Anthropic has overtaken OpenAI in self-reported annual revenue, reaching a $47 billion run rate by May 2026, compared to OpenAI’s projected $25 billion to $33 billion. The gap widens when you look at valuation: Anthropic recently raised $65 billion in Series H funding at a $965 billion valuation, now making it the world’s most valuable AI startup, surpassing OpenAI’s current $852 billion valuation.

The shift reflects not just growth, but a fundamental difference in how the two companies make money. Anthropic earns roughly 85% of its revenue from enterprise and developer customers who use Claude through APIs. OpenAI earns roughly 85% from ChatGPT consumer subscriptions. One company is selling to businesses. The other is selling to people.

How Anthropic Got Here

Anthropic started in 2021 as a research lab spun out of OpenAI by former VP of Research Dario Amodei and others. The company raised modest funding initially, then picked up speed. By April 2026, Anthropic had crossed the $30 billion annual revenue threshold. By May, it hit $47 billion.

The company reported profitability by the fourth quarter of 2025 and projects it will become sustainably profitable by 2029. OpenAI hasn’t reached profitability yet and hasn’t given a timeline for when it will.

Anthropic’s customer base includes enterprise clients using Claude for customer service, code generation, content analysis, and other high-volume tasks. These contracts are large, recurring, and defensible—the bread and butter of enterprise software.

What the Revenue Race Says About AI

Consumer adoption of AI has plateaued faster than expected. ChatGPT grew fast initially, but user growth slowed dramatically after mid-2023. The app still has millions of subscribers, but the gold rush for consumer AI is over. Enterprise, by contrast, is hungry. Companies are racing to embed AI into internal workflows, and they’re willing to pay.

Anthropic bet on enterprise earlier and more deliberately. The company focused on Claude’s reliability, safety, and performance on practical business problems. That focus is paying off in contract value.

OpenAI, meanwhile, is trying to diversify. The company built a consumer app that generated hype but limited revenue per user. It’s now pushing GPT integrations into enterprise, but Anthropic already has momentum there.

The Stakes Ahead

Both companies will release new models this year. OpenAI’s upcoming GPT-5.5 is expected soon, with hints at a more advanced version called GPT-6. Anthropic hasn’t announced its next-generation model publicly, though it’s clearly in development.

The revenue gap matters because it funds research. Anthropic’s $47 billion run rate translates to more money for compute, talent, and R&D. OpenAI still has access to capital from Microsoft and other investors, but the revenue comparison signals that enterprises trust Anthropic’s models more for their core workloads.

The narrative of AI competition is shifting from who built the smartest model to who built a model businesses can’t afford to replace.