Today, Chai Discovery announced a $400M Series C, now valuing the business at $3.8B. The raise caps a run of momentum for the company, including the newly announced collaboration with Novartis to advance AI-driven antibody discovery, and existing partnerships with Pfizer and Eli Lilly. Together, these deals underscore a shift Chai's founders describe in the conversation below. Today, AI-driven drug discovery is moving from research promise to real deployment inside the world's largest pharma and biotech companies.
We sat down with Chai co-founders Joshua Meier and Jack Dent to talk about that shift, and what it takes to build a platform company in one of the hardest problem spaces in science. At Oak, we are focused on life sciences and the ways AI is poised to remake the industry. We've been closely tracking companies working to bring engineering-grade precision to drug discovery. Few are doing it with the momentum, or the partner list, that Chai has built.
You have been in this industry for quite a while, being at FAIR and then Absci before founding Chai. How has the industry changed over the past five years, and what has fundamentally shifted to enable a company like Chai to emerge with the momentum it has?
Josh: Five years ago, the ambition was already there, but the models weren’t good enough. Existing systems could help you understand a molecule or rank candidates, but they couldn’t create a new molecule with the properties you wanted.
What’s shifted is the ability to move from prediction towards design. Predicting the structure of an existing protein is like producing a blueprint of something that already exists. Drug discovery requires you to create something new: a molecule that binds the right target, at the right location, with the right biochemical and developability properties.
A second shift is scale. Models are more capable, compute infrastructure is much better, and we can now run these systems across many targets rather than treating every program as a bespoke research project.
You spent time at phenomenally successful companies like Stripe, where you got a front-row seat to how foundational next-generation companies are built. How are you bringing those perspectives to life sciences, and what ingredients have been missing?
Jack: Stripe was a formative experience in every way but if I had to pick just a couple of things that really influenced my thinking they’d be how to go after a problem of incredible complexity, and the importance of talent density.
Stripe took a process that used to be complicated and cumbersome and turned it into something fast and straightforward. It sounds simple but it's one of the hardest things for any company to actually achieve and maintain over time.
There are few problem spaces as complex as biology, and we want to apply the precision and scale of modern engineering to it to fundamentally accelerate how new therapeutics are developed.
On the second point, talent density has been a priority since the very start. We’re still a small team, and each person has this incredible expertise, scope and ownership, whether they’re from the biology side or incredible ML engineers. When you sit in front of a team of scientists at a pharma company and see their reaction to what we’ve built for them, it’s a great feeling.
What has been the most challenging aspect of building the company so far?
Josh: Focus. There are so many things we could be doing at any particular moment, and it takes real discipline to say no. Putting more wood behind fewer arrows can be painful, since the work we do is so exciting and we know we can have impact in so many areas, but I also think this is one of the main reasons the company has been successful to date. We don’t stop working on a problem until we feel like we’ve really solved it.
What consensus view within the industry do you think is wrong?
Jack: A lot of companies are focused on building their own drug pipelines, and have ambitions to put their own molecules into the clinic. From the start, we’ve focused on being a platform company that builds a computer design suite for molecules. Our priority is being the best partner we can be to the pharma companies we work with, since that’s the way we think we’ll have the biggest impact. There are too many promising directions to pursue on our own so we scale our impact through our partners. A single drug can cost hundreds of millions of dollars to go through trials and we’d rather put that money toward improving our models and products. Our collaborations look much more like large software partnerships than pharma deals - that wasn’t very common before.
What has surprised you the most while building Chai?
Jack: You asked earlier about the consensus opinions inside the industry that we’ve challenged. The thing that surprised me the most is the consensus view outside the pharma industry that there’s been a decade of hype and no real progress when it comes to AI’s impact in drug discovery. I think there’s a bit of a “Boy Who Cried Wolf Effect” at play. Certainly, people might have gotten a little overexcited a few years ago about the sparks of promise they saw, and appropriate skepticism is always healthy, but I don’t think most people have realised just how much has changed in the last year. From the start of 2025 to the start of 2026, it feels like we’re living in totally different universes. We like to say that last year was the year of breakthrough research, and this is the year that pharma is adopting these technologies at scale.
A decade from now, what do you think drug discovery looks like, and where is Chai?
Josh: I think the earliest stages of drug discovery will look much more like an engineering discipline.
Today, scientists often begin by generating or screening a very large number of possibilities and then iterating experimentally. A decade from now, I expect that search to happen computationally. Scientists will specify the biological target, mechanism, and desired molecular properties, and models will generate a much smaller set of candidates with a substantially higher prior probability of success.
That doesn’t remove the lab. Experimental validation will remain necessary. But the role of the lab changes, and instead of being where initial trial and error searches start, it will become the place to test and validate a smaller number of higher conviction bets.
I also expect to see models move up the biological stack. Right now we’re working on biomolecules, but over time systems will reason across proteins, cells, tissues and beyond. We want Chai to be the foundational design platform scientists use to create safer, more effective medicines.
Mikael Dolsten recently joined your board. What practical insight has a traditional pharma giant’s executive brought to an AI-first software startup, and how has that shifted your product roadmap?
Josh: Mikael is one of a relatively small number of people who actually knows what it takes to advance a molecule into clinical trials and end up with a new medicine approved at the end. As Pfizer’s CSO for 15 years, he advanced 150 molecules to clinical trials and delivered 36 approved medicines, including vaccines and cancer therapies. His belief in what we’re building has been incredibly powerful, and of course he’s been an invaluable sounding board.
Are you competing for elite machine-learning talent with companies such as OpenAI, Google and Anthropic? When recruiting researchers who may not have a background in biology, how do you pitch them on the challenge?
Josh: It’s extremely competitive out there for AI talent, of course. From day one it really helped that our co-founders Matt McPartlon and Jacques Boitreaud were both very respected for their research work, people fully understood their caliber. We tend to attract people who want to work on solving something genuinely hard, which could have a disproportionately high impact.
Jack: I’d love to take this opportunity to put on the record that you don’t have to be a structural biologist to work at Chai! It’s the biggest myth we hear during the recruitment process. I’d say well over half of our team doesn't have a biology background at all - they have backgrounds in ML, infrastructure, product engineering, even pure mathematics. The things we look for when making hires are an intense intellectual curiosity, a desire to work on hard problems, and the ability to learn fast and adapt.
People who’ve joined the team have learned the key primitives of biology and chemistry very quickly—I think that’s so much easier when you come in wanting to learn, and in an environment where you’re surrounded by domain experts. It’s actually helpful to have a balance of insiders and outsiders, since it allows us to tackle new problems from first principles while also building on established knowledge.
