Anthropic's IPO filing taps brakes on biology claims

Anthropic has spent years promoting its models as engines of biological discovery. Two years ago, CEO Dario Amodei wrote that he was "open" to the idea that AI systems could produce a thousand years of progress in biology in just five to 10 years. In August, he said on X that the company was ramping up its efforts in biology and medicine, predicting AI would make it possible to cure most human diseases in the next five to 10 years. The company also promoted Claude's discovery of a "Crispr-like" enzyme in September.

The IPO filing Monday takes a more measured tone. While it still touts AI's ability to "profoundly accelerate the rate of progress in biology and medicine," it acknowledges that the company will need to engage in activities outside its historical experience as a software and research organization, including handling hazardous materials and operating lab equipment. It also reframes the enzyme discovery as something that "might have the potential" to become a biotechnology tool, a notable retreat from earlier enthusiasm. Researchers had previously told WIRED they were skeptical of the enzyme's value, and one scientist suspected Claude had used his research (the company denied access).

The filing also signals a strategic shift: Anthropic no longer sees itself solely as an outside AI provider to biotech, but as a company for which biotech discovery is a core part of its business model. That brings it face-to-face with the same long timelines as any biotech firm—test tube, animal, then human trials—and the filing warns that any drug candidates may never generate revenue. SEC filings are typically more cautious than executive statements due to legal risks, but the document makes clear that proving worth in trials cannot be hastened simply by adding computers.