Latest / The Series A Show with Fexingo: Early-Stage Funding, Pitch Decks, and Startup Milestones / How Biotech Startups Are Cashing In on AI Drug Discovery
Transcript
- Lucas: So the ARKG Genomics ETF is up over ten percent in the last five days, which is a pretty dramatic move even for a sector that's used to volatility. Luna: Yeah, I saw that. ARKK also up three and a half percent. Biotech and tech are having a moment, but what's driving the genomics angle specifically? Lucas: A lot of it is the AI drug discovery narrative finally translating into real pipeline progress. We're seeing startups that five years ago were just academic white papers now raising serious Series A rounds based on preclinical data generated entirely by machine learning models. Luna: Right, so the classic 'software ate the world' but now it's eating wet lab biology. Lucas: Exactly. And the funding dynamic is interesting because biotech has always been capital intensive — you need millions just to get to phase one trials. But AI is compressing that timeline. A startup like Recursion Pharmaceuticals, which went public a while back, started by using AI to screen billions of compounds. Now newer companies are doing that at the seed stage. Luna: So what's the typical check size for an AI biotech Series A these days? Is it higher or lower than traditional biotech? Lucas: It's actually comparable on the surface — we're seeing ten to fifteen million dollars as a typical Series A. But the difference is that ai native startups can get to that point with less capital because they're not renting lab space and hiring armies of PhDs from day one. They're running simulations. Luna: But investors still want to see wet lab validation, right? You can't just say 'our model says this molecule works' and expect a check. Lucas: Absolutely. The key metric that Series A investors are looking for now is what they call 'preclinical validation in a relevant model.' That means you've shown your ai discovered drug works in cells or in animals, not just in silico. I spoke to a partner at a top biotech VC last month, and he said the bar is actually higher now because there's so much noise. Luna: So the 'fake it till you make it' approach doesn't fly in biology. Lucas: It really doesn't. But the companies that do have that validation are getting premium valuations. Take Insitro, which is a bit more mature — they raised a Series C at a billion-dollar valuation a couple years ago. Their whole thesis is using machine learning to identify drug targets and then design molecules. They have partnerships with big pharma like Bristol Myers Squibb. Luna: I remember that deal. It was structured as a traditional pharma partnership — milestones and royalties — not just equity. Lucas: Right, and that's actually a great signal for early-stage investors. If a startup can land a partnership with a major pharma company before Series A, that de-risks the whole thesis. I've seen a seed-stage company called Chimera Biosciences, not a real name but let's call it that, that got a two million dollar upfront payment from Novartis just for access to their AI platform. Luna: That's smart. So the startup doesn't have to carry the full cost of development. Lucas: Exactly. And that's becoming a pattern. For founders looking to raise a Series A in this space, my advice is: get a pharma partnership first, even if it's small. It validates your tech in the eyes of VCs who may not have the scientific background to judge your model themselves. Luna: Now, on the flip side, we've seen some high-profile AI drug discovery startups fizzle out. What's the common pitfall? Lucas: The biggest one is overpromising on timelines. Biology is messy. You can predict a molecule's binding affinity, but you can't predict how it will behave in a human body. A lot of startups have gone under because they told investors they'd be in the clinic in two years, and then it took five. Luna: So the valuation gets ahead of the science. Lucas: Exactly. And that's why the current market is actually healthy. The ARKG jump this week might feel like hype, but I think it reflects a maturing understanding. Investors are rewarding companies that have real data, not just cool demos. Luna: What about the funding environment specifically for AI biotech startups in May 2026? Are we seeing more crossover investors — the Tiger Globals and the like — coming in? Lucas: Yes, and that's new. Historically, biotech was the domain of specialist VCs. But now generalist tech funds are writing checks because they understand the AI side. I saw a deal last month where a Series A was led by a tech-focused growth fund at a ninety million valuation for a company that hadn't even picked a lead candidate yet. Luna: That sounds frothy. Lucas: It is, but the company had a really strong team — former executives from Genentech and DeepMind — and they had a clear path to a partnership. So it's not just blind speculation. But it does mean the market is bifurcated: the top quartile of startups are getting premium terms, while everyone else struggles. Luna: So what should a founder listening today take away? If you're building an AI biotech startup and aiming for a Series A in the next twelve months, what's the one thing you should focus on? Lucas: I'd say: get your preclinical data package tight. That means choosing the right disease model, running the right controls, and ideally getting a contract research organization to validate your results. Then, start talking to pharma business development teams early — before you even need the money. Because a partnership letter of intent can transform your Series A from a grind into a competitive process. Luna: And maybe don't name your company after a mythical creature? I feel like every other biotech startup is called something like 'Phoenix Therapeutics.' Lucas: Ha, that's a fair point. But seriously, the naming is less important than the science. And as we've seen this week, the market is rewarding good science with strong capital inflows. The question is whether that momentum can sustain into the second half of the year. Luna: Do you think it will? Lucas: I think the secular trend is real. Drug development costs have been ballooning for decades, and AI is one of the few genuine solutions. But there will be bumps. The companies that survive will be the ones that marry technical rigor with realistic timelines. It's not easy, but the payoff — both financial and societal — is enormous. Luna: Alright, so for our listeners who are founders or investors in this space, what's a concrete metric they should track that most people overlook? Lucas: I'd say the 'hit rate' from your AI predictions to actual experimental validation. If your model says a compound will work, and it works in the lab sixty percent of the time, that's world-class. Most startups are below thirty percent. If you can show a hit rate above forty percent, investors will pay attention. That's the kind of number that turns a conversation into a term sheet. Luna: Love that. So track your hit rate, get a pharma partner, and don't overpromise on timelines. That's the recipe. Lucas: That's the recipe. And maybe avoid naming your company Phoenix Therapeutics. Luna: I'll make a note. Thanks Lucas. Lucas: Thanks Luna. We'll be back next week with another angle on the startup funding landscape.