Anthropic has opened a physical biology lab in the San Francisco Bay Area, marking a move beyond computer-based modeling as the AI company deepens its push into drug research. Eric Kauderer-Abrams, the company's head of life sciences, confirmed the facility to Reuters, making its existence public for the first time.

In an interview with the news agency, Kauderer-Abrams said the lab is already up and running and that Anthropic still sees hands-on work as essential to testing biology. Computational models can generate hypotheses, but testing them against real biology physical experiments. Asked whether the company was conducting real-world experiments, he confirmed it was “absolutely” doing so today.

A spokesperson told CNBC that the facility was not set up specifically for drug discovery but declined to provide further details, according to Quartz.

Focus on diseases drugmakers overlook

The lab will pursue an area of medicine and drug research that has traditionally been less attractive to pharmaceutical companies. Kauderer-Abrams first discussed the plan at a June gathering in San Francisco, where he said Anthropic would concentrate on early-stage research in areas that established drugmakers had largely overlooked because they offered little financial incentive.

The company views AI as a way to speed up treatments for conditions long considered too difficult to tackle, including complex therapies. The work could include bispecific and trispecific antibodies, which are designed to bind to more than one site on a protein or cell. Anthropic has not identified a specific disease it is targeting, nor has it disclosed any drug candidate from the research — leaving the lab's eventual disease focus and first findings as the key signals to watch.

Anthropic is also drawing a clear line between its research and the business of bringing drugs to market. “We're not competing with pharma and biotech companies that make their business in bringing drugs to market,” Kauderer-Abrams told Reuters. The company is not running clinical trials yet, and any treatment resulting from its research would still need to undergo preclinical testing, human trials and regulatory review.

Robots, Claude and a growing division

Another part of the experiment is testing whether Claude can run the lab entirely on its own. Anthropic is exploring whether its models can direct robotic systems to carry out experiments with minimal staffing, according to Quartz.

Kauderer-Abrams described the work as still in its early stages, saying the company is “in the very early innings of using AI to automate the execution of lab work.” A spokesperson stressed that human oversight remains a safety requirement.

The opening builds on a series of recent life sciences moves by Anthropic. The company launched Claude Science, a research workbench that connects to more than 60 scientific databases covering genomics, proteomics and structural biology. It also added Novartis chief executive Vas Narasimhan to its board and acquired startup Coefficient Bio for about $400 million in stock.

In August, Anthropic introduced a Model Hardware Standard designed to help AI systems communicate with lab equipment. It has also been building partnerships across the pharmaceutical industry, including a deal with Novo Nordisk this week, access to Claude for Bristol-Myers Squibb staff and AI services supplied to Roche's Genentech. Kauderer-Abrams said life sciences has become one of the company's most heavily staffed and resourced units.

The lab's opening follows recent warnings from Anthropic about the risks posed by increasingly capable AI. CEO Dario Amodei has recently called for a slowdown in advanced AI development to reduce the risk of losing control over frontier systems. In recent weeks, the company has said advanced AI could carry serious risks, including the possibility that its systems could be used to help develop biological weapons. Anthropic is thus expanding hands-on biology work even as it cautions about the risks advanced AI could pose in the same domain.