Anthropic Expands Beyond Digital Simulations: Establishing a Physical Wet Lab in the San Francisco Bay Area
Anthropic is no longer limiting its ambitions in life sciences to digital screens and computer simulations. The American startup has established an experimental biology wet lab in the San Francisco Bay Area, where physical experiments are conducted to complement its computational work in biology and pharmaceutical research, according to Reuters. The existence of this wet lab was confirmed by Eric Kauderer-Abrams, Anthropic’s Head of Life Sciences.
From In Silico to Physical Experimentation
The transition from in silico research (computer-based studies) to physical experimentation marks a significant evolution for a company historically specialized in artificial intelligence. Anthropic clarifies that its model relies on both work conducted in its own facilities and collaborations with external partners.
However, the link between this laboratory and drug development requires nuance. Following the release of this information, an Anthropic spokesperson clarified that the lab is not specifically dedicated to drug discovery, without providing further details on its activities. Nevertheless, the company clearly lists health, life sciences, and rare disease research among its strategic priorities. This strategic shift has accelerated over the past several months.
- June: Anthropic launched Claude Science, a platform designed for researchers that aggregates scientific tools, databases, and computational capabilities. The company states its goal is to use AI to accelerate various stages of scientific research, particularly in biological sciences.
- July: Anthropic launched a program dedicated to rare genetic diseases, offering up to $50,000 in Claude credits for selected projects. The company believes AI can help identify common mechanisms across different rare diseases and generate new research hypotheses.
Strategic Value: Bridging AI and Biological Reality
The strategic interest of the laboratory lies precisely in this bridge: while an AI can analyze publications, process biological data, formulate hypotheses, or design molecules, these results must ultimately be tested against the actual behavior of biological systems. Anthropic is already working on this critical step.
- August: The company published results on the use of Claude to design proteins capable of binding to target molecules.
- September: Anthropic presented work in which Claude contributed to optimizing more than 30 models used for biomolecular modeling.
Automating Experimental Workflows
In parallel, the company is seeking to further automate experimental work. According to Reuters, Anthropic is exploring the possibility of operating robotic systems capable of executing experiments based on instructions generated by its AI models. However, this project remains in its early stages, and human intervention is still necessary.
