Anthropic acquires biotech startup to test Claude in drug discovery

A move into life sciences

Anthropic is making a significant push into the healthcare and life sciences sector. The company recently filed an S-1 document listing this industry as its primary target for new market growth. This move is notable because the global pharmaceutical market is expected to exceed $2 trillion by 2026.

The company aims to use its work in biology to improve public sentiment toward artificial intelligence. By applying its technology to life-saving medicine, Anthropic hopes to demonstrate the practical and positive utility of its models. They are targeting a space where nine of the ten largest pharmaceutical companies are already potential customers.

This strategy follows a previous acquisition of Coefficient Bio, a startup with fewer than 10 people. Anthropic spent about $400 million on that deal to acquire tools for planning drug research and managing clinical regulatory strategies. These moves show a coordinated effort to build a comprehensive suite of biotech tools.

Testing Claude on real drugs

Anthropic recently acquired Mithril, a one-person biotech company, to test its Claude model on actual drug programs. The company stated that running these programs will show them exactly where the AI holds up and where it fails. This real-world testing provides the company with necessary data to refine its technology for complex scientific use cases.

The acquisition also allows Anthropic to build case studies that it can share with other biotechnology firms. By documenting how the AI handles specific hurdles, they can provide a roadmap for other companies looking to integrate Claude into their workflows. This creates a feedback loop where real-world results inform the next generation of AI capabilities.

Mithril currently possesses a lead antibody that has only been tested in mice. Anthropic plans to keep the work mostly in the preclinical stage for now while it extends Claude's capabilities into clinical trials. If the drug candidates work, Anthropic could eventually own the medicines produced through these programs.

Efficiency in paperwork

The practical benefits of using Claude in biotech are already appearing in administrative tasks. Enveda, a company that uses AI to turn nature's chemistry into medicine, partnered with Anthropic to handle manufacturing records. These records are mandatory documents that drugmakers must file before they can begin a clinical trial.

Before using the AI, Enveda's chemists spent up to 75% of their time reviewing these documents. The Claude-based tool they built was able to catch all 11 issues found by human reviewers across a 1,061-page filing. It also flagged seven issues that the human reviewers had completely missed during their manual check.

The tool reduced the review time from almost 26 reviewer-hours to just under three hours. This represents an 89% reduction in time for that specific use case. Such efficiency gains are significant because healthcare buyers are already beginning to budget for AI-driven administrative work.

The competitive landscape

Anthropic is not the only major AI player moving into the pharmaceutical space. OpenAI recently agreed to acquire Orchestra Bio, a platform that uses AI agents to monitor study timelines and workflows. While Anthropic is focusing on drug discovery and production, OpenAI is betting on the management of the long drug development process.

Moving a drug from discovery to regulatory approval typically takes between 10 and 15 years. Orchestra Bio's agents watch these long timelines and can identify issues, such as a failed manufacturing batch. When a failure occurs, the agents identify which dates and activities are affected and present options to the human leaders.

OpenAI is integrating Orchestra into its Rosalind line of tools, which is already used by companies like Amgen and Moderna. This competition suggests that the biotech industry is becoming a primary battleground for AI providers. Both companies are trying to capture a piece of a massive market by solving different parts of the drug development lifecycle.