Anthropic’s IPO filing reveals its strategy for keeping enterprise customers hooked

Buried in Anthropic's IPO filing is a clear picture of how it grows inside companies. More than half of customers spending at least $100,000 a year now use both the Claude platform and a first-party app like Claude Code or Cowork, up from about 30% at the end of last year. The company reports net revenue retention of about 900% as of the end of the third quarter.

That is an extraordinary expansion number, and it explains how revenue reached $16.3 billion in the first half of 2026, up from $1.2 billion a year earlier. It also tells enterprise buyers something important: once Claude is in, it tends to spread.

We'll give Anthropic credit first. Customers do not expand spending like this on something that does not work. The filing says the top 5% of Claude Enterprise customers who used Claude Code spent about 20 times as much per month in Q3 as the average enterprise user who did not use the coding tool. That is a signal that people are finding real work for these tools, which is the version of AI we want to see win.

The strategy is also plainly about stickiness. TechCrunch compares it to AWS and Uber, and the comparison fits. A company starts with the API, rolls Claude Code to its engineers, then Cowork to everyone else. Anthropic adds Applied AI teams that help wire Claude into existing systems, and Ode with Anthropic, a venture with Blackstone and other private equity firms that embeds engineers to build custom tools. Each step deepens the integration.

That is where the caution belongs. When one vendor supplies the model, the developer platform and the apps on top, switching gets expensive. The filing says Anthropic will spend less effort than a typical business entrenching current products and more on new ones, which is a nice line, but lock-in does not require intent. It comes from integration depth. Buyers should be planning their exit options now, while they still have leverage.

The risks Anthropic names cut both ways. The same cloud giants that resell Claude, AWS, Google Cloud and Azure, increasingly see it as a competitor, and the filing warns they may prioritize their own needs or rivals they have invested in when handing out compute. It also flags concentration: losing or shrinking one or more of its largest customers could hurt results. For a company whose customers are betting their workflows on it, those are vendor risks too, not just investor risks.

Our read: this is a strong, real business built on a classic land-and-expand model. If it scales, a handful of AI vendors could end up as deeply embedded in corporate operations as the cloud providers are today, with the same pricing power. That may be fine. It just should not happen by accident.

Anthropic filed its S-1 with the SEC on Monday. As of September 30, it reported more than 11,900 customers spending at least $100,000 a year, more than 1,900 spending over $1 million, and about 250 spending $10 million or more. Under its joint go-to-market approach, Anthropic holds the direct contract with customers who come through AWS, Google Cloud and Azure, and pays each cloud provider a fee, which it says lets it avoid channel conflict and set pricing.