Skip to content
AI.info

companies

Chai Discovery

AI lab building foundation models that predict and design biomolecular structures, from the open Chai-1 model to the antibody-design system Chai-2.

Chai Discovery

Chai Discovery was founded in early 2024 by Josh Meier, Jack Dent, Matthew McPartlon and Jacques Boitreaud, a team with backgrounds spanning OpenAI, Meta's FAIR lab, Google X and Stripe, on the bet that biomolecular structure prediction and design could move as fast as language modeling had. The San Francisco company released its first model, Chai-1, within months of founding in late 2024, open-sourcing both code and weights under an Apache 2.0 license and offering a free hosted web tool that quickly became a talking point among computational biologists as a viable alternative to Google DeepMind's AlphaFold 3. In mid-2025 the company followed with Chai-2, a closed model for de novo antibody design that reported a roughly 16-20% experimental hit rate against previously unaddressed antigens — more than a hundred times higher than the sub-0.1% hit rates typical of prior computational methods, according to the company's published validation across dozens of novel targets. Chai Discovery has since signed drug-discovery collaborations with pharma partners including Eli Lilly, and in July 2026 raised a $400 million Series C led by Index Ventures with Kleiner Perkins, Sequoia Capital and Dimension participating, valuing the two-year-old company at roughly $3.8 billion.

Founded
2024
Headquarters
San Francisco, United States
Sector
lab

Tools

  • Chai-2

    AI protein-design model for generating de novo antibodies, nanobodies, miniprotein binders, and epitope-specific designs.

  • Chai-1

    Chai-1 predicts molecular structures and interactions across proteins, small molecules, DNA, RNA, and covalent modifications.

Official website