Anthropic has completed a $65 billion Series H financing round at a post-money valuation of $965 billion, moving ahead of OpenAI as the world’s highest-valued AI startup. The company, founded less than three years ago, said its annualized revenue has already exceeded $47 billion. The deal ranks among the largest private financings on record.
Investor list spans major funds and sovereign capital
The round was led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital. Co-leads included Capital Group, Coatue, D1 Capital Partners, Singapore’s GIC, ICONIQ, and XN. Blackstone, Fidelity, and Temasek were also listed among participants.
Of the $65 billion raised, $15 billion came from commitments previously made by hyperscale cloud providers. According to the report, those commitments were converted from compute procurement agreements into equity, with Amazon accounting for $5 billion. Sequoia partner Alfred Lin said Anthropic is “redefining the boundaries of what AI applications can do.” Altimeter founder Brad Gerstner described the financing as “the most important infrastructure investment of this generation.”
Compute expansion moves to multi-gigawatt scale
Anthropic CFO Krishna Rao said the new capital will be directed toward three areas: safety and interpretability research, compute expansion, and broader product and partnership development. On infrastructure, the company signed an agreement with Amazon for up to 5 GW of new capacity. It also secured 5 GW of next-generation TPU capacity through Google and Broadcom, and obtained GPU capacity from SpaceX’s Colossus 1 and Colossus 2.
The report compares 5 GW of sustained compute power to the output of a mid-sized nuclear power plant used entirely for AI training and inference. By comparison, most leading AI training clusters in 2023 were below 100 MW. That gap shows how aggressively Anthropic is scaling for next-generation model demand.
Claude is now available across the three largest clouds
On the cloud side, Claude has become the first frontier model available on AWS, Google Cloud, and Microsoft Azure at the same time. That gives enterprise customers access to the model without locking themselves into a single cloud provider, a point that may matter as AI deployment decisions become more closely tied to infrastructure choice.