Researchers and industry partners must adjust how they work with Anthropic now that the company confirmed it operates a Bay Area wet laboratory where its AI models can direct physical experiments. The change ties Anthropic’s computational research directly to hands-on biology, shifting responsibility for experimental validation and biosafety oversight onto the company and any collaborators.
Anthropic says the facility focuses on foundational biology rather than on producing finished drugs, a distinction it highlights alongside existing commercial ties to pharmaceutical companies. The company recently announced a collaboration with Novo Nordisk on joint drug discovery, and in April it acquired Coefficient Bio, a stealth AI biotech it has integrated into its life sciences efforts. Anthropic also published technical reports on accelerating protein design and improving bimolecular modeling, which it presents as tools to support discovery rather than completed therapeutics.
Eric Kauderer-Abrams, Anthropic’s head of life sciences, framed the lab as a bridge between models and measurement, saying, "We believe that to do biology, the final test is still, and will be for a while, in real lab work. We absolutely are doing that today." To control access, Anthropic this week launched a Life Sciences Verification Program intended to let vetted biological researchers use the company’s most capable models.
The announcement arrives amid prominent internal and public debates over AI risk at Anthropic. A recent resignation by researcher Jacob Coxon included his warning that "the people building AI earnestly believe that it could kill us all by the end of the decade." Separately, the company’s alignment lead has estimated a greater than 10% chance that AI could exterminate humanity within the next decade. CEO Dario Amodei has called publicly for a slowdown in development and repeatedly listed bioterrorism among the major risks posed by advanced AI.
The combination of active wet-lab work and high-profile safety warnings has already prompted outside comment, including a post on X from investor Chamath Palihapitiya: "The group behind such hits as: 'We’re All Going To Die' and 'Regulate Me Now' are building a wet lab in SF. I do not recommend this." What happens next will depend on how Anthropic balances in-house experiments, partner collaborations and the oversight mechanisms it has announced, and how regulators and the research community respond to an AI company operating at laboratory scale.
