On June 30, 2026, Anthropic launched Claude Science at a San Francisco event — a dedicated AI research workbench that integrates more than 60 scientific databases and tools into a single environment. The announcement also came with something less expected: Anthropic disclosed it is running its own internal drug discovery program, targeting neglected diseases that traditional pharmaceutical companies don't find commercially attractive.
What Claude Science Actually Does
Drug discovery research is famously fragmented. A medicinal chemist might move between PubMed for literature, a protein structure viewer for 3D biology, a cheminformatics tool for molecular properties, R or Python for statistical analysis, and a figure-generation tool for publications — all in a single afternoon. Claude Science collapses this into one environment built on top of Anthropic's existing Claude models.
The platform integrates genomics, proteomics, and cheminformatics databases alongside commonly used research tools like PubMed, Jupyter, and R. Researchers can search literature, run statistical analyses, generate molecular visualizations, and produce publication-ready figures without switching applications. The 60+ built-in capabilities cover everything from BLAST sequence alignment to ADMET property prediction for drug candidates.
Anthropic's Own Drug Pipeline
More surprising than the product itself is what Anthropic is doing with it internally. The company's life sciences head Eric Kauderer-Abrams confirmed that Anthropic has started its own drug discovery program, focused on treatments for rare and neglected diseases — conditions like Chagas disease, leishmaniasis, and schistosomiasis, which collectively affect hundreds of millions of people but attract minimal pharmaceutical investment because patients in endemic regions can't pay drug prices that make clinical development financially viable.
This positions Anthropic alongside Google DeepMind (which recently partnered with A24 and has its own protein folding-derived drug programs) and OpenAI (which has backed biotech firms using GPT-5.5 for target identification) in a three-way AI-pharma race. The difference is that Anthropic is running the discovery program directly rather than licensing technology to a pharma partner.
Open to External Researchers
Claude Science is currently in beta and available to biopharma researchers. Anthropic is also supporting up to 50 external research projects through a grant program offering up to $30,000 in Claude Science credits per project, with cloud compute credits from Modal. Applications are open until July 15, 2026, with recipients announced by July 31.
The eligibility scope is deliberately broad: academic labs, nonprofit research organizations, and independent researchers can all apply, not just established pharma companies. The stated goal is to direct AI research capacity toward problems that wouldn't otherwise attract it.
Why This Matters Beyond Pharma
The pharmaceutical application is the headline, but the underlying product is something more general: a domain-specific AI workbench that integrates scattered professional databases and tools into one coherent interface. If Claude Science works well for drug discovery, the template — unified access to specialized data sources, AI-assisted analysis, publication-ready outputs — is directly applicable to materials science, climate research, genomics, and academic research broadly.
Anthropic's move into science also signals something about where the company sees its differentiation. Competing on general chatbot capability against OpenAI and Google is expensive and incremental. Owning a vertical where deep domain integration creates genuine workflow value — and where the stakes of getting answers right are measurably high — is a different game entirely.