Anthropic runs a wet lab in the Bay Area and uses its models to design and run physical experiments

Anthropic confirmed to TechCrunch that it operates a wet lab in the Bay Area where its models are used to plan and execute real-world biological experiments. Eric Kauderer-Abrams, the company's head of life sciences, told Reuters that final validation in biology remains and will remain bench work for the foreseeable future, and that the company is already doing it today. The lab runs much like a conventional biotech lab: some research is internal, some is conducted with external partners.

The acquisition that preceded the build-out

In April Anthropic acquired Coefficient Bio, an AI-driven biotech startup that had been operating in stealth, and that purchase cleared the path for the lab. The company says its primary focus is fundamental biology, not drug discovery — a distinction intended to distance Anthropic from the perception that it is competing with its large pharma customers. Only this week the company announced a partnership with Novo Nordisk for joint drug discovery, and Anthropic has drawn criticism in the past for launching products that were seen as competing with those of its customers.

Verification program and technical reports

To strengthen ties with the research community without stepping into drug developers' shoes, Anthropic this week launched the Life Sciences Verification Program, which gives vetted biological researchers access to its most powerful models. At the same time the company published reports on its work in the field: one on accelerating protein design and one on improving biomolecular modeling. The publications are meant to demonstrate technical capability without blurring the line between infrastructure provider and drug developer.

The tension between existential warnings and hands-on experiments

Underlying the move is a conspicuous contradiction. Anthropic researcher Jacob Coxon resigned and warned that the people building AI genuinely believe it could drive humanity to extinction by the end of the decade. The company's head of alignment put the probability of that outcome at more than 10% over the next ten years. Chief executive Dario Amodei published a call over the weekend for the industry to slow down and adopt self-regulation, naming bioterrorism as one of AI's greatest risks. Investor and entrepreneur Chamath Palihapitiya summed up the irony in a post: the group that signed "we all die" and "regulate me now" is building a wet lab in San Francisco, and he does not recommend it.