Anthropic confirms wet lab for AI-supported biology experiments
Anthropic has confirmed that it operates a wet biology laboratory in which its AI models can be used to run physical experiments. Eric Kauderer-Abrams, Anthropic’s head of life sciences, told Reuters that the company is already conducting this work. He said the facility operates like a typical biotech lab, combining Anthropic’s own research with work involving external partners.
The company did not disclose the lab’s specific research programmes. It said the facility is not focused on drug discovery, an area where Anthropic has major customers and partners. Anthropic recently announced joint drug-discovery work with Novo Nordisk, while maintaining that the wet lab is not intended to make it a direct competitor to pharmaceutical companies.
Physical testing adds a laboratory layer to AI work
The confirmation establishes that Anthropic’s life-sciences activity extends beyond providing models for analysis, hypothesis generation or other software-based tasks. Kauderer-Abrams said that, for biology, the final test remains real work in a laboratory. The lab gives the company a setting in which AI-generated ideas can be evaluated through physical experiments.
Anthropic acquired Coefficient Bio, a stealth AI biotech company, in April. That acquisition provides context for the laboratory capability, although Anthropic has not detailed what research is being carried out there. The company’s commercial position also makes the boundary significant: it sells and partners in sectors where customers may be sensitive to a model provider developing adjacent internal capabilities.
Research expansion alongside access controls and risk warnings
Anthropic has also launched a Life Sciences Verification Program intended to give vetted biology researchers access to its most powerful models. The programme arrives as the company continues to frame bioterrorism as one of the major risks associated with advanced AI. CEO Dario Amodei has called for the industry to slow down and adopt self-regulation, while also arguing that AI could help cure most major diseases within five to ten years.
The company’s growth and its expanding product reach are part of the context in which Anthropic’s growth beyond model performance outlines why Anthropic’s momentum is not solely a question of model performance. The wet lab adds a physical research dimension to that broader life-sciences strategy, while the verification programme places an access-control mechanism around powerful models for qualified researchers.
What customers should clarify
For businesses using AI in regulated or research-intensive fields, Anthropic’s disclosure underlines the need to distinguish model access, joint projects and a provider’s in-house research. Procurement and governance teams can ask where experimental work is performed, what is shared with external partners, and how a supplier defines work outside its stated commercial focus. Those questions help organisations assess competitive boundaries and controls before deploying advanced AI tools in sensitive workflows.