18 July 2026 · 1 min read
$3.8bn valuation, well done, Chai Discovery
I look at why Chai Discovery just raised $400M at a $3.8B valuation by deliberately refusing to build its own drug pipeline, and what that strategic choice means for who owns molecule selection in pharma R&D.
Author
Share
$3.8bn valuation, well done, Chai Discovery.
My former colleagues at Novartis expanded its year-long partnership with Chai Discovery. The day after, Chai raised $400 million at a $3.8 billion valuation. Lilly and Pfizer are already customers.
The company is barely two years old and has no drug of its own. That is the point. Chai sells the design machinery and has been explicit that it will not build a pipeline. Pharma can run its general models, train private versions on proprietary data, and wire them into internal workflows without funding a future competitor.
That is the opposite of the move Anthropic made when it launched Claude Science and said it would develop drugs itself. A company thinks twice before handing its data and workflows to a partner that also wants the resulting medicine.
The founders and team explain the valuation as much as the customers do. CEO Joshua Meier helped build ESM-1, one of the first protein language models, at Meta, then led AI research at Absci. OpenAI backed the seed round. Former Pfizer R&D chief Mikael Dolsten sits on the board.
The $190 billion question (2024 combined R&D spending of top 20 pharma) is who owns molecule selection. Chai just priced its answer at $3.8 billion.
Related insights
31 Mar 2026
Eli Lilly and Insilico Medicine: what three AI drug discovery deals signal
17 Feb 2026
Big Pharma AI deals: why binding intelligence matters more than “AI”
20 Nov 2025
Jeff Bezos just bet $106M on AI that designs proteins that don't exist in nature
9 Jul 2026
AI Memory Will Become the Real Enterprise Platform Lock-In