Anthropic just told the world it plans to develop drugs of its own. The company made the announcement at its event “The Briefing: AI for Science” earlier this week, according to The Verge AI, alongside the launch of Claude Science, a new “AI workbench for scientists.” That’s the headline most people missed. The workbench is interesting. The drug ambition is the real story.
What Anthropic announced
Claude Science pulls fragmented research tools and datasets into one environment and generates figures and visuals for scientists. Anthropic framed it around AI’s potential to “dramatically accelerate the pace of scientific discovery and the development of healthcare interventions,” and pointed to a long list of biotech and pharma customers already using Claude.
Then it went further. Head of life sciences Eric Kauderer-Abrams said Anthropic will develop treatments for “neglected” diseases itself. As The Verge AI reports, that puts Anthropic in an unusual spot: selling software to drugmakers while becoming a drugmaker that competes with them.
Why this matters
Every major AI lab is chasing science and pharma customers. OpenAI, Amazon, and Google all have life sciences tools. What stands out here is that Anthropic isn’t just selling picks and shovels. It wants to mine the gold too.
That’s a direct move. It drops Anthropic into a crowded race that already includes AI-first drug companies like Insilico, Google DeepMind spinout Isomorphic Labs, and Big Pharma outfits building or buying AI tools of their own.
Here’s the catch: Anthropic has shared almost no specifics. Kauderer-Abrams didn’t say what happens if the company finds a promising candidate. Anthropic didn’t respond to The Verge AI’s questions about which diseases it will target first, or whether it will partner for lab work, animal testing, clinical trials, or manufacturing.
The reality check from experts
Experts told The Verge AI that “AI drug discovery” is a fuzzy term. Namshik Han, a University of Cambridge professor and cofounder of AI biotech startup CardiaTec, called it “a really broad term,” noting AI now touches “every single stage of drug discovery.” Matthew Todd of University College London called it a “catchall phrase.”
Here’s what AI is genuinely good at today:
- Suggesting new molecules that could interact with known disease targets
- Speeding up research and helping “road test” new ideas
- Finding connections across vast chemical and biological data
- Spotting new uses for existing drugs
And here’s where it hits a wall. Todd said the field is “a long way off” from a regulator approving an AI-designed drug for human use, and that the process still needs human supervision throughout. Frank von Delft, a University of Oxford professor, put it plainly: AI models “haven’t yet come close to making experiments unnecessary.”
Drug candidates still have to be tested in the real world for efficacy, toxicity, and whether they can be stored and delivered safely. That takes skilled people, serious money, and time, especially the human clinical work where most promising candidates fail. If Anthropic wants a real drug, von Delft said, it’s “going to have to spend a lot on experiments.”
Anthropic looks serious about spending
This isn’t just talk. Over the past year, Anthropic has been hiring biologists and building its own wet labs, with several live listings for life sciences roles. Han said the company has been “actively recruiting” and that several of his academic colleagues were approached. He believes Anthropic has already pulled a few hires from Big Pharma and top academic institutions.
One more constraint worth naming: both Han and Todd flagged a shortage of high-quality, publicly available experimental data on how chemicals behave in the body. That gap slows everyone down, no matter how strong the model.
What to expect
Don’t hold your breath for a Claude-designed pill. Any payoff is likely the better part of a decade away, given how long clinical trials take. “It takes time to show experimentally that something’s safe,” Todd said. No AI-designed drug has made it through clinical trials yet.
So watch two things. First, whether Anthropic names a disease target and a partner for the physical lab work. Second, whether its pharma customers get nervous about buying tools from a company that now wants to beat them to market. You can find the full details at the original report from The Verge AI.