Anthropic is operating a wet biology laboratory (wet lab) in the Bay Area, where the company can use its AI models to run physical experiments. Head of Life Sciences Eric Kauderer-Abrams confirmed to Reuters: “We’re doing that today.”
This does not come as a surprise. Anthropic acquired Coefficient Bio, a stealth AI biotech startup, in April 2026 for $400 million. The new laboratory is a natural next step following that deal.
Fundamental Research, Not Drug Development
Anthropic emphasizes that the lab focuses on fundamental biology, not pharmaceutical development. The stated reason is that Anthropic has numerous major pharmaceutical clients and partners, including Novo Nordisk—which recently signed a joint drug discovery partnership agreement. If Anthropic were to develop drugs itself, it would compete directly with its own customers.
The company has previously faced backlash when launching products perceived as competing with its customers. The biology lab is not intended to repeat that mistake.
In parallel, Anthropic launched the Life Sciences Verification Program, allowing vetted biology researchers access to its most capable models. The company also published a report on accelerating protein design and upgrading molecular biology modeling.
A Controversial Contrast
Just a week earlier, Anthropic researcher Jacob Coxon resigned with a warning that “the people building AI genuinely believe it could kill us all this decade.” Anthropic’s head of alignment put the probability of AI wiping out humanity within the next 10 years at over 10%.
CEO Dario Amodei has posted calling on the industry to slow down and self-regulate. He has repeatedly cited bioterrorism as one of AI’s greatest risks.
Yet Anthropic is still building a biology lab. Investor Chamath Palihapitiya quipped on X: “The group behind hits like ‘We’re all gonna die’ and ‘Please regulate me’ is building a wet lab in San Francisco. I do not endorse this.”
This contradiction reflects a larger question in the AI industry: Can companies simultaneously warn of existential risks while expanding capabilities into the very areas they deem most dangerous? Anthropic is likely betting that a deeper understanding of biology will make AI safer, not more dangerous. But the line between “research to protect” and “research to exploit” has never been clear.