Anthropic S-1 Shows How Claude Locks In Enterprise Customers
October 26, 2026
Anthropic's S-1 shows that more than half of customers spending at least $100,000 a year now use both the Claude platform and a first-party app such as Claude Code or Cowork, up from about 30% at the end of last year. The headline revenue figure, $16.3 billion in the first half
Anthropic's S-1 shows that more than half of customers spending at least $100,000 a year now use both the Claude platform and a first-party app such as Claude Code or Cowork, up from about 30% at the end of last year. The headline revenue figure, $16.3 billion in the first half of 2026 against $1.2 billion a year earlier, is less revealing than that adoption curve: the filing is a blueprint for turning an API contract into a multi-product dependency. For enterprise buyers, it explains how a pilot becomes a platform decision, and why the time to set negotiating terms is before the second product lands.
What's new
Anthropic's filing with the SEC describes Claude as three layers: the models, the developer platform (the API, or programming interface, that developers use to call the models), and the applications users interact with directly. Anthropic calls this the "integrated approach," and the filing says "an advance at any layer compounds across all three layers, accelerating the pace at which we can deliver more capable intelligence." The commercial logic follows from the architecture. A customer can start with the API, then deploy Claude Code to engineers, then Cowork to the broader workforce.
Anthropic PBC, the filer, is not leaving that progression to chance. An internal Applied AI team (staff who place Claude inside customers' existing technology stacks) works with startups and enterprises on deployment. The company also recently launched "Ode with Anthropic" with Blackstone and other private equity firms, a venture that embeds forward deployed engineers (engineers who work inside the customer's business rather than from a vendor office) to build bespoke AI tools with Claude.
Distribution runs through the hyperscalers (the largest cloud providers) as well. Under what the filing calls a "joint go-to-market approach" with AWS, Google Cloud and Microsoft Azure, Anthropic holds the direct contract with the customer and pays each cloud provider a fee. The filing says the structure lets Anthropic "avoid channel conflict" (a seller competing with its own distributors) and "set pricing."
Why it matters
The stickiness data is the core of the case. As of September 30, Anthropic reports more than 11,900 customers spending at least $100,000 a year, more than 1,900 at $1 million or more, and about 250 at $10 million or more. Net revenue retention, the measure of how much existing customers spend this period relative to last, stood at about 900% at the end of the third quarter. By that figure, the average cohort is spending roughly nine times what it did before.
- $100,000+ annually: more than 11,900 customers
- $1 million+ annually: more than 1,900 customers
- $10 million+ annually: about 250 customers
Revenue tells the same story at company scale: $16.3 billion in the first half of 2026 against $1.2 billion in the same period a year earlier, a jump of more than thirteenfold. Growth that steep is rarely driven by new logos alone, which is why the expansion inside existing accounts matters more than the headline.
The app layer appears to drive that expansion. In the third quarter, the highest-spending 5% of Claude Enterprise customers who used Claude Code spent about 20 times as much per month as the average enterprise user who did not use it. That is a comparison of a selected top slice against an average, so it shows the ceiling of expansion rather than the typical outcome. The filing itself attributes momentum to "significant usage expansion as customers realize the value of deploying Claude across their organizations."
Competitive Landscape
The filing positions Claude as "frontier AI with rapidly advancing capabilities," not a "static product that serves a single use case." The contrast implies a target: single-purpose software that cannot absorb new model capability. Bundling is the mechanism. A company running the API, Claude Code and Cowork under one direct contract, with cloud-marketplace billing layered on top, faces higher switching costs at each added layer.
Scale supports the pitch. Coverage of the filing cites $430 billion in locked-in AI computing and $65 billion raised at a $965 billion post-money valuation (the company's value once the new funding is counted). Those figures describe capacity and capital, not product quality, but they signal Anthropic intends to stay at the frontier.
- OpenAI: the largest commercial rival, widely described as Anthropic's biggest challenger, with the two competing fiercely for business customers, talent and influence in Washington.
- Amazon (AWS), Google (Google Cloud), Microsoft (Azure): partner-competitors, not product alternatives. The filing warns they are increasingly seeing Anthropic as a competitor, which could make compute harder to secure if they prioritize their own needs or the rivals they have invested in.
What's next
Anthropic is expected to go public in November, raising "more than $100 billion," according to press reports.
Buyers should watch three signals. First, whether cross-product adoption keeps climbing past half of $100,000-plus customers. Second, whether net revenue retention of about 900% holds as the base grows. Third, whether the expansion pattern described in the customer-expansion language continues once the largest accounts finish their initial rollouts.
Practical advice: price the second and third products into the first contract. Ask what the direct-versus-cloud-marketplace choice does to your pricing leverage, and keep an exit path for the model layer even if you adopt the apps.
The prospectus is written for investors, but its subtext is a sales strategy aimed at your budget. The same filing that celebrates customers who spend more each quarter also admits that losing a few of the largest would hurt. Lock-in is a two-way street, and buyers who remember that will negotiate better.
For a CIO whose engineers start on the API with a small monthly bill, the filing's data suggests the real budget line is the app layer. If Claude Code spend behaved like the top-5% cohort, at about 20 times the average enterprise user who does not use it, a pilot could grow into a far larger line item, though that is the high end of the range, not the typical outcome. Procurement should negotiate seat-tier price protection, a cap on year-over-year renewal increases, and data-portability terms before Cowork is rolled out to the wider workforce, not after.
-- Aria Lin, Enterprise Technology Analyst
Sources: Claude / Anthropic · TechCrunch · Microsoft Foundry