Anthropic S-1 Reveals Ecosystem Strategy: Cross-Product Stickiness Drives ~900% Net Revenue Retention

By IT之家Original title: Anthropic 招股书披露生态打法:打造高粘性企业客户体系
78AI Score

Anthropic's S-1 filing reveals a strategy to boost enterprise stickiness by integrating models, platforms, and applications. Over half of customers spending $100k+ annually now use first-party apps like Claude Code, driving net revenue retention to nearly 900%. This approach mirrors Amazon's AWS expansion but faces risks from cloud partner competition and high customer concentration.

Full Report

October 27 — Anthropic is building a cross-product ecosystem, the kind that drives strong user stickiness for tech giants like Amazon and Uber. According to its S-1 filing, this strategy is already showing results. Currently, more than half of customers spending at least $100,000 annually use both the Claude platform and first-party applications such as Claude Code and Cowork; by the end of last year, this proportion was only around 30%.

Anthropic's IPO Filing Reveals Ecosystem Strategy: Building High-Stickiness Enterprise Customer Base

The S-1 document filed with the U.S. Securities and Exchange Commission on Monday (local time) states: "The speed and depth of enterprise adoption strongly demonstrate that Claude is at the forefront of the artificial intelligence field."

As of September 30, Anthropic had over 11,900 customers with annual spend of at least $10 (IT Home note: approximately 66.8 RMB at current exchange rates); over 1,900 customers with annual spend exceeding $100 (approximately 667.7 RMB); and roughly 250 customers with annual spend of $10 million or more (approximately 66.774 million RMB).

Anthropic attributes the sustained growth in customer spending to "broad deployment of Claude within enterprises, recognized value, and significantly increased usage," which drove the company to achieve a net revenue retention rate of approximately 900% by the end of the third quarter.

Anthropic subtly challenges the traditional SaaS industry by positioning Claude not as a "static product serving a single scenario," but as a "cutting-edge AI system with rapidly evolving capabilities," covering the model layer, developer platform, and various applications used directly by users.

The filing states: "Technical progress at any one layer creates compounding effects across the three-layer architecture, accelerating our ability to deliver stronger intelligent capabilities." The company claims that this integrated approach is driving an increasing number of customers to adopt both the Claude platform and first-party applications simultaneously.

This playbook is familiar. Uber initially offered only ride-hailing services before expanding into food delivery and freight logistics; Amazon leveraged AWS cloud services to attract customers, then extended relationships into storage, databases, and other products, gradually building an ecosystem that makes it difficult for customers to leave. Like these companies, Anthropic does not need to find entirely new buyers for each product: an enterprise can start by calling APIs, then deploy Claude Code for engineers, and subsequently roll out Cowork to all employees.

Users of these applications hold significantly higher value. The prospectus reveals that in the third quarter, the top 5% of Claude enterprise customers using Claude Code spent approximately twenty times more per month than average enterprise customers who did not use the coding tool. This data indicates substantial room for growth in per-customer spending and explains why Anthropic is eager to introduce first-party applications to accounts that started with API services.

Anthropic states that its Applied AI teams have further amplified this natural penetration effect. Internal Applied AI teams engage with startups and large enterprises, assisting customers in embedding Claude into their existing technology stacks. Recently, the company also launched a new initiative called Ode with Anthropic in partnership with Blackstone and other private equity firms, deploying dedicated engineers on-site to build customized AI tools based on Claude for clients.

Anthropic also relies on third-party distribution partnerships with hyperscale cloud providers AWS, Google Cloud, and Microsoft Azure. The filing describes this model as a "co-marketing arrangement": Anthropic signs contracts directly with customers and pays service fees to the cloud providers. The company claims this approach allows it to "avoid channel conflicts" and maintain "control over pricing."

As this AI giant continues to launch "new category products aimed at generating economic value far exceeding existing businesses," this nascent stickiness strategy may become crucial. The company states that, unlike typical enterprises, it will not devote excessive effort to reinforcing existing products and business models, but will instead allocate more resources to defining new businesses.

Despite impressive revenue growth—reaching $16.3 billion in the first half of 2026 (approximately 108.842 billion RMB), far exceeding $1.2 billion (approximately 8.013 billion RMB) in the same period last year—the prospectus lists several headwinds. Cloud partners that help sell Claude are increasingly viewing Anthropic as a competitor, which Anthropic believes could make it difficult to secure necessary compute capacity from these cloud providers. The filing notes that Amazon, Google, and Microsoft may prioritize their own needs or favor competitors they have invested in.

Customer concentration is another risk. The risk factors section states: "Adverse changes in demand from key customers... could negatively impact our business and prospects. Therefore, if one or more core customers significantly reduce spending or terminate cooperation, without being offset by new customers or increased orders from other existing customers, it would adversely affect our business, financial condition, and operating results."