Anthropic

AI · Biotech · Drug Discovery

Expansion

Anthropic reportedly built a secret Bay Area wet lab for AI drug work

Turning AI predictions into real medicines still requires human hands testing molecules at a physical lab bench.

• Anthropic quietly set up a physical wet lab in the San Francisco Bay Area for biology experiments, confirmed by its head of life sciences, Eric Kauderer-Abrams, in a Reuters interview.
• Anthropic, the AI safety-focused maker of the Claude models, is pushing beyond computer-only research into targeting rare diseases it says the pharmaceutical industry has neglected.
• One source said Anthropic aims to have Claude eventually direct robots inside the lab, but the company plans to stop short of running clinical trials to avoid competing with drugmakers.
• An Anthropic spokesperson later clarified the lab is not specifically for drug discovery, without elaborating, leaving the exact scope of the facility unclear.
• Claude-designed protein binders hit a 22-35% success rate against 15 targets in recent experiments, nearly double the industry's typical 10-15% benchmark.
• The lab follows Anthropic's June launch of Claude Science, already used by Novo Nordisk and AstraZeneca, and its September 17 Life Sciences Verification Program, which loosens safety limits for biology-related model use.
• An AI lab built around safety warnings is now betting that owning physical bench science, not just software predictions, is what's needed to prove AI can deliver real cures before public trust in the technology erodes.