Anthropic estimates that revenue from customers it later found to be using its models in violation of its usage policy or commercial terms was about 5% to 7% of its revenue for the nine months ended 30 September 2026, according to the S-1 registration statement it filed with the SEC on 26 October 2026.
The disclosure sits in the filing’s discussion of the factors that affect its results. Kalshi posted it on X on 29 October, describing it as a share of Anthropic’s 2026 revenue; the filing’s period is the first nine months of 2026.
Business Pill · Run-Rate Revenue
An evergreen Business Engineer explainer on run-rate revenue. It explains the general idea only and does not discuss this filing.
The key insight: Anthropic’s own estimate puts revenue from customers later found to breach its usage policy or commercial terms at about 5% to 7% of nine-month revenue; by our arithmetic, that is no less than roughly $814 million to $1.14 billion.
The S-1 states: “Based on our detection and review process, we estimate that revenue from such customers represented approximately 5% to 7% of our revenue for the nine months ended September 30, 2026.”
It explains the timing: “Because detection, investigation, and enforcement take time, we recognize revenue from such customers when the related services are provided, before restricting or terminating their access.”
Anthropic adds that “This estimate requires judgment and may change as our detection methodologies evolve and we identify previously undetected usage.” It also writes: “We believe this dynamic is common in our industry.”
Anthropic’s estimated range applied to its first-half 2026 revenue of $16,286 million, as a lower bound; the nine-month revenue total is not in the S-1. Source: Anthropic Form S-1, 26 Oct 2026. Dollar figures are our arithmetic.
What Was Posted on X
BREAKING: 5% to 7% of Anthropic’s 2026 revenue came from customers violating its usage policy
Kalshi’s post says 5% to 7% of Anthropic’s 2026 revenue came from customers violating its usage policy.
What Anthropic Expects Next
The filing says Anthropic expects the share “to decline in future periods as we have enhanced our detection and enforcement efforts, in particular in August and September 2026; however, the timing and magnitude of any resulting adverse effect on our revenue and revenuegrowth are uncertain.”
Elsewhere in the S-1 it warns that “customers whose access is restricted or terminated may re-register under new accounts” and that the enforcement work “may reduce our overall revenue and growth rates in future periods”.
The Size of the Range
The S-1 reports revenue of $16,286 million for the six months ended 30 June 2026, and quarterly revenue of $4,731 million for the March quarter and $11,555 million for the June quarter. It does not report a revenue total for the nine months ended 30 September.
Because the nine-month total is at least the first-half figure, our arithmetic puts the range at no less than roughly $814 million to $1.14 billion of revenue (5% and 7% of $16,286 million). The real figure depends on September-quarter revenue, which the filing leaves out.
Related Context
The filing does not say which customers or what kinds of violation the estimate covers. Anthropic has separately written about distillation, saying: “Early evidence suggests that these methods have reduced the amount of data harvested by distillers by around 90 to 99%.” According to The Information, as reported by Reuters, DeepSeek and Moonshot AI relied heavily on outputs from OpenAI’s GPT-6 Astra to train their latest models.
The Structural Read
As we read the filing, the disclosure is about timing: Anthropic says it books the revenue “when the related services are provided, before restricting or terminating their access”.
The S-1 says the estimate “requires judgment and may change as our detection methodologies evolve and we identify previously undetected usage”, so the range can move in later filings.
The filing links the expected decline to enforcement stepped up “in August and September 2026”, and says the effect on revenue and growth is uncertain in timing and size.
Anthropic Form S-1, 26 October 2026
“We believe this dynamic is common in our industry.”
Three Implications
FOR ANTHROPIC’S GROWTH RATE The S-1 says enforcement “may reduce our overall revenue and growth rates in future periods and contribute to variability in our results from period to period”.
FOR IPO INVESTORS The filing gives a percentage range for nine months and no dollar figure; the September-quarter revenue needed to convert it is not in the S-1.
FOR ENFORCEMENT The S-1 warns that “customers whose access is restricted or terminated may re-register under new accounts”.
The framework argues that Anthropic “is becoming the first major AI lab actively attempting to define the governance layer of the AI stack.”
The framework does not discuss this S-1 disclosure or the revenue estimate.
What Is Not Established
The filing gives a range, not a dollar figure, and does not break it down by violation type, product or customer. It does not say how much of the estimated revenue came from accounts already terminated, or whether any of it was refunded.
The September-quarter revenue figure needed to turn the percentage into dollars is not in the filing. The lower bound above uses first-half revenue only and is our arithmetic, not an Anthropic disclosure.
This Week In The AI Supercycle
Execution Fault Lines, Concentrated Demand, and Geopolitical Firewalls
This week’s issue: the AI build is not short of demand for its parts. It is short of cheap money, and its end revenue rests on a narrow base of large, often venture-funded buyers.
Anthropic’s S-1 estimates that about 5% to 7% of its revenue for the first nine months of 2026 came from customers it later found to be breaching its usage policy or commercial terms, and says it expects that share to fall after stepping up enforcement in August and September.
A note on sourcing. The estimate and every quotation come from Anthropic’s Form S-1 as filed with the SEC on 26 October 2026, read on 29 October 2026. The dollar lower bound is our arithmetic. The distillation context is attributed to Anthropic’s own post and to The Information as relayed by Reuters.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
