Two questions have followed the AI labs over the last several years as they grew: Are businesses really using advanced AI models, and can they ever make any money selling them? Anthropic's S-1 statement gets a little closer to the answer.
Anthropic outlined its strengths Monday with the release of the long-awaited S-1, including rapid revenue growth, compute efficiency, and operating efficiency, but it also laid out the numerous risks associated with its IPO, reportedly set for mid-November.
Enterprise demand for its AI models has been a key part of Anthropic's growth: Its customer base as of September 30, 2026 included more than 11,900 customers that generate $100,000 or more in revenue, up from 950 as of the same period last year. In addition, it has more than 1,900 customers that produce $1 million or more of revenue on a twelve-month trailing (TTM) basis, and 250 customers that spend $10 million or more on a TTM basis.
“The speed and depth of our enterprise customer adoption are powerful evidence of Claude’s position at the frontier of AI,” it said in the filing, adding that as of September 30, 2026, it had net revenue retention of 900%.
In other words, as Michael Sikand, owner of Sikand Media, posted on X, “This means customers who were paying Anthropic a year ago now spend about 9x as much. That's 6x higher than any public software company.”
In contrast with more established companies, Palantir boasts a 157% revenue retention, while Snowflake follows with 126% — both of which are quite good by most enterprise tech standards.
“This customer momentum is driven by significant usage expansion as customers realize the value of deploying Claude across their organizations,” according to the filing.
