Anthropic is building an internal drug discovery program, putting its own researchers to work developing treatments rather than only selling AI tools to the companies that do. Eric Kauderer-Abrams, the company’s head of life sciences, announced the effort at a San Francisco event on Tuesday, framing it as a way to test Anthropic’s own tools under real conditions. The company is aiming squarely at “neglected” diseases, the ones traditional biopharma passes over because the commercial math does not work.
That framing matters more than it first appears. A public benefit corporation, Anthropic is not bound to the same shareholder-return logic as a conventional biopharma company, and its spokesperson leaned on that structure directly, telling CNBC the company “can choose programs on patient benefit, including work the commercial market overlooks.” That is a real structural advantage. It is also a claim that has not yet produced a drug.
“We’re doing this because we believe first and foremost that to build the right models, products and tools to accelerate the industry, we need to live it along with all of you,” Kauderer-Abrams said, according to CNBC. “We believe in the power of tight feedback loops, and there’s no substitute for having our own experiences alongside you all in the trenches trying to develop drugs.”
The unresolved question is what happens after a candidate molecule looks promising. Traditional drugmakers push candidates into clinical trials, a process that costs hundreds of millions of dollars and takes years even for well-funded sponsors. Kauderer-Abrams did not say whether Anthropic intends to run trials itself, license candidates to a pharma partner, or spin out a separate entity. An Anthropic spokesperson added only that “we’re at the start of this, and we’ll share more as the work progresses.”
Anthropic’s leadership also tied the program directly to commercial incentives. Jonah Cool, the company’s head of life sciences partnerships, said the neglected-disease focus doubles as a proof point for Claude Science, the product Anthropic is selling to drugmakers. Running its own discovery work alongside the companies it hopes to sign as customers gives Anthropic a credibility argument that a slide deck cannot: evidence that its models work on a real pipeline, not just a benchmark.
Anthropic is not the first technology company to bet on healthcare, and the sector’s track record with outside entrants is uneven. Alphabet and Apple have both pushed into health in different ways, with results that have not reshaped either company’s core business. Amazon took the acquisition route instead, buying One Medical and PillPack and folding both into Amazon Health Services, a unit still working to prove it can outperform standalone healthcare operators. Drug discovery is a different and harder target than either primary care or pharmacy logistics, since it demands sustained capital and regulatory expertise that has nothing to do with training language models.
The announcement gives no benchmark data, no named disease targets, and no timeline for when a candidate might emerge. Judged purely on what was said Tuesday, this is a positioning move: it tells drugmakers that Anthropic is willing to put its own AI to work on the hardest, least profitable version of their problem before asking them to buy the tool. Operators evaluating Claude Science should treat the program as a credibility signal for now, and watch for the first disclosed candidate as the real test of whether Anthropic’s models translate into anything a regulator would recognize as a drug.
Reported by CNBC on June 30, 2026.
