Chai Discovery raises $400M amid AI drug discovery boom
▼ Summary
– Chai Discovery raised $400m in Series C funding, tripling its valuation to $3.8bn in seven months.
– The company uses generative AI to design new antibodies and proteins from scratch, rather than screening existing molecules.
– Its Chai-3 model improves on earlier versions, and its technology is already deployed at Eli Lilly, Pfizer, and Novartis.
– The AI drug discovery market is projected to grow from $2.35bn in 2025 to $13.7bn by 2033, with competitors including Isomorphic Labs and Xaira Therapeutics.
– Despite rapid AI-driven molecule generation, candidate drugs still require years of lab work, clinical trials, and regulatory approval before reaching patients.
The artificial intelligence drug discovery sector has minted another fast-rising star. Chai Discovery has closed a $400 million Series C funding round that values the San Francisco-based startup at $3.8 billion, the company announced. That valuation represents nearly a tripling of the $1.3 billion price tag it carried just seven months ago.
Index Ventures led the latest financing. Kleiner Perkins, Sequoia Capital, and Dimension also participated, alongside new investors including Bain Capital Ventures, Battery Ventures, and Baillie Gifford. Returning backers OpenAI and Thrive Capital increased their stakes as well.
The round pushes Chai’s total capital raised past $600 million in roughly eleven months. The company secured a $70 million Series A in mid-2025, followed by a $130 million round at a $1.3 billion valuation last December.
Founded in 2024 by Joshua Meier, Jack Dent, Matthew McPartlon, and Jacques Boitreaud, Chai draws on experience from OpenAI, Meta’s FAIR lab, and Stripe. The company’s core mission is to design new medicines rather than discover them through traditional screening methods.
Conventional drug discovery involves sifting through vast libraries of existing molecules. Chai takes a different approach, using generative AI to create new antibodies and proteins from scratch, engineered for a specific disease target before any lab work begins.
Its 2025 model, Chai-2, became the first zero-shot system for fully de novo antibody design to achieve double-digit success rates in laboratory testing. The company describes this as a hundredfold improvement over older methodologies. The newer Chai-3 model delivers even tighter binding to more challenging targets.
Pharmaceutical companies are already adopting the technology. Chai reports that its models are deployed at Eli Lilly, Pfizer, and Novartis. This week, it announced a new collaboration with immunology firm argenx.
That commercial traction matters. Large drugmakers rarely trust unproven software tools. Having paying customers at that scale distinguishes Chai from the crowded field of AI startups chasing the same problem.
Capital continues to flood into AI-designed medicine. Isomorphic Labs, Google DeepMind’s drug discovery spinoff, raised $2.1 billion in May and also counts Novartis as a partner. Xaira Therapeutics launched with $1 billion, and Recursion has accumulated more than $1 billion in funding.
The major AI labs are moving in too. Anthropic spent roughly $400 million acquiring a small drug design startup, and ByteDance has entered the race. Academic researchers like Stanford’s James Zou are building their own AI biology firms.
Market analysts estimate the AI drug discovery market was worth about $2.35 billion in 2025, with projections reaching $13.7 billion by 2033. Chai’s value tripling in seven months suggests investors are betting faster than the forecasts predict.
The real challenge lies ahead. A model can generate candidate molecules in minutes. Those same molecules still require years of laboratory work, clinical trials, and regulatory approval before reaching patients.
Chai and its competitors are wagering that the next generation of medicines will be designed, not discovered. Whether that conviction survives the inevitable wave of clinical failures is the question that a $3.8 billion valuation is really putting to the test.
(Source: The Next Web)