Anthropic says revenue will top $10.9 billion, first profit looms

Anthropic told investors it expected second-quarter revenue of about $10.9 billion and its first-ever operating profit, a striking signal that one of the most closely watched AI companies was moving from breakneck expansion toward actual business durability. The expected operating profit was about $559 million, a rare result for a frontier AI firm still spending heavily on computing power and model development.

The scale of the jump was unusual even by Silicon Valley standards. Anthropic’s first-quarter revenue was $4.8 billion, which meant sales would have more than doubled in just a few months if the company hit its latest target. At Anthropic’s developer conference in San Francisco, chief executive Dario Amodei said the company had seen 80-fold growth on an annualized basis in the first quarter, underscoring how quickly demand for Claude and Claude Code had accelerated.

Those numbers matter because they point to a possible turning point in the economics of generative AI. The industry has been defined by explosive usage growth, but also by staggering infrastructure bills, expensive model training cycles and intense competition for enterprise customers. Anthropic’s results suggested that revenue can scale fast enough, at least for now, to outrun some of those costs.

That does not mean the path ahead is cheap. Anthropic said Claude remained available across all three major cloud platforms, AWS Bedrock, Google Cloud Vertex AI and Microsoft Azure Foundry, reinforcing the company’s pitch to large businesses that want flexibility and broad distribution. It has also been tied to rising enterprise demand, especially for Claude and Claude Code, as companies look for tools that can be plugged into existing cloud workflows.

The company’s ambitions were rising alongside its costs. Recent reporting said Anthropic was in a funding round that could value it above OpenAI, a reminder that investors still see frontier AI as a winner-take-most market. But the company’s own infrastructure commitments showed how expensive that race remained. Anthropic agreed to pay SpaceX $1.25 billion monthly for computing power, a figure that highlighted the sheer scale of the compute required to build and serve advanced AI systems.

For Anthropic, profitability would not only mark a company milestone. It would be evidence that the leading AI labs can, at least in some cases, build a sustainable business before infrastructure spending and model-training costs consume the gains.