How Anthropic Quietly Passed OpenAI on Revenue

Anthropic went from an underdog to a serious threat in about three years. It closed a 20x revenue gap down to roughly 2x, and it did that while growing far faster than OpenAI.

That pace is the part people miss. Consumer growth is loud and visible. Enterprise growth is quiet, and it compounds through renewals and expanding contracts.

My read is that the market watched ChatGPT's user count and assumed that was the whole story. Revenue was telling a different one the entire time.

OpenAI built for the world. Anthropic built for the finance team. That single decision explains most of the revenue mix.

A free ChatGPT account is a real expense. Every query costs compute, and most of those users never pay. You need a huge base to convert a thin slice into subscribers.

Anthropic skipped the stadium. Roughly 85% of its revenue comes from enterprise and developer customers, while OpenAI leans heavily on consumer subscriptions. Fewer users, far more revenue per customer, and much lower churn once a contract is signed.

For years, Anthropic's caution read like a handicap. Slower to ship the flashy consumer features, more restrained in what its models would say.

Then the buyer changed. When the person signing the check is in legal or compliance, they do not care which model writes the wittiest email. They care whether it will get the company sued.

That is where safety turned into a sales pitch. The trait critics called overly cautious became the reason a bank or an insurer felt comfortable rolling it out at scale. Positioning matters, and Anthropic positioned the exact thing enterprises were nervous about.

Enterprise revenue is stickier by design. Once a company builds workflows around a model through the API, switching means re-testing, re-approving, and retraining teams. That friction protects the incumbent.

There is nuance here worth saying out loud. OpenAI still leads in raw scale and brand reach, and consumer dominance has its own advantages in mindshare and data. Anthropic has not won the war.

But on the metric that pays the bills, one focused bet beat a broader one. Loud and profitable are not the same thing, and the AI race is starting to prove it.