What Chai Discovery does
Chai Discovery builds multimodal AI foundation models and a “computer-aided design suite” aimed at accelerating molecular and biologics discovery—especially de novo antibody design—by learning biological structure and interaction principles and using them to propose new molecules from scratch. The company presents its models as engineering tools that industrial life-science teams can apply during early-stage discovery workflows, including zero-shot antibody design and design-criteria control (e.g., framework/species/ortholog specificity and cross-reactivity).
More
Chai’s product approach is organized around model generations (Chai-1, Chai-2, and a more advanced model series discussed as Chai-3) that are made available through licensing/access for commercial and limited non-commercial use. The company describes its current platform as a “computer-aided design suite for molecules,” and it emphasizes adoption by large pharma organizations (naming Eli Lilly, Pfizer, and Novartis) as part of its strategy to deploy the models at scale across partner discovery teams. Business model (as described publicly): Chai provides controlled access to its models for commercial organizations (via an access form/call) and limits academic use; it also publishes use restrictions/terms around prohibited uses and training/derivative restrictions for its released AI models and outputs. Current strategic position: Chai reports having raised multiple rounds in quick succession and publicly announced a $400M Series C in July 2026 at a $3.8B valuation, positioning the additional capital for increased compute/data, faster model development, and broader deployment of its technology across large pharmaceutical partners.