Where AI-driven drug discovery is heading: Murat Tunaboylu's three generations of tech bio

On the latest episode of Beyond the Purpose, Murat Tunaboylu, CEO and co-founder of Antiverse, set out a framework for how AI-driven drug discovery companies have evolved, and where they're heading next. It's a useful lens for anyone assessing this space, whether as an operator, an investor or a leader hiring into it.

The earliest wave of tech bio companies, Murat explains, made a straightforward pitch: with AI, we can do things the industry couldn't do before. The promise was volume, where a traditional biotech might run three programmes, these companies claimed they could run a hundred with the same resources.

Many of these companies raised significant capital and went public on that promise. But according to Murat, it largely didn't translate. Most ended up advancing only one or two programmes to the clinic, much like a traditional biotech would. He points to the underlying complexity of biology as a core reason, and says the shortfall did real damage to confidence in the model.

The companies Murat describes as the second generation took a different approach, working the problem from two angles. The first is building foundational models to understand the biology itself, how a disease state differs from a healthy one, and how a molecule needs to operate to change that state. In his view, these AI-driven tech bio companies are, in essence, data companies, and the differentiation between underlying models is narrowing quickly as new ones emerge every few months.

The second angle addresses where these efforts have historically failed: toxicity and efficacy. Murat notes that the assays used to test for these aren't high-throughput, which limits the data available for models to learn from. He expects second-generation companies to put substantial capital into closing that gap, building both foundational models and the high-throughput assays needed to validate them at scale.

Even with that progress, he's clear that the fundamentals of drug development don't disappear. A molecule can be designed in months rather than years, but it still faces the same 10 to 12-year path to market through clinical development.

The most striking part of Murat's framework is what he sees coming next: a shift towards regulatory approval of the production process itself, rather than each individual product. He points to precedents already in motion, CAR-T therapies, and antisense oligonucleotides in the UK, where the principle is to gain regulatory sign-off for the method of producing a treatment, rather than for each treatment individually.

Under this model, a company would still go through the full 10 to 12-year development and approval process once. But subsequent molecules produced through that validated process would carry the same efficacy and toxicity profile on paper, without needing to repeat the full testing cycle each time, potentially aided further by targeting smaller, more defined patient populations that contain the risk of side effects.

Murat is careful not to put a firm date on when this will play out at scale, but he sees it as the shift that would allow tech bio companies to build something of lasting value, rather than repeating the same lengthy cycle for every new molecule.

Asked where tech bio activity is concentrated globally, Murat points to Boston and Silicon Valley as still dominant, driven less by any national advantage and more by the depth of investor networks and capital based there. He notes an interesting divergence, though: while capital remains concentrated in the US, the technical expertise within these companies is increasingly distributed worldwide, with lean teams often 30 to 40 people, built from talent wherever it happens to be.

Murat's generational framing offers a useful way to think about where the value in AI-driven drug discovery is heading, and what it might mean for the leadership these companies need as they mature, from scientific and platform leadership in the early stages, through to the regulatory and commercial expertise required to bring a validated process to market at scale.

Listen to the full conversation with Murat Tunaboylu on episode two of Beyond the Purpose, available on Spotify and Apple Podcasts.

To learn more about how BioX helps organisations secure the leaders who drive performance across the full drug development life cycle, visit BioX.