Anthropic's Boldest Move Yet

On June 30, 2026, at the AI for Science event in San Francisco, Anthropic made a declaration that sent ripples across both the tech and pharmaceutical industries: it is launching its own internal drug discovery program for neglected diseases, and releasing Claude Science — a purpose-built research workbench designed to transform how scientists discover new medicines.

This is not a sideline experiment. Anthropic has been methodically building toward this moment. In April 2026, it acquired Coefficient Bio, a stealth drug discovery startup. Shortly after, John Jumper — co-creator of Google DeepMind's Nobel Prize-winning AlphaFold — joined Anthropic to lead its life sciences division. The pieces were quietly falling into place.

What Is Claude Science?

Claude Science is not a chatbot with a science theme. It is a purpose-built research environment where 60+ preconfigured tools — scientific database connectors, genomics pipelines, sequence analysis utilities, and proteomics tools — operate as first-class citizens alongside the AI model.

Key capabilities:

- 60+ integrated tools spanning genomics, proteomics, and drug discovery - Multi-source computing — local, remote, and high-performance computing from a single interface - Research environment — tools, data, and AI working together rather than through chat - Available now in beta for Pro, Max, Team, and Enterprise users on macOS and Linux

"To build the right models and tools to accelerate the industry, we need to live it alongside all of you," said Eric Kauderer-Abrams, Anthropic's head of life sciences.

An Internal Drug Program — for Credibility

Rather than just selling tools to pharmaceutical giants, Anthropic is using Claude Science itself to develop drug candidates for neglected diseases — rare genetic disorders and tropical diseases that traditional biopharma companies won't prioritize.

This is a credibility play, and a smart one. By generating real scientific results internally, Anthropic can demonstrate to potential pharma customers that its platform works on real problems — not just benchmarks. As one industry analyst put it: "It's easy to sell shovels during a gold rush. It's harder to prove your shovel works by digging gold yourself. That's what Anthropic is doing."

The Impact: Faster Drugs, Lower Costs

Novartis CEO Vas Narasimhan has stated that new AI models could cut drug development time from twelve years to seven or eight years. Better safety predictions could double the success rate from 8% to 16%.

In June, Anthropic published VirBench, showing that deterministic tools pushed AI biology accuracy from 16.9% to 92.8%. The technical foundation is there — now it needs to deliver real molecules.

The Biology Benchmark War

OpenAI responded within days by publishing GeneBench-Pro, a harder biology benchmark for AI agents. The pattern mirrors what happened with SWE-bench in coding: one company establishes a benchmark, the other responds with a harder version. Biology has officially become the next AI frontier after coding.

What's at Stake

The pharmaceutical industry represents a $1.5 trillion market. If Anthropic can help cut development times and improve success rates, the financial implications are staggering. Claude Science's introductory pricing and the VirBench results suggest Anthropic is aiming to dominate this vertical the way it has with enterprise coding.

Claude Science is available now in beta. Anthropic's own drug candidates are expected to enter preclinical development within the next 12-18 months.

Sources