Anthropic sets up biology lab as AI drug research expands

Scientist working in a laboratory during a research experiment. ILLUSTRATIVE IMAGE

SAN FRANCISCO — September 18, 2026

Anthropic has established a physical biology laboratory in the San Francisco Bay Area as the artificial intelligence company expands its work in life sciences and explores how AI can be used in real-world biological research.

The company confirmed that it has moved beyond computer-based, or “in silico,” research and is now conducting physical laboratory work. Eric Kauderer-Abrams, Anthropic’s head of life sciences, said real laboratory experiments remain an essential part of biological research, while the company also works with outside partners.

An Anthropic spokesperson later said the new laboratory is not specifically for drug discovery, but the company has been developing a broader life sciences program that includes research aimed at accelerating the development of treatments for diseases that have received less attention from pharmaceutical companies.

Anthropic is also exploring ways for its Claude AI systems to work with laboratory robots. One person familiar with the company's plans said Claude could direct robotic equipment to carry out experiments with limited human intervention. Anthropic said human involvement and oversight remain important for safety.

Kauderer-Abrams described the use of AI to automate laboratory work as being at an early stage. The company believes such systems could speed up a range of scientific processes by allowing experiments to be carried out more quickly and at greater scale.

The expansion comes as Anthropic continues to increase its investment in life sciences while also warning about potential risks from advanced AI. The company has recently discussed the possibility that AI systems could be misused for biological weapons development, while CEO Dario Amodei has called for a slower approach to the development of increasingly powerful AI systems.

Anthropic has also been building the tools and partnerships needed for its biological research. In June, Kauderer-Abrams said the company planned to conduct preclinical work in areas where established drugmakers had limited financial incentives to invest. Preclinical research takes place before testing a potential treatment in humans.

The company has launched Claude Science, an AI workbench designed to support scientific research, and has added Novartis CEO Vas Narasimhan to its board. Anthropic also confirmed its acquisition of Coefficient Bio, a biotechnology startup, although it did not disclose the purchase price. Media reports have put the value of the stock deal at about $400 million.

Anthropic is hiring for additional life sciences roles as it expands its laboratory operations. Recent positions have included work involving laboratory procurement and the analysis of proteins and nucleic acids, while the company has described its broader objective as significantly accelerating progress in life sciences.

The company is interested in conditions sometimes described as “undruggable,” meaning diseases or biological targets that have been particularly difficult to treat with existing approaches. Anthropic has pointed to areas such as the development of complex antibodies that can act on multiple targets as examples of where AI could potentially assist researchers.

However, Anthropic has not disclosed which specific diseases it is pursuing through its own research or what progress it has made toward particular treatments. It is also not running clinical trials, meaning the company has not moved into the stage where experimental medicines are tested in humans.

That boundary is important as Anthropic works with major pharmaceutical companies, including Genentech, Bristol Myers Squibb and Novo Nordisk. Kauderer-Abrams said the company does not want to compete directly with pharmaceutical and biotechnology companies whose businesses depend on bringing medicines to market.

Anthropic is therefore combining in-house laboratory work with outside partnerships. The company says its own facilities can allow it to carry out some experiments more quickly, while external partners remain useful where they can perform work more efficiently.

The approach also creates questions about trust among pharmaceutical customers. Companies using Anthropic's AI tools may be concerned about protecting commercially sensitive information from competing drug programs, although Anthropic says customer data is separated and protected.

For now, the company's biology operation remains in an early stage. Anthropic is attempting to combine AI models, laboratory automation and traditional scientific research while keeping human oversight in place and limiting its role in the later stages of drug development.