Anthropic’s S-1/A gives no revenue line for Claude Code, but it ties the product to three parts of Anthropic’s business model: how customers first find Claude, how much they spend once they use it, and how Anthropic writes its own software. Over the period from the quarter Claude Code entered research preview to the three months to 30 September 2026, Anthropic’s quarterly revenue rose by $16.58 billion, and about 74% of that rise came in the latest two quarters, by our arithmetic from the filing.
The filing calls Claude Code “our coding agent and the product we believe has most concretely defined what autonomous AI work looks like.” This analysis puts what the S-1/A says next to what our own reporting on Claude Code has found, from spending per user to job ads and departures.
Business Pill · The Model and the Harness
A short Business Engineer explainer on the difference between an AI model and the harness around it. It teaches the general idea only and says nothing about the people in this piece.
The key insight: Anthropic’s S-1/A gives no revenue line for Claude Code, but it ties the product to how customers arrive, how much they spend and how Anthropic writes its own code; company-wide quarterly revenue rose by $16.58 billion from Claude Code’s research-preview quarter, about 74% of it in the latest two quarters, by our arithmetic.
Claude Code entered research preview in February 2025, the S-1/A says. In the three months to 31 March 2025, Anthropic’s revenue was $376 million. In the three months to 30 September 2026 it was $16,956 million, about 45x, by our arithmetic.
The quarterly steps in between were $787 million, $1,382 million, $2,040 million, $4,731 million and $11,555 million. Of the $16.58 billion added over the whole period, $12.23 billion, about 74%, came between the three months to 31 March 2026 and the three months to 30 September 2026, by our arithmetic.
Costs moved more slowly. Compute and infrastructure expense rose from $944 million to $9,863 million a quarter, about 10x. By our arithmetic, that is $2.51 of compute and infrastructure per dollar of revenue at the start and $0.58 in the latest quarter; revenue first passed that cost line in the three months to 31 March 2026. The filing reports these lines for Anthropic as a whole and does not attribute any of the change to Claude Code.
Average monthly spend per Claude Enterprise user in the three months to 30 September 2026: users who did not use Claude Code spent less than $75; the top 5% of users who used Claude Code spent approximately 20 times that amount. Anthropic’s own metric, not independently verified; circle areas are our arithmetic. Source: Anthropic S-1/A, filed 2 November 2026.Anthropic’s quarterly revenue and compute and infrastructure expense, $ millions, in the three months to 31 March 2025 (the quarter Claude Code entered research preview) and the three months to 30 September 2026. The 45x, 10x and per-dollar figures are our arithmetic; company-wide, the filing gives no Claude Code revenue line. Source: Anthropic S-1/A, filed 2 November 2026.
How Claude Code Earns Money
The filing says Anthropic earns revenue “as customers use Claude through our platform and applications, including Claude Code, Claude Cowork, and vertical solutions for key industries.” Customers pay per use, through subscriptions, or both, depending on the product they choose.
“Revenue is recognized when customers use our intelligence service, based on the quantity of tokens consumed,” the S-1/A says. Subscriptions “are subject to usage limits that reset over short periods”. As we read it, that makes a coding agent that runs for hours a direct driver of metered revenue: the more work Claude Code is allowed to do, the more tokens a customer uses.
Our piece on Cat Wu, whose X profile reads “claude code + cowork @anthropicai”, quoted Wu on that shift: “When we first launched, most people approved every permission prompt by hand and mostly used Claude to answer questions and write tests.” Wu added: “Today, the default is trusting Claude’s judgment to auto-approve. Claude does large migrations and builds new products.”
The Funnel: Developers First
The S-1/A describes Claude Code as the way into the enterprise. “The adoption of Claude Code began with individual software developers who noticed the potential of autonomous coding agents before enterprise procurement teams had considered the category,” it says, and calls prosumers “a powerful top-of-funnel engine”.
Our reporting shows how one large employer now manages the tool. Goldman Sachs limited Claude Code and Microsoft’s GitHub to its engineers after months in which bankers could use them, Bloomberg News reported; chief information officer Marco Argenti put the firm’s engineers at 12,000 in a July interview, Bloomberg reported, according to PYMNTS.
Anthropic’s own hiring points the same way. By our count, Anthropic names Claude Code in 79 of its 681 open job ads, about 11.6%, and 20 of the 26 posted in the 30 days to 4 November sit in Sales or Applied AI, the teams that sell the product and help customers deploy it.
The Expansion: Spend Per User
The clearest Claude Code number in the filing is about spending. In the three months to 30 September 2026, “Claude Enterprise users who did not use Claude Code spent less than $75 per month on average, while the top 5% of Claude Enterprise users who used Claude Code spent approximately 20 times that amount.” By our arithmetic, that puts the top group somewhere under about $1,500 a month.
