Anthropic announced on May 28, 2026 that it raised $65 billion in Series H funding at a $965 billion post-money valuation. The round was led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, with Capital Group, Coatue, D1 Capital Partners, GIC, ICONIQ, and XN among the co-leads.
The number that makes the round more than a valuation headline is revenue. Anthropic said its run-rate revenue crossed $47 billion earlier in May. It also said the funding will support safety and interpretability research, compute expansion, product scaling, and partnerships. Read together, this is a capacity raise: Anthropic is funding the infrastructure and distribution needed to keep Claude inside enterprise workflows.
Compute is the operating constraint
Anthropic’s post names several compute relationships alongside the funding. The round includes $15 billion of previously committed investments from hyperscalers, including $5 billion from Amazon. Anthropic also said it signed agreements with Amazon for up to five gigawatts of new capacity, with Google and Broadcom for five gigawatts of next-generation TPU capacity, and with SpaceX for access to GPU capacity in Colossus 1 and Colossus 2.
Those details matter because frontier AI companies no longer raise only to hire researchers or subsidