Anthropic launches Claude Science, a research workbench integrating 60+ scientific databases for pharma and biotech

Anthropic launched Claude Science on June 30, 2026, a dedicated AI workbench for scientific research that integrates more than 60 scientific databases into a single environment, enabling researchers to pull data from disparate sources and work on it in one place rather than switching between dozens of disconnected tools. The product was announced at Anthropic's "The Briefing: AI for Science" virtual event, where the company stated: "AI could unlock the greatest era of scientific discovery in human history. Getting there depends on what we choose to build, and who we build it with."

What's new

Claude Science is a standalone workbench product — not a new AI model — that packages Claude's reasoning capabilities alongside an integrated layer of scientific data infrastructure:

• 60+ scientific database integrations: Researchers can query and reason across databases covering genomics, chemistry, biology, and other life sciences domains without manually exporting and importing data between tools
• Auditable outputs: All results are auditable, allowing researchers to validate, rerun, and trace back Claude's conclusions to the underlying data sources
• Biology and chemistry automation: The workbench can automate tasks such as protein structure prediction, gene expression analysis, and compound screening workflows
• Compressed research timelines: Anthropic frames the tool as addressing a core problem in scientific research — that researchers currently work across "dozens of disconnected tools, databases, and compute environments" — and positions Claude Science as compressing research timelines from weeks into hours

Target users include senior executives and scientists at pharmaceutical and biotech companies (CIOs, Chief Scientific Officers, VPs of R&D), principal investigators at academic research institutions, and biotech founders. The event featured demonstrations and customer spotlights from leading pharma, biotech, and research institutions.

Context

Anthropic already runs an AI for Science program offering free API credits (up to $20,000 over six months) to academic and nonprofit researchers in biology, chemistry, medicine, and related fields. Claude Science appears to be a commercial product tier aimed at industry — pharmaceutical companies, biotech firms, and commercial research organizations — that need an enterprise-grade, pre-integrated scientific data environment rather than raw API access.

The launch also positions Anthropic more directly in competition with OpenAI's science-focused offerings. OpenAI has partnered with life sciences organizations and offered models in biology research contexts, while Google DeepMind's AlphaFold and related infrastructure have dominated the structural biology and protein prediction space. Claude Science enters that market with a workbench product approach rather than a standalone model, betting that scientific data integration is the friction point for enterprise adoption.

The timing connects to Anthropic's broader revenue expansion ahead of an anticipated IPO. Science and healthcare represent large, high-value enterprise markets where AI can justify premium pricing based on research outcome value rather than pure token consumption.

Why it matters

Scientific research has historically been siloed: genomics tools don't talk to chemistry databases, which don't connect to clinical trial records. Claude Science's workbench approach — where 60+ databases are already integrated and Claude can reason across them simultaneously — addresses a real integration bottleneck that has slowed AI adoption in pharmaceutical and academic research settings.

The auditability emphasis is also strategically important. Scientific credibility requires that AI-generated outputs can be traced to their sources and independently verified. By building audit trails into the product, Anthropic positions Claude Science for regulated industry use cases (drug discovery, clinical research, regulatory submissions) where traceability is a compliance requirement, not just a nice-to-have.