Understand any stock in simple English
Enter a ticker to learn what the company does and how it makes money.
Chai Discovery builds artificial intelligence models that design new antibodies and drug molecules for pharmaceutical companies.
Chai Discovery is an AI drug-design company. It builds computer models that dream up brand-new antibodies and other drug molecules from scratch, then hands those designs to pharmaceutical companies to test in the lab and push toward real medicines. The core pitch is simple: designing a good antibody has traditionally been slow, expensive, and partly guesswork, and Chai thinks software can do a big chunk of that thinking faster and cheaper.
The company is best known for Chai-1 and its successor, Chai-2, protein-structure prediction models that guess how biological molecules fold and fit together. In late 2024, Chai made the unusual move of releasing Chai-1 openly, free for anyone to use, including for commercial work. That looked generous, but it was really a land-grab: it put Chai's name in front of every drug researcher on earth and built credibility fast. When Chai-2 arrived, the company showed it could design working antibody binders from scratch with a hit rate that would have sounded absurd a few years earlier — the model generates candidates, and a meaningful share of them actually stick to their targets in the lab. Chai does not develop its own drugs or take them through clinical trials. Its customers are the big pharmaceutical and biotech companies that pay to use its models and platform, because a faster, cheaper route to a promising molecule is worth a lot when a single failed program can burn years and hundreds of millions of dollars. The company raised serious venture money to fund this push, with investors betting that AI design becomes a standard tool in drug discovery the way modeling software became standard in engineering. The obvious competition comes from other AI-biology players and from the internal teams at large drugmakers, some of which have struck partnerships with rivals like Isomorphic Labs and Nvidia-backed efforts.
The promise is enormous and the risk is real. A model that predicts structures brilliantly still has to survive messy biology, safety testing, and clinical trials, and the big checks come only when a designed molecule becomes an approved drug. Chai's bet is that it can be the design engine underneath many drugmakers at once, collecting fees and partnerships along the way, rather than gambling everything on one molecule of its own.
Saved answer · last researched 15 Nov 2026