From the offices of OpenAI to the deal with Eli Lilly — how Chai Discovery became one of the flashiest names in AI drug development

Drug discovery, the art of identifying new molecules to develop medicines, is a notoriously slow and difficult process. Traditional techniques such as high-throughput screening offer expensive scatter-shot approaches and are rarely successful. But a new breed of biotech companies is leveraging AI and advanced data technologies to accelerate and streamline processes.

Chai Discovery, an AI startup founded in 2024, is one such company. In just over 12 months, the company’s young co-founders have raised hundreds of millions of dollars, amassed the support of some of Silicon Valley’s most influential investors, and built it into one of the most glamorous companies in a growing industry. In December, the company completed its Series B, raising an additional $130 million, giving it a valuation of $1.3 billion.

Chai also announced a partnership with Eli Lilly last Friday. This is an agreement in which pharmaceutical giant Eli Lilly and Co. will use the company’s software to develop new drugs. Chai’s algorithm, called Chai-2, is designed to develop antibodies, the proteins needed to fight disease. The startup says it wants to function as a kind of “computer-aided design suite” for molecules.

This is a critical moment for certain areas of chai. The startup’s deal was announced shortly before Eli Lilly announced it would also work with NVIDIA in a $1 billion partnership to establish an AI drug discovery lab in San Francisco. This “collaborative innovation lab” aims to combine big data, computational resources and scientific expertise to accelerate the speed of new drug development.

The industry is not without its detractors. Some industry veterans seem to feel that these new technologies are unlikely to have a major impact, given how difficult traditional drug development is. But for every naysayer, there seem to be just as many believers.

Elena Viboch, managing director of General Catalyst, one of Chai’s key backers, told TechCrunch that her company is confident that companies that adopt the startup’s services will see results. “We believe that the biopharmaceutical companies that are the first to partner with companies like Chai will be the first to bring molecules to the clinic and produce important medicines,” Bibock said. “In practice, this means a partnership in 2026 and a first-in-class drug entering clinical trials by the end of 2027.”

Aliza Apple, director of Lilly’s TuneLab program, which uses AI and machine learning to discover drugs, also expressed confidence in Chai’s product. “By combining Chai’s generative design model with Lilly’s deep biologics expertise and proprietary data, we are pushing the frontiers of how AI can design better molecules from the start, with the ultimate goal of accelerating the development of innovative medicines for patients,” she said.

Chai may be less than two years old, but the startup’s origins began about six years ago in a conversation with co-founder and OpenAI CEO Sam Altman. One of these founders, Josh Meier, worked on OpenAI’s research and engineering team in 2018. After leaving the company, Altman messaged Meyer’s college friend Jack Dent, asking him about potential business opportunities. Meyer and Dent originally met in a computer science class at Harvard University, when Dent was an engineer at Stripe (another company that Altman supported early on). Altman asked if he thought Meyer would be open to working with a proteomics startup, a company focused on researching proteins.

“Everyone at OpenAI sent me a message saying they thought highly of him and asked if I would be interested in working with him on a proteomics spinout,” Dent said. Dent told Altman, “Sure,” but there was just one problem. Meyer didn’t feel the technology was completely “there” yet. The AI ​​technology that powers these companies—leveraging powerful algorithms—remains a growing field and far from a necessary one.

Meyer is also pretty determined to join Facebook’s research and engineering team, which is what he will continue to do. At Facebook, Meyer helped develop the first transformer protein language model, ESM1. This is an important precursor to Chai’s current research. After working at Facebook, Meyer will spend three years at Absi, another AI biotech company based in drug discovery.

By 2024, Meyer and Dent finally felt ready to take on the proteomics company they had originally discussed with Altman. “Josh and I reached out to Sam and told him we should pick up the conversation where we left off, and that we would start having chai together,” Dent said.

OpenAI ultimately became one of Chai’s first seed investors. Meyer and Dent, along with co-founders Matthew McPartron and Jacques Boitlot, actually founded Chai while working out of the AI ​​giant’s offices in San Francisco’s Mission District. “They were kind enough to provide us with office space,” Dent revealed.

More than a year later, as Chai basks in the glow of its new partnership with Eli Lilly, Dent says the key to the company’s rapid growth has been assembling a team of extremely talented people. “We really put our heads down and pushed the frontiers of what these models can do,” says Dent. “All of the code in our codebase is homegrown. We don’t take off-the-shelf LLMs that are in the open source (ecosystem) and tweak them. These are highly custom architectures.”

General Catalyst’s Viboch told TechCrunch that he felt Chai was ready to hit the ground running. “There are no fundamental barriers to deploying these models in drug discovery,” she said. “While companies will continue to need to recruit drug candidates through trials and clinical trials, we believe that companies that adopt these technologies will have significant benefits, not only in accelerating discovery timelines but also in developing classes of drugs that have historically been difficult to develop.”