Anthropic has made its S-1 registration statement public, giving investors the first audited look inside one of the world's most valuable AI companies. The Claude maker filed with the US Securities and Exchange Commission on October 26 for a proposed initial public offering of Class A common stock and plans to list on the Nasdaq Global Select Market under the ticker ANT. The prospectus shows revenue growing at an extraordinary pace alongside a very large paper loss, heavy compute commitments and a governance structure that keeps control firmly with the founders.
What Anthropic Filed
Anthropic's own announcement is brief. The company says the number of shares and the price range have not been set, and that the registration statement is not yet effective, so no shares can be sold or bought until it is. Goldman Sachs and Morgan Stanley are listed as the contacts for the preliminary prospectus, consistent with reports that they lead the underwriting group. Coverage of the filing says the listing is planned for later this fall, but no trading date has been announced.
The company was valued at $965 billion in its last private round. Reports cite IPO valuations of $1.5 trillion to $2 trillion under discussion, but those figures are not in the filing itself.
The Numbers: Fast Growth, Large Paper Loss
According to coverage of the prospectus by Yahoo Finance, the headline figures are:
- First-half 2026 revenue of $16.3 billion, compared with $4.6 billion for all of 2025
- Second-quarter revenue of $11.6 billion, up from $4.7 billion in the first quarter
- Net loss of $80.7 billion for the half, versus $42 billion for all of 2025
- Second-quarter operating income of $974 million, after a $1.7 billion operating loss in the first quarter
- Compute costs of $5.6 billion in the second quarter, up from $3.4 billion
The swing to an operating profit in the second quarter is the more telling number for investors trying to judge whether frontier AI can be a profitable business. Revenue per employee rose to more than $6.6 million, with headcount of about 5,300 as of September 30.
Compute Commitments and Customer Concentration
The filing makes clear how much capital Anthropic has promised to future infrastructure. Fortune reports that the company has committed $542 billion to $567 billion to future compute and infrastructure. Yahoo Finance highlights an agreement to spend up to $120.5 billion with SpaceXAI through 2029, and a hybrid model the company calls Anthropic Direct, in which it buys accelerators directly and leases data centre space rather than relying only on cloud providers.
Concentration is another theme. According to Fortune's reading of the prospectus:
- The top five customers accounted for about 25% of revenue in the first nine months of 2026
- Cloud marketplaces run by Amazon, Google and Microsoft generated 48% of first-half revenue
- Between 5% and 7% of revenue came from customers later found to be breaking usage policies or commercial terms
Governance, Risks and an Unusual Lockup
Anthropic is a public benefit corporation, and the S-1 is built around that identity. Seven co-founders will hold a special class of stock carrying 50.1% of shareholder voting power, and the Long-Term Benefit Trust selects a majority of the board. The filing warns that this structure may effectively remove Class A shareholders' ability to influence major decisions, including a change of control.
The risk factors run to roughly 94 of the filing's approximately 300 pages, according to Fortune. They include a warning that AI could pose a "catastrophic or existential risk to humanity," the company's ongoing dispute with the Department of War over its "supply-chain risk" designation, June export restrictions that forced it to disable its Fable 5 and Mythos 5 models, and the possibility that insurers cannot cover a major AI incident. The company also acknowledges that safety-driven slowdowns may not maximise shareholder value.
The lockup is also unusual. Some employees and insiders can sell from the first trading day, with unlocks staggered across more than 20 windows over about a year, rather than a single standard 180-day lockup.
Why It Matters
Anthropic's filing is the first time a frontier model developer of this scale has had to open its books to public-market disclosure standards. It gives the wider AI economy hard data on three questions investors have been debating: how fast enterprise and developer demand for models is growing, how much compute that growth requires, and whether margins can turn positive. The second-quarter operating profit suggests they can, at least for now.
The filing also lands in a fragile market. The Nasdaq Composite rose about 0.8% on the day of the filing as chip stocks rallied, but AI-linked shares have been volatile this month and Big Tech earnings are testing investor patience with capital spending. How Anthropic prices its offering, and how the stock trades, will shape the funding environment for every AI company that follows, including OpenAI, which is reportedly preparing its own listing.