The filing gives that example alongside its Net Revenue Retention of approximately 900% as of 30 September 2026, and says: “We see this expansion as customers adopt agentic products like Claude Code.” Customers with $100,000 or more of revenue over the trailing twelve months rose from more than 950 to more than 11,900 in the year to 30 September 2026, and over 50% of them used Claude Platform and at least one first-party application in the latest quarter, up from approximately 30% in the three months to 31 December 2025, it says.
The Engine Room: Anthropic’s Own Code
Claude Code also sits inside Anthropic’s cost base. The S-1/A says over 90% of the contributions to its codebase in September 2026 were written by Claude, and that average daily code volume per engineer in the September quarter was over 45 times the 2024 average.
From 31 December 2024 to 30 September 2026, Anthropic’s headcount grew over five times, while revenue for the latest quarter was 86 times the revenue of the three months to 31 December 2024, the filing says. The filing says it believes that operating leverage “has been driven by the use of Claude throughout our organization.” The filing also says the Agent SDK makes the same harness that powers Claude Code available to all developers.
The Market and the People
Outside the filing, the main market estimate combines Claude Code with a rival. Menlo Ventures estimates that Claude Code and OpenAI’s Codex together account for $24.0 billion, or 70%, of a $34.0 billion enterprise AI coding market, up from about $1 billion a year earlier. The figures are survey-based estimates for the two products combined, and Menlo Ventures is an investor in Anthropic.
The product’s people have also become a story. By our count, three of the four named exits at Anthropic we could source from 5 September to 3 November 2026 worked on Claude Code, including Wu. No source we read calls the moves a reorganization, and none of the three Claude Code posts gives a reason for leaving.
The Business Lever
The lever here is paid subscribers. For readers who follow Anthropic’s businessmodel, Claude Code is where the filing’s growth story becomes concrete: developer adoption, per-user expansion and internal engineering leverage all point to the same product, while the revenue it brings in stays inside the company-wide totals.
Connecting the Coverage
Our reporting on Claude Code and Anthropic’s filing that this analysis draws on:
As we read the S-1/A, Claude Code is the product the filing uses to explain its growthmodel: developers adopt it first, their companies expand, and usage is billed by the token.
As we read the spending example, the gap between users who did not use Claude Code and its heaviest users, about 20 times, is the filing’s clearest product-level number, though it compares the top 5% with an average.
As we read our own reporting next to the filing, the hiring, the departures and the market estimates all centre on one product whose revenue Anthropic does not report separately.
Anthropic S-1/A, 2 November 2026
“We see this expansion as customers adopt agentic products like Claude Code.”
Three Implications
FOR IPO READERS The S-1/A reports revenue for Anthropic as a whole; the Claude Code evidence is a spending example, adoption comparisons and internal engineering figures.
FOR ENTERPRISE BUYERS The filing says revenue is recognized on the quantity of tokens consumed, so the work an agent is allowed to do shows up in the bill.
FOR RIVALS IN AI CODING Menlo Ventures estimates Claude Code and Codex together at 70% of a $34.0 billion enterprise AI coding market; Menlo is an investor in Anthropic.
The framework puts it this way: “Anthropic is betting on many harnesses, embedded where work already lives.”
As we read it, Claude Code is the first of those harnesses, and the S-1/A shows the pattern spreading: the Agent SDK makes the same harness available to all developers, and Claude Cowork extends it to knowledge workers. The framework was published in July 2026 and does not discuss the S-1/A.
What Is Not Established
The S-1/A gives no revenue, run-rate, user count or price for Claude Code, so we do not know what share of the $16.58 billion rise came from it. The 20x is Anthropic’s own metric, compares the top 5% of Claude Code users with the average non-user, and has not been independently verified.
Menlo’s figures are estimates that combine Claude Code with Codex. The job-ad count is a text match on open postings, not hires. All shares, multiples and the 74% are our arithmetic from the filing. We did not contact Anthropic.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growth story. This week it started being priced as a credit story.
Anthropic’s S-1/A ties Claude Code to how customers arrive, how much the heaviest users spend, about 20 times the average non-user, and how its own engineers write code, but it reports revenue only for the company as a whole. Its quarterly revenue rose by $16.58 billion from Claude Code’s research-preview quarter to the three months to 30 September 2026, about 74% of it in the latest two quarters, by our arithmetic.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. This analysis is built from Anthropic’s S-1/A filed with the SEC on 2 November 2026 and from FourWeekMBA’s own earlier reporting on Claude Code, each linked in the text. All shares, multiples and the revenue bridge are our arithmetic from the filing. We did not contact Anthropic.
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.
