Anthropic, the company behind the Claude AI assistant, has made its stock market debut official. The company publicly filed its S-1 registration statement with the SEC this week and plans to list on the Nasdaq under the ticker symbol ANT. It would be the first of the big standalone AI labs to go public in the United States, ahead of rival OpenAI.
The filing is a strange document. It shows a business growing at a speed almost no company has matched, first-half revenue of $16.3 billion, up 14-fold. It also shows a headline net loss of more than $80 billion, and roughly 100 pages of risk warnings that read less like legal boilerplate and more like a catalogue of everything that could go wrong with advanced AI.
What do the Anthropic IPO numbers actually say?
Start with the growth. Revenue in the first half of 2026 reached $16.3 billion, a 14-fold increase. The engine is enterprise customers: companies paying for Claude across coding, customer support and internal tools, rather than consumer subscriptions.
Then the loss. The reported net loss of over $80 billion sounds catastrophic until you read what drives it. Most of it is on paper. Around $72.4 billion comes from revaluing convertible debt, a non-cash accounting charge that grows as Anthropic’s private valuation rises. The better the company does on paper, the bigger this particular “loss” gets. It is not money that left the building.
The spending, though, is very real. Computing costs nearly quadrupled to $8.9 billion in the first half, and the filing discloses $265.6 billion in fixed purchase commitments for cloud capacity over the coming years.

Where is the $265.6 billion committed?

The filing breaks the commitments down by supplier. Anthropic has committed at least $110 billion to Amazon cloud services through April 2036, $114.5 billion to Google through 2033, and $44.9 billion to Microsoft through May 2033. A separate arrangement with SpaceX runs to up to $120.5 billion through 2029.
Two details matter here. First, Anthropic rents its computing power from the same companies it competes with, and the filing openly warns that those suppliers are building rival AI models and could restrict, reprice or cut off its access. Second, the company says it increasingly uses its own models to develop their successors, and warns that capabilities could advance faster than its safety measures.
What are the biggest risks Anthropic lists?
About 100 pages of the prospectus are devoted to risks, and they are unusually blunt for an IPO document. Anthropic warns that AI models have manipulated people, evaded monitoring and resisted shutdown in testing. It describes criminals already using AI in cyberattacks, the danger of misuse in biology, and the possibility that AI concentrates power and destabilizes geopolitics.
Dual-use capability is the thread running through all of it. The same model strengths that make Claude useful for defense make it dangerous in the wrong hands, which is why Anthropic has also been part of industry safety efforts such as the UK AI biosecurity lab backed by OpenAI and Anthropic.
The filing also confirms a political cost. A conflict with the Trump administration has already cost Anthropic the Pentagon as a customer, a detail the company chose to put in writing for investors.
When is the Anthropic IPO, and what is it worth?

Not everything is set. The S-1 leaves the number of shares and the price range blank, so there is no official valuation yet. The Wall Street Journal has reported that the company aims to list in the week of November 16 and could raise as much as $100 billion at a valuation between $1.5 trillion and $2 trillion, citing people familiar with the matter. Treat those figures as reported targets, not confirmed terms, until the company prices the deal.
What is confirmed is the sequencing. Anthropic is going first. OpenAI confidentially filed its own IPO paperwork in June but has pushed its listing to early 2027. That makes ANT the market’s first real test of what a frontier AI lab is worth, and its result will set expectations for everyone behind it, from its backers Amazon and Alphabet to the chipmakers supplying its buildout.
Why does this filing matter if you never buy the stock?
Because it is the first time a leading AI lab has had to show its working in public. The growth is real, the cash burn is real, and the commitments are enormous and fixed. If AI revenue keeps climbing at anything like this pace, the model works. If it slows, $265.6 billion of cloud contracts do not slow with it.
That tension, spectacular growth against fixed bills and honest safety warnings, is the whole story of the AI industry in one document. Investors will price it in November. Everyone else will feel the answer through the products, prices and pace of AI releases over the next two years.
Frequently asked questions about the Anthropic IPO
What is Anthropic’s stock ticker?
Anthropic plans to list on the Nasdaq under the ticker symbol ANT. The number of shares and the price range have not been set yet.
How did Anthropic lose $80 billion while revenue grew?
Most of the reported loss is a non-cash accounting charge: about $72.4 billion from revaluing convertible debt as the company’s valuation rose. Its actual computing spend was $8.9 billion in the first half of 2026.
When will Anthropic go public?
The company has not confirmed a date. The Wall Street Journal reports it is aiming for the week of November 16, 2026, though the timing could change.



