Anthropic's IPO filing reveals in-house drug lab and nine of top 10 pharma as customers

Anthropic filed its S-1 registration statement with the US Securities and Exchange Commission on 26 October 2026, and the document sets out a bigger role in drug research than a software supplier would normally claim. The company plans to list on Nasdaq under the ticker ANT and has not yet disclosed a price range, a share count or a listing date¹².

What the filing says

The prospectus describes a two-sided life sciences business. On one side, Claude is sold to drugmakers: Anthropic says it already works with nine of the top 10 big pharma groups, and it has partnerships with Novo Nordisk and Bristol Myers Squibb¹². Its Claude for Life Sciences platform is pitched across the development chain, from early discovery and preclinical work to clinical trial operations and regulatory submissions, and links into electronic lab notebooks and scientific databases¹.

On the other side, Anthropic is doing research of its own. The filing lists hiring industry specialists, setting up a molecular biology laboratory and buying Coefficient Bio, a computational biology start-up¹². That deal, reported in April at $400m, was paid for with 15.4 million Anthropic shares, part of which depends on staff staying on¹. A separate Model Hardware Standard is meant to let AI agents run lab and manufacturing kit such as microscopes, liquid handlers and robotic arms¹. In the company's words, "Claude is used in drug discovery and development by our customers, our biotechnology partners, and our own researchers"².

The financial backdrop is large on both lines. Anthropic reported revenue of $16.3bn for the six months to 30 June 2026, against losses of more than $80.7bn². Estimates of the raise differ: a New York Times report from August put it above $100bn with a valuation approaching $2 trillion, while Renaissance Capital has pointed to about $75bn¹². The filing frames the sector opportunity by citing expected global pharmaceutical spending of more than $2 trillion in 2026¹².

What it means for the market

For AI-first drug discovery companies, the filing confirms that the largest model developers intend to compete for discovery work, not only sell tools to those who do it. Anthropic says the platform supplies drugmakers while its own lab builds the capacity to pursue discoveries directly¹. Pharma buyers that use Claude will need to weigh that dual role when they decide what data and workflows to put through it.

For investors, the listing would be an early public test of a pure-play AI developer, with OpenAI also planning to go public². The scale of the losses relative to revenue will feed the argument over whether AI valuations have run ahead of earnings, a question on which economists and investors remain divided².

The filing also sets out regulatory exposure that health and pharma buyers will recognise. Anthropic warns it could fall under rules on drugs, biologics, medical devices and protected health information, overseen by the FDA and its counterparts abroad, and that regulators could require new authorisations or restrict some uses¹. It flags that errors, inaccurate results and hallucinations could contribute to patient harm or flawed research, and states: "We may be unable to fully understand or explain why our models produce particular outputs or behave in particular ways"¹.

Caveats

The company concedes that lab work involving biological and chemical materials sits outside its history as a software organisation and could bring liabilities such as workplace injury and contamination¹. Any drug candidate it finds would face "lengthy, costly, and uncertain development and regulatory pathways" and might never earn revenue¹. Its standing with the US government is also uneven: the Department of Health and Human Services barred staff from using Claude in March, and the FDA has since moved its Elsa review tool to Google's Gemini¹. The listing could take place in November².