Anthropic's amended S-1, filed November 2, reports $17.0 billion in revenue for the quarter ended September 30, up from $1.4 billion a year earlier, a $567 million operating loss driven by $5.0 billion of short-term compute, and a risk factor saying it has changed and may change usage limits. The Nasdaq ticker is ANT; no price range yet.
Update, November 2: the same filing discloses about $895.4 billion in compute and infrastructure commitments. Details are in the commitments section below.
Anthropic, the company behind Claude, filed an updated IPO registration with the SEC on November 2 that adds its third-quarter numbers. Revenue for the three months ended September 30 was $17.0 billion, up from $1.4 billion a year earlier. The quarter ended with an operating loss of $567 million after a profitable second quarter, and the filing says the main cause was $5.0 billion of extra compute bought on short-term terms. If your team pays for Claude seats or the API, the useful part is in the risk factors: Anthropic says it has changed usage limits, rate limits and model capabilities before and may do so again.
What the filing says
The amended Form S-1 updates the original S-1 Anthropic filed publicly on October 26. The numbers, all from the filing:
- Revenue: $17.0 billion in the third quarter of 2026, and $33.2 billion for the first nine months, up from $2.5 billion in the same nine months of 2025. Full-year 2025 revenue was $4.6 billion.
- Operating result: an operating loss of $567 million in the third quarter, after operating income of $974 million in the second. Anthropic blames $5.0 billion of compute under short-term commitments, which it says “currently costs significantly more” than compute on its multi-year contracts, plus $2.0 billion of non-cash charges tied to its donation matching program.
- Its own adjusted measure: Anthropic reports Adjusted Operating Income of $2.5 billion, a 15% margin, for the quarter. That is a company-defined figure that leaves out items such as stock-based compensation.
- Net loss: $104.0 billion for the nine months. Of that, $94.3 billion is changes in the fair value of convertible notes, derivatives and SAFEs, accounting items tied to its funding instruments.
- Enterprise customers: more than 11,900 customers paying $100,000 or more a year, up from more than 950 a year earlier. More than 1,900 pay $1 million or more, and about 250 pay $10 million or more.
The company has applied to list Class A shares on the Nasdaq Global Select Market under the symbol “ANT,” with Goldman Sachs and Morgan Stanley leading the underwriters. The price range and share count are still blank. Bloomberg reports Anthropic could start marketing the offering as soon as the week of November 9, according to IPOScoop. Anthropic has not confirmed a date.
What the $895 billion in commitments means
The filing also puts a number on what Anthropic has promised to buy. As of and after September 30, it says it has entered into purchase commitments and lease agreements, including short-term deals and leases not yet started, with an aggregate committed spend of about $895.4 billion. It says substantially all of that is for compute and infrastructure. CNBC reported the figure on November 2.
- Timing: about $447.5 billion is expected before December 31, 2030, and substantially all of the rest between 2031 and 2045.
- Next 12 months: about $69.8 billion falls due in the year after September 30.
- Terms: the filing says some agreements carry minimum purchase and take-or-pay terms, so Anthropic pays whether or not it uses the capacity. It calls these “substantial fixed obligations.”
Anthropic also lists the total as a risk: it “may substantially increase our operating expenses and cash requirements,” and the spending may not produce more revenue.
For a buyer, this adds weight to the usage-limit point above. A company with fixed multi-year compute bills has a strong reason to manage how much a flat-rate plan includes. Nothing in the filing says prices will change, and it announces no plan changes. Treat it as a reason to get limits in writing, not as a prediction.
Claude seat buyers should expect limits to move
Two lines in the filing matter more to Claude customers than the revenue headline.
The first is compute. Anthropic says the extra short-term compute “reduced revenue per dollar of compute compared to the preceding quarter.” In plain terms, demand ran ahead of the cheap capacity it had locked in, and it paid more to keep serving it.
The second is the risk factor on product terms. Anthropic writes that it has “changed, and may in the future change, usage limits, rate limits, or model and product capabilities,” and that some of those changes have led to claims about its disclosures. Subscriptions, the filing says, “are subject to usage limits that reset over short periods.”
My read: a company preparing to sell shares while compute costs squeeze its margin has every reason to keep tuning how much usage a flat-rate plan includes. Anthropic loosened limits on its top consumer plan last month, when Ultra dropped the 5-hour cap for verified accounts. OpenAI did the same for ChatGPT Pro. Neither company publishes hard numbers for those allowances, so neither change is locked in.
What to do: if Claude Code or Claude agents are part of how your team ships, track your actual usage for a few weeks and keep that record. Before you sign an annual Team or Enterprise contract, ask your Anthropic rep to put rate limits and model access in writing. A monthly plan keeps you able to switch if the terms change after the IPO.