$63 billion-revenue Pfizer has signed an AI drug-development partnership with Chai Discovery, giving the startup a second major pharma deal after Eli Lilly.
Chai said its new Chai-3 model drove the agreement by doubling the success rate of its prior system and generating antibodies that bind 100 times more tightly to targets.
The upgrade could shorten a key refinement step in drug discovery: Chai said roughly half of Chai-3 molecules bind as tightly as approved drugs, improving on Chai-2's 16% design success rate.
More than 15 pharma companies are also in talks with Chai, which has raised over $225 million and is now seeking another $400 million at a $3.4 billion valuation.
AI drug discovery funding hit $11.4 billion globally in 2025, underscoring rising pharma demand even though AI-designed medicines are still likely years from market.
Chai's AI binds targets 100 times tighter, but can it predict the human side effects that derail most drugs?
Can a software-only model outpace rivals already testing AI-designed drugs in human trials?
With AI-designed drugs a decade away, what is the biggest hurdle between today's code and a future cure?
Chai Discovery’s Chai-3: The $1.3B AI Engine Powering Next-Gen Biologics and Pharma Partnerships
Overview
Chai Discovery is transforming biology into a precise engineering discipline by leveraging advanced artificial intelligence to predict and reprogram biochemical molecules. Founded by experts from leading AI institutions, the company’s vision is to develop better and faster therapeutics, especially for previously 'undruggable' conditions. Chai Discovery has developed sophisticated AI models, with Chai-3 as the latest iteration, aiming to serve as a computer-aided design suite for molecules. Building on the success of Chai-2, which demonstrated strong capabilities in designing novel therapeutic molecules like nanobodies and miniproteins, Chai-3 represents a significant step forward in AI-driven drug discovery.