Anthropic Runs Bay Area Wet Lab as Alignment Lead Cites 10% AI Risk

Anthropic confirmed to TechCrunch that it operates a wet biology lab in the Bay Area where its AI models are used to run physical experiments, according to a TechCrunch report published Sept. 18. The disclosure ties together two moves the company made earlier this year: its April acquisition of stealth AI biotech firm Coefficient Bio and the launch this week of a Life Sciences Verification Program that gives vetted researchers access to its most capable models.

What does the lab actually do?

Eric Kauderer-Abrams, Anthropic's head of life sciences, told Reuters, as quoted by TechCrunch, that the lab lets Anthropic move past theory. "We believe that to do biology, the final test is still, and will be for a while, in real lab work," he said. "We absolutely are doing that today." Kauderer-Abrams said the lab functions like a typical biotech facility, running some research internally while also collaborating with outside partners. Anthropic declined to specify current projects but said the primary focus is fundamental biology, not drug discovery.

How does this connect to Coefficient Bio?

Anthropic acquired Coefficient Bio, a stealth AI biotech startup, in April, per TechCrunch. The wet lab disclosure suggests that acquisition gave Anthropic in-house capacity to physically test hypotheses generated by its models rather than relying solely on outside labs. Separately, Anthropic published research reports on accelerating protein design and on "uplifting" biomolecular modeling, work the company frames as supporting drug discovery broadly rather than running its own pipeline.

Why does Anthropic say it isn't competing with pharma partners?

Anthropic has drawn criticism before for launching products that compete with paying customers, TechCrunch reported. The company counts pharmaceutical firms among its major clients and partners, including a recently announced joint drug-discovery deal with Novo Nordisk. By publicly framing the wet lab's work as fundamental biology rather than therapeutics development, Anthropic is drawing a line meant to reassure those commercial partners that it isn't building a competing drug pipeline.

What safety concerns has the lab raised?

The lab's existence lands amid internal warnings from Anthropic's own staff. Researcher Jacob Coxon resigned, saying "the people building AI earnestly believe that it could kill us all by the end of the decade," according to TechCrunch. Anthropic's alignment lead has put the odds of AI causing human extinction within a decade at greater than 10%, and CEO Dario Amodei has repeatedly named bioterrorism among AI's biggest risks. Amodei also posted last weekend calling on the industry to slow down and adopt self-regulation. Investor Chamath Palihapitiya captured the tension on X: "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."

Amodei told reporters last week that "AI could cure most major diseases in the next 5-10 years," a timeline that requires exactly the kind of physical verification the Bay Area lab is meant to provide, per TechCrunch.

What to watch

• Whether Anthropic discloses specific research projects or publications originating from the wet lab.
• How the Life Sciences Verification Program's researcher vetting criteria are defined and enforced.
• Whether biosecurity concerns tied to the lab factor into Amodei's proposed industry self-regulation push.
• Whether pharma partners, including Novo Nordisk, comment on the boundary Anthropic has drawn between fundamental biology and drug discovery.