Anthropic Builds Physical Biology Lab Where Claude Directs Robots to Run Drug Experiments

Anthropic is building a physical biology lab in the San Francisco area where Claude will direct robots through drug development experiments, according to a Reuters report. The move marks a shift from computer-based AI drug discovery simulations toward hands-on laboratory work.

The plan calls for Claude to guide robotic systems through experiments with minimal human involvement, though Anthropic says human oversight will remain in place for safety reasons. The company has already released two tools to support this effort: Claude Science, described as an AI workspace built for researchers, and the Model Hardware Standard, a protocol for controlling lab equipment.

Targeting "undruggable" diseases

Eric Kauderer-Abrams, who leads Anthropic's life sciences division, told Reuters the biggest opportunity lies in diseases long considered "undruggable" — conditions for which no treatment currently exists. According to Kauderer-Abrams, AI could accelerate the design of complex antibodies capable of hitting multiple biological targets simultaneously, a process that is difficult and slow using conventional methods.

Built on a $400 million acquisition

The lab initiative builds on moves Anthropic made earlier this year. In April, the company acquired biotech startup Coefficient Bio for approximately $400 million and appointed Novartis CEO Vas Narasimhan to its board — a signal of Anthropic's deepening ambitions in pharmaceutical research.

Despite the scope of the buildout, Anthropic says it will not run its own clinical trials for now. According to the Reuters report, the company wants to avoid competing directly with pharmaceutical companies that would otherwise be customers or partners for AI-designed compounds.

What this means

This is a strategic bet that large language models can do more than analyze biological data — they can direct the physical experiments that generate it. If Claude can reliably run robotic lab equipment through the Model Hardware Standard, Anthropic gains a feedback loop that pure simulation cannot offer: real experimental results to validate or correct AI-generated hypotheses.

The approach also positions Anthropic as an infrastructure and tools provider to pharma rather than a competitor. By avoiding clinical trials, the company can sell AI-driven discovery capabilities to drug makers without threatening their downstream business — a distinction likely aimed at securing partnerships rather than adversarial relationships with companies like Novartis.

The real test will be whether AI-guided robotic experimentation actually shortens the notoriously long and expensive drug discovery timeline, or whether physical lab work — with its contamination risks, equipment failures, and biological variability — proves far harder to automate than digital simulation ever was. Anthropic has not disclosed a timeline for when the lab will be operational or what specific disease targets it will pursue first.