Chai Discovery

Chai Discovery builds frontier artificial intelligence models that predict and reprogram the interactions between biochemical molecules, with the aim of redesigning drug discovery. The company applies machine learning to protein structure prediction, molecular interaction modelling and antibody design, targeting the atomic-precision engineering of drug-like antibodies against challenging targets. Its stated ambition is to turn biology from a science into an engineering discipline and to shorten the path from concept to therapeutic.

Its product line includes Chai-1, a model for predicting and reprogramming interactions between biochemical molecules, encompassing protein structure prediction and molecular design. Chai-2 is reported to achieve success rates of more than 85% in computationally designing molecules that meet drug-like properties, among them functional GPCR agonists generated with minimal screening. In principle, such models could replace the traditional screening of millions to billions of protein sequences, a process that typically takes years and costs billions.

The company was founded by researchers who co-invented protein language modelling and built state-of-the-art folding algorithms. It counts OpenAI, Thrive Capital and General Catalyst among its investors and has a partnership with Eli Lilly. Relevant technical domains include protein language modelling, protein structure prediction, molecular interaction modelling, antibody design and biomolecular engineering, serving the pharmaceutical, biotechnology and therapeutics industries.