Anthropic has quietly constructed a wet lab in the San Francisco Bay Area as it aggressively pushes artificial intelligence into drug development—a move that raises urgent questions about corporate accountability, public benefit, and who actually bears the risk when tech companies enter medicine.
The startup's head of life sciences, Eric Kauderer-Abrams, confirmed the lab's existence to Reuters on Tuesday, describing it as essential for moving beyond computer simulations into physical experiments. "We believe that to do biology, the final test is still and will be for a while in real lab work," Kauderer-Abrams said. Yet Anthropic has offered remarkably little transparency about what diseases it's targeting, what progress it's made, or how it will ensure public safety as it automates drug discovery with AI systems that some of its own researchers warn could pose existential risks.
The company frames this work as addressing market failures—diseases the pharmaceutical industry considers unprofitable. That's a compelling narrative. But it obscures a deeper structural problem: Anthropic is building pharmaceutical capabilities while remaining accountable primarily to investors, not patients or regulators. The company is preparing for a potentially $2 trillion IPO. That's a staggering amount of capital raised on the promise of AI transformation, yet the actual therapeutic benefit remains speculative.
The Public Health Gap
Anthropic says it wants to unlock treatments for rare diseases the industry neglects. That's genuinely important work—there are thousands of conditions affecting small populations that receive minimal research funding. But Anthropic isn't a nonprofit. It's a for-profit company racing to demonstrate that AI justifies its astronomical valuation and the jobs it displaces. One person familiar with the effort told Reuters the company is trying to deliver tangible results "before staff and the public lose faith that AI justifies erasing jobs and possibly human life."
That's not a health mission. That's a legitimacy crisis dressed up as medicine.
The company acquired Coefficient Bio for roughly $400 million in stock and added Novartis CEO Vas Narasimhan to its board—moves designed to accelerate its entry into drug development. It launched Claude Science software and is working with major pharmaceutical companies including Roche's Genentech, Bristol Myers Squibb, and Novo Nordisk. Yet the precise diseases Anthropic is targeting remain unclear. Its progress is opaque. And once a molecule is identified, it can take years to bring a drug to market through human trials—a challenge Anthropic hasn't yet undertaken.
Automation Without Democratic Oversight
Anthropic is embracing physical automation, with plans to use Claude AI to direct robotic units to carry out experiments with limited human intervention. The company launched the Model Hardware Standard in August to help AI operate laboratory equipment. Kauderer-Abrams described this as "an area that has the potential to bring about meaningful acceleration in so many different processes."
What he didn't describe: the regulatory framework. Who oversees these experiments? What happens if an AI system makes an error in a lab that could have biological consequences? Anthropic says human oversight and involvement are essential for safety, but the company is racing toward automation precisely to reduce human involvement and accelerate timelines.
This isn't unique to Anthropic. Isomorphic Labs, Alphabet's drug discovery arm, has spent years trying to move from discovery to clinical trials and earlier delayed that goal to the end of 2026. But Alphabet operates within established corporate structures with clearer accountability lines. Anthropic is a relative newcomer to biology, moving fast, and offering limited transparency about how it's balancing speed against safety.
The Trust Problem
Anthropic faces what two people familiar with its efforts called a "trust problem." Pharmaceutical companies worry that Anthropic will learn from their proprietary drug programs, even if the AI startup claims to wall off their data. That concern cuts both ways: Anthropic's customers can't verify the company's data protections, and the public has no mechanism to monitor what Anthropic learns or how it might use that knowledge.
Kauderer-Abrams said Anthropic has set boundaries, including not running clinical trials for now and focusing on diseases industry wouldn't address. "We're not competing with pharma and biotech companies that make their business in bringing drugs to market," he said. But that statement assumes Anthropic's intentions remain fixed. A company planning a $2 trillion IPO has powerful incentives to expand its addressable market.
CEO Dario Amodei framed this work as personal—a disease killed his father only a few years before a cure emerged. That's a human motivation. But it's also a reminder that even well-intentioned companies operate within market logic. Amodei has also called for AI development slowdowns, warning that future AI systems could cause catastrophic damage. Yet Anthropic is simultaneously racing to scale its capabilities and prepare for a massive public offering. Those positions aren't easily reconciled.
Why This Matters:
When private companies move into drug development, the stakes shift from profit margins to human lives. Anthropic's entry into biology could genuinely help patients with rare diseases—that's possible and worth monitoring. But it's happening in a context of minimal public oversight, opaque decision-making about which diseases matter, and corporate incentives that may not align with public health. The company is hiring aggressively to "speed up progress in the life sciences by an order of magnitude," yet it hasn't committed to bringing a single drug through human trials. Most drugs fail clinical testing. The success of any program Anthropic runs is far from certain. What's certain is that if something goes wrong—if an AI-directed experiment causes harm, if proprietary data leaks, if the company pivots toward more profitable diseases—the public will bear the consequences while Anthropic's shareholders capture the gains. That asymmetry demands stronger regulation, clearer accountability mechanisms, and genuine transparency about what Anthropic is building and why.