Anthropic reached a $47 billion revenue run rate by May 2026, more than doubling from $9 billion a year earlier. The AI startup’s growth has caught investor attention—one venture capitalist said it is growth they have never seen in 25 years of backing startups.

The San Francisco company makes Claude, a large language model that competes with OpenAI’s GPT and Google’s Gemini. Claude became the second most-used AI assistant globally this spring, behind OpenAI’s ChatGPT. That shift in market share came fast.

How Anthropic Got Here

Anthropic was founded in 2021 by former OpenAI executives, including Dario and Daniela Amodei. The company took time building. For two years, it barely existed outside of research papers. Then Claude launched publicly in March 2023. Growth was slow at first, then accelerated.

The recent jump reflects two shifts. First, enterprise customers now pay for AI. Early skeptics thought Claude was a toy. Now banks, law firms, and manufacturers test it on real work. Second, Anthropic released Claude 3 in March 2024, then Claude 4 and Claude Sonnet in 2025. Each release got cheaper and faster. Cheaper always wins.

The Funding Bets

Google invested $5 billion last year. Amazon committed $4 billion. Both bets are partially hedged—neither wants one AI vendor owning the future. But they are also real money. Anthropic uses it to rent GPU time from cloud providers and hire the researchers who know how to make models bigger without breaking them.

What It Means

Fifty billion a year would rank Anthropic ahead of most software companies. The company is not selling that much today—Wall Street projects it will. The gap between what it makes now and what that run rate assumes matters. If enterprise adoption stalls, if competitors release better models, the number shrinks fast.

Anthropic is also still private. No IPO yet. That runway extends for years if burn rate holds.

Anthropic’s fastest growth comes from customers who stopped treating Claude like a demo and started building products around it.