not focus on the best chemistry? And that was the insight that led to us realizing that at the end of the day, one can start the drug discovery process by having a biological hypothesis in hand, or like, you know, the drug discovery systems of old, like we learned from rapamycin, aspirin, artemisinin, you know, chemical discoveries that on average, I think yielded a Nobel prize every other year. and let that teach us the biology. So that is the core thesis for the company. And so we inevitably end up with first in class medicines, where complex, unique chemistry needs to love a biology. And the bet was that this novel biology will yield differentiated medicines. So what was the bottleneck? The bottleneck was actually a few things. I would say three main things. One is where to look within nature. Two is how to look. within nature, so annotating the structure and function of net new molecules. And three is once you figure out what these molecules do and their potential for a drug, identifying the precise molecular mechanism of action so that they can be developed with the same rigor as, you know, biologically originated synthetic compound. And for us, it felt like the 21st century was a great time to give this old but validated yet unfinished idea another try. And here we are applying AI to the world's largest data sets of mass spectrometry to try and do to chemical code what next generation sequencing did to DNA. Fascinating. Fascinating. Okay. So as you're saying, there's so much potential for what can be done here, and especially within the natural space where a lot of other companies have shied away or have decided not to be going into this area. So let me ask you actually, from your vantage point, are we in an AI drug discovery bubble or are we still early in the value creation curve? and what metrics do you watch to decide? Yeah, thanks for that question. think it's a, there is clearly a lot of excitement around here. We've seen now a lot of activity, even over the last 12 months. But that doesn't necessarily mean there's a problem. I think there's a lot of validation points that come, which have gotten people excited about how AI actually uses drug discovery. But when you look at what the most sophisticated investors, buyers are doing, how much deals are happening, there's a decent bit of discipline around that. What you're seeing is a lot of gating towards microphones and proof points for how capital is being bought. So I think that from that point there is hype, there is excitement, but there are also good symbols to say that some of this is proven now. And what you're waiting for is I think the translation part of it to convert these signal structural drugs. Yeah, I think ⁓ in neither in a bubble ⁓ The excitement is justified, I would say, from the early stage perspective, whether we have really new age biotechs, we have to keep going. We are excited to see that happen. So let me ask you two just a bit about that from Vishwa, from your perspective. If you had to explain the edge to a skeptical pharma R &D leader, let's say in one slide, what would be on it and what would you leave out? You know, I think it would be precisely the... metrics that convinced Akshay, it would be that the proof is in the proverbial pudding, but the proverbial pudding is the pipeline in Biotech. I think in four years from our series A financing, we became a clinical stage company. Actually three years from our series A financing, we became a clinical stage company. And five years on, we have three assets in the clinic, number that I think we will continue to increase in the very short order here. And it's important that not just we have medicines. and that they are in big markets, which is a choice that we've made is attack big markets and still show that we'd be able to beat out the competition. It is the fact that these medicines could not have been discovered with an alternate thesis or an alternate technology. So you have real drugs that show that they can potentially deliver differentiated value in the most competitive spaces in the market. For example, our lead asset, ENV294, we just released phase 1b. moderate to severe atopic dermatitis data. And we've seen some of the best signals numerically of any drug, both on the efficacy and safety side, both in development and in approval. And we are set to repeat that in hopefully obesity, IBD and liver disease in the near future. So that's number two. And number three, that these drugs are fully originated from the platform. The thesis of the drugs is intrinsic to why we started the company and we couldn't have gotten them any other way. So it'd be one slide with the drugs we made and the clinical data that we got. So how did you decide which indications though were the best proving grounds for the platform? Because what I'm hearing from both of you, and I've heard this many times, is that just saying that you have an amazing AI drug discovery platform is not enough. You need to be having the proof in the pudding. So for yourself, Ashwag, why decide on those three indications here to move forward? For us, it was actually really simple. It is that if at the end of the day, if you abstract away the team, the thesis, the technology, the black or white box, if you will, of what every founder and CEO is pitching to VC. The pitch at the end of the day is that we can make better medicines slash faster slash cheaper now in the AI era, but the real piece is can we make better medicines that shift not just