Anthropic launches Claude Science, an AI workbench for researchers

On 30 June 2026 Anthropic launched Claude Science, an AI workbench aimed at life-sciences researchers who otherwise split their work across PubMed, Jupyter, R and cluster terminals. It folds more than 60 scientific databases, computing tools and research workflows into one environment. A grants programme accompanies the launch: up to 50 projects, each eligible for as much as $30,000 in Claude credits, plus up to $2,000 of Modal compute for select projects; applications run through 15 July 2026, awards by 31 July, and projects from 1 September to 1 December 2026.

The launch is the visible edge of a larger bet — Anthropic's ~$400M all-stock acquisition of Coefficient Bio and its hire of AlphaFold lead and 2024 Nobel laureate John Jumper, in service of Dario Amodei's goal to compress life-sciences R&D by 10x. But the moat here is workflow and data integration, not model IQ, and it isn't secured yet. It's a three-way race with OpenAI (which shipped GPT-Rosalind in April 2026) and Google DeepMind, and customers overlap: Novo Nordisk partners with both Anthropic and OpenAI, and the Allen Institute has worked with both — pharma is running parallel evaluations, not committing.

Watch whether the grants programme converts into sustained lab use once awards land at the end of July, whether any pharma partner commits exclusively rather than evaluating rivals side by side, and how OpenAI and Google DeepMind answer. AI-for-science is now an explicit frontier-lab battleground; the launch is a strong opening move, but adoption over the coming months — not the announcement — will decide who leads.