Anthropic’s S‑1 filing lays out a company that has transformed in a few years from a research startup into a revenue-generating AI business — while simultaneously committing to staggering multi‑year compute deals and reporting enormous accounting losses. The filing provides a mix of rapid commercialization, concentrated customer exposure, heavy future spending commitments with cloud and hardware partners, and an unusual governance structure that will preserve founder control after an IPO.
Rapid revenue growth
Anthropic’s reported revenue shows an accelerating trajectory: $41 million in 2023, $386 million in 2024, and $4.59 billion in 2025. The firm reported $16.29 billion in revenue in the first six months of 2026, up from $1.16 billion in the same period the prior year — roughly a 1,300% year‑over‑year increase for the half‑year comparison.
Massive losses and compensation liabilities
Despite the top‑line surge, Anthropic disclosed an $80.75 billion net loss for the first half of 2026, versus a $12.97 billion loss in the same period the year before. About $72.39 billion of the 2026 loss was tied to changes in the fair value of convertible notes and other financial instruments. The company also warned that stock‑based compensation expenses will rise substantially once the IPO completes, noting approximately $26.6 billion in unrecognized stock‑based compensation associated with restricted stock units as of June 30, 2026.
Customers and revenue mix
Anthropic reported a sizable base of high‑value customers: as of September 30, 2026, 11,900 customers generated at least $100,000 in annual revenue; 1,900 customers produced at least $1 million in trailing annual revenue; and 250 customers spent at least $10 million annually. In the first half of 2026, roughly 80% of revenue came from consumption‑based pricing; of that consumption revenue, about 48% flowed through cloud marketplaces such as Amazon Web Services, Google Cloud, and Microsoft Azure. The filing also noted customer concentration: approximately 20% of first‑half 2026 revenue came from two customers, and for the nine months ended September 30, Anthropic’s five largest customers accounted for roughly 25% of revenue.
Enormous compute and vendor commitments
A central theme of the filing is Anthropic’s long‑term commitments for compute capacity and hardware. The company disclosed multi‑year agreements and leases that, in aggregate, total in the hundreds of billions of dollars. Specific items listed include $161.2 billion in equipment leases involving Broadcom and Google TPUs, $114.5 billion with Google Cloud, $110 billion with Amazon, and $44.9 billion with Microsoft. Anthropic also identified GPU capacity deals of up to $120.5 billion with SpaceXAI, $44.6 billion with Nscale, and $41.3 billion with Lambda. Beyond contracted deals, the filing says Anthropic expects to spend another $125 billion to $150 billion on TPUs and more than $20 billion on AMD‑based systems.
Governance, control, and risk disclosures
Anthropic’s proposed governance structure would leave its seven co‑founders with 50.1% voting power after the IPO via a single special Class F share held by a founder‑controlled entity, meaning public investors would not gain majority voting control. The company’s Long‑Term Benefit Trust would also retain the ability to elect a majority of the board. The S‑1 includes more than 90 pages of risk factors, covering financial risks, competitive pressures, computing constraints, cybersecurity, and the potential for advanced AI systems to behave in unintended ways.
Outstanding IPO details
The filing did not specify the number of shares to be offered, an IPO price range, or a confirmed trading date. Press reports cited alongside the filing suggested the company might seek roughly $100 billion in proceeds at a potential valuation between $1.5 trillion and $1.75 trillion, but those figures were not finalized in Anthropic’s S‑1.
Licensed under CC BY-ND 4.0 International