the floor, but the ceiling and have a better chance of bridging the translational chasm. Right? So if you say that is the... that is essentially the promise, meta promise that every team is making while they seek capital, then the question is how is that promise best proven out? And we decided to set the very highest bar for the company, especially as the sentiment around some rare diseases, et cetera, changed in the early part of the company and pivoted to going after the biggest indications where there's the most amount of commercial and discovery attention to prove. that we can shift the ceiling in the most competitive spaces. And so far, we've been extremely excited. Early on, it was intrigue, and now it is impatience by what we've been able to find. And so I think over the next couple of years, we have over a dozen clinical signals that we're excited to read out ⁓ and hopefully repeat what happened with our first asset. Yeah, yeah, very much so. I mean, when we talk about All these platforms out there, what I referred to at the beginning, and you have dozens of AI drug discovery enabled in biology companies. So let me ask you, Ash Kaksha, is when you look at a new platform, what are three or four non-negotiable signals that tell you this is real versus, as you said, there's hype? So you go back to what we supposed to have tried. You look for small signals which kind of talk about platform improvement and the platform of learning. What we really focus on is when we did the first And last one, I know that it was more about how good is the platform in terms of learning from what kids can do. So we look for like quantifiable metrics where how is time to DC shrunk over time? How good is the platform at predicting? Not just learning, we also look for like prospector validation. It's easy to look at a data set and force fit your answers to confirm. But... where I would say even in the data platform really shine was they were really good at predicting structures and over time that prediction actually improved. And we could see that over a 10, 12, 24 month period, they were able to predict structures better and then they could actually go put it through an NMR, look at the structure and say match it, how close they were in terms of that's a very good thing. The third thing is again, goes back to biology, a good translation part, either validated biology or good models to validate that the targets you're going after are the right ones. Be it good animal models, year, silicone models, but just being able to prove that the targets are going after the right ones. And this will kind of translate well into kitmuses. and eventually I think it's going to be less about models and more about proprietary data. How much do you kind of keep investing and building that more data on the proprietary data and that proprietary data is not. just this combination of a lab and this data that we bring. So yeah, if I think about it, like three or four signals, learning that the platform improves over time. It has real predictive power. It can translate what is software output to actual translate and create a modern on. I would say those four things probably are important. Yeah. That does make a lot of sense. I think just to continue that thought a bit, just because you look at the metrics that determine that investment decisions. So on that point, what are the most common strategic mistakes you see AI platform founders make in the first three to five years? And how do you try to them away from those? You may also have some insight upon this too as a company that's doing very well. Yeah, I think sometimes we need to rein back enthusiasm from founders. There is like so much to do that you can do too many things at the same time. But you need to sometimes prioritize going after too many therapeutic areas, too many candidates versus generating clear signals that provide validation and medication for the next set of investors. Second is probably confusing models versus data. You can have great test models, but the way we seeing outside healthcare in a sense is that models eventually catch up. So like I said, it's more about creating that data modes which help you be the leader in Yarsper's versus just thinking about the most sophisticated model that you can build. That's what I would say. I think they underestimate the need for traffic. think today in my stuff are very, very focused on how does you can create that. think nobody really questions to the top of the funnel, it will expand. You will just have more candidates to look at and evaluate, but that doesn't mean that they all are transmitting to the point. I think translation is a barrier where lot of AI markets will still kind of make a break. And being thoughtful about partnerships. So I'll probably talk about this so that we might have many chances to kind of maybe partner with our company. But at the same time, we've been very thoughtful about how we think about it, how we monetize it. There is obviously the failure of non-debate or capital, but at the same time, collaboration partnerships need a lot of thinking and resources in those two. really determines how you think about them. Absolutely. I think that was going to be one of my primary points is you have to decide the business model you're going to try and follow. And it's far too easy in some instances, especially if your thesis or technology is a shoe in for where Pharma is or where Pharma is going in the short term to do a number of partnerships. And