Anthropic to Develop Its Own Drugs for Neglected Diseases
SkimNews Take
<think>The article says Anthropic is developing its own drugs while also selling software to drugmakers — a vertical integration play that creates a competitor-customer tension. The second-order insight here is about the structural conflict of interest this creates: by becoming a drug developer itself, Anthropic potentially compromises its neutrality as a software vendor, which could push rival pharma companies toward competitors or in-house solutions. This is grounded in the article's explicit mention of this "unusual position.
Get the Health newsletter
Daily health & science — research, biotech, public health, the studies worth knowing. Free.
- Anthropic unveiled Claude Science, an "AI workbench for scientists" at "The Briefing: AI for Science," pulling fragmented tools and datasets into one environment with figure and visual generation
- Anthropic said it will develop drugs of its own, focused on "neglected" diseases, per head of life sciences Eric Kauderer-Abrams — one of the most direct public moves by a frontier AI lab to actually create pharmaceuticals
- Anthropic now sells AI tools to drugmakers that could become its competitors, joining AI-first drug companies like Insilico, Google DeepMind spinout Isomorphic Labs, and Big Pharma firms building their own AI capabilities
- AI drug discovery already pervades "every single stage" of development according to University of Cambridge professor Namshik Han, but no AI-designed drug has yet cleared clinical trials and FDA approval
- Oxford professor Frank von Delft cautioned that AI models "haven't yet come close to making experiments unnecessary" — drug candidates still require costly real-world testing for efficacy, toxicity, and safety
- Anthropic has spent the last year hiring biologists, building wet labs, and recruiting from Big Pharma and prestigious academic institutions, with several live applications for life sciences roles
Why it matters: Anthropic offered almost no specifics on which diseases, whether it will partner for clinical trials, or what happens if it finds a promising candidate — and no AI-designed drug has ever reached FDA approval, putting any payoff at minimum "the better part of a decade" away per UCL professor Matthew Todd. The company is simultaneously selling Claude to pharma customers and competing with them, a conflict no other frontier AI lab has pursued as publicly.
Ask SkimNews
