So here's a headline I genuinely did not expect on a Friday morning: Anthropic — the company whose CEO spent last week asking the entire AI industry to slow down — has been quietly building an Anthropic biology lab. Not a simulation, not a software product — an actual wet lab with benches and pipettes in the San Francisco Bay Area, confirmed by Reuters in an exclusive report on September 18.
01.What Anthropic just confirmed
Quick primer on what a "wet lab" means, because the term gets thrown around loosely. Most AI-for-biology work so far has been in silico — simulations running on computers. A wet lab is where you do physical experiments with real cells, real proteins and real chemicals. Until now, that was the domain of pharma companies and biotech startups, not AI labs.
Anthropic's head of life sciences, Eric Kauderer-Abrams, confirmed the lab to Reuters and was direct about why: "We believe that to do biology, the final test is still and will be for a while in real lab work." He described the setup as typical of most biotech companies — some work in Anthropic's own facilities, some with external partners. A spokesperson later clarified that the lab is not for drug discovery specifically — which raises more questions than it answers.
02.Claude will direct robots in the lab
Here's the detail that should stop you mid-scroll. According to one of Reuters' sources, Anthropic wants its Claude AI to direct robotic units carrying out science experiments with limited human intervention. The company insists human oversight remains essential for safety, and Kauderer-Abrams admits they're in "the very early innings of using AI to automate the execution of lab work."
Anthropic already launched the Model Hardware Standard in August — a way for AI to operate equipment. This is the same basic idea as AI "computer use" (which we cover in depth in the GPT-6 Astra — Zero to Hero Guide), except instead of clicking buttons in a browser, the AI is running an experiment bench. If you've been wondering how far AI agents have moved beyond chat windows — this is your answer. They're holding the pipettes now, or at least telling the robots which ones to hold.
03.The target: diseases pharma ignores
Why would a company valued at nearly $2 trillion bother with test tubes? Two reasons.
First, the commercial logic: Anthropic says it will focus on preclinical work in areas that traditional pharma doesn't find financially attractive — rare diseases and conditions previously thought "undruggable," including complex molecules like bispecific and trispecific antibodies. To avoid competing with its own customers (it already works with Roche's Genentech, Bristol Myers Squibb and Novo Nordisk), Anthropic has drawn a hard boundary: no clinical trials, no bringing drugs to market itself.
Second, the human part: for CEO Dario Amodei, it's personal. A disease killed his father only a few years before a cure arrived. "By far, we see the biggest opportunity for AI in the life sciences," Kauderer-Abrams told Reuters, "and that is motivating everything that we're doing."
04.The quiet money trail
Connect the dots and a picture emerges: the acquisition of startup Coefficient Bio for roughly $400 million in stock (Anthropic confirmed the deal but not the price), Novartis CEO Vas Narasimhan joining its board, the launch of Claude Science software, and job postings that read "our goal is to speed up progress in the life sciences by an order of magnitude." Life sciences, Kauderer-Abrams says, is already one of Anthropic's biggest investment areas by headcount and resources.
If you've been following our blog, you'll remember we wrote about Anthropic's life sciences program opening up Claude for bio research — that was the software side. This wet lab is the physical escalation of the same strategy.
05.The timing is the strange part
Now hold on, you might say. Isn't this the same Anthropic whose researchers spent the past two weeks warning about extinction risk? The same one whose CEO just published an essay calling on the industry to slow down, days before a potential $2 trillion IPO (which we covered in detail earlier)?
Yes. And that's exactly what makes this story worth reading twice. Kauderer-Abrams' own framing is that Anthropic is "always balancing" — making models better at accelerating medicine while being careful about deployment to limit risk. Whether you find that reassuring or worrying probably says a lot about how you feel about AI in general. My honest take: an AI company doing real, careful lab work on neglected diseases is one of the better applications of this technology I've read about all month. The open question is whether the pace of ambition stays matched to the pace of caution.
For the academic take on the same idea, read how Stanford's virtual biotech ran 37,000 AI agents to design a cancer drug.
Key Takeaways
- Anthropic has confirmed a physical wet lab in the San Francisco Bay Area — beyond computer simulations
- Claude is being positioned to direct robots running experiments with limited human intervention
- Focus areas: rare and "undruggable" diseases, preclinical work only — no clinical trials
- The push is backed by the ~$400M Coefficient Bio acquisition, Claude Science software and Novartis' CEO on the board
- It lands amid extinction warnings, a slowdown call and a potential $2 trillion IPO — an unusual combination
Frequently Asked Questions
What is Anthropic's biology lab?
Anthropic has built a physical wet lab in the San Francisco Bay Area where Claude will direct robotic units running experiments with limited human intervention.
What kind of work will the lab do?
It will focus on preclinical work on rare and “undruggable” diseases. Anthropic says it will not run clinical trials itself.
How is Anthropic building out its life sciences push?
The company already bought Coefficient Bio for about $400 million and added the Novartis CEO to its board.
What This Means for You
For Indian readers, this story matters more than it first appears. India is one of the world's largest producers of generic medicines, and our pharma hubs in Hyderabad and Bengaluru are already investing heavily in AI-driven research. If AI-directed labs compress early-stage drug discovery, Indian companies and researchers who understand both biology and AI will be in enormous demand. For students, the signal is clear: the intersection of life sciences and AI is about to become one of the hottest career paths of the decade. And if you want to be ready for it, start by learning to direct today's AI agents — the GPT-6 Astra — Zero to Hero Guide (₹149) is the fastest structured path we know, covering prompting, agents, cost optimization and safety.