I think that if you want to be a multi pharma partnership, platform company that has a little bit in every drug and is so successful that you overcome all of the usual barriers of, you know, long time cycles, low royalty returns, no control over whether pharma will ultimately prioritize. Then that's great, but you have to be very deliberate about understanding that that is what you're going after. And I think mapping out your path to venture scale returns. But I think the one business model that does have. I think all of the data, literally maybe 100 % of the data going for it is being a drug company that has drugs. So if you're, if you're choose that, you know, at the end of the day, you're going to make drugs with your technology, then you must understand that any partnerships you do are effectively diluting other things for cash and potential validation. And I think it's very easy to get caught up in round after round momentum with a deal or a partnership. you know, some progress on the pipeline, and then you hit sort of the mid growth stage and people say, great, what company are you and how can I underwrite your valuation and it's multiple in the long term. And if you find yourself as a series B, series C company without real new business model traction, a reasonable valuation, but no drugs. I think people have found that to be sort of the valley of death in BC. So picking your primary business model, realizing that any deviation from that is a trade off and you can choose, I think your, your pill and there's a reasonable chance. I think that you would be successful if you focused on it, on it, despite, you know, the odds in biotech. My only second piece of advice would be if you choose the path of we're going to be a drug company that makes drugs. I think it's very important to be clear-eyed about where your alpha comes from. Right. Like, why do you believe, let's assume your technology works exactly as you hope it would into the future. Why do you believe that the drugs that you will make have a better chance of working? And that can be as contrarian. So only you can be the only one in the world that believes it or ask consensus as possible. But I think it's important to articulate and really examine that opinion, taking it out into the veranda and hitting it with a cricket bat, if you will. Yeah. mean, speaking of the other. the fundraising actually aspect and how intertwined it is with business development. Question for you, Akshay, is there any trends you're observing in the current funding environment that you think founders should be aware of? Yeah, I think the biggest one we are seeing is capital concentration. I think that's true for across segments, but definitely true here. One, round sizes are getting larger earlier, but at the same time, it also feels like the number of deals that are happening is relatively stable. More capital is going into fewer names, maybe either the team or the tech or the staff is driving that. But we are seeing that. Second is the piece I talked about earlier, model versus data. I think there is a concept of a model. Model itself is not defensible. And there will be commoditization. You have to decide early what you want to be and how you want to raise. Last thing we're seeing is I think there is some excitement around pharma ⁓ &E, but pharma is really in keen on procuring more assets in the department. that's driving a And those are the three things, mainly, traffic concentration, to your company, raising bigger amounts. And second, there is a lot of things. At the same time, the bar for a new AR company in mobile tech space is much higher than the first. Now, final question for both of you. Let's say like a 30-second lightning round. How do you expect the bar for AI platform validation to change over the next five years? especially in terms of clinical data, regulatory interactions, and partnerships with big pharma. So I think we've hit almost all those topics within this discussion, but Chris LaValle, what do you say? Vishwa, start with you. For us, it's always been very simple. Can you make medicines that move the needle for patients? I'm going to borrow something I heard from an incredible CEO, Will Lewis of InceMed, who said it, your medicines you make must pass the Christmas table test. So not that you can administer less frequently or slightly more conveniently, you should be able to stand up and pass along the medicine that you made with pride to a close friend or a loved one because you know it's going to change their day-to-day life. And I think it doesn't matter AI or not AI or somewhere in between, you need to show that you can move the ceiling for care for patients. Akshay, wrap it up, please. Yeah. What I would say is we're also seeing a lot of regulatory tailwinds. that AI wants to make it easier to get to commercial or single trials for phase three. So we are, I'm really excited over the next five years. think we're going to see both better medicines and treatments for cliques for diseases which have been historically untreated. So I'm super excited about how AI is going to change the discovery and development here. And even for the ride, we have a lot of new products coming out from multiple portfolio companies and ⁓ really exciting. Perfect. Well, that's a great optimistic way to end this podcast. So thank you so much, Vishwa, and thank you, Akshay, for joining this episode of Denatured. If you'd like to listen to more episodes, please turn into biospace.com. Thank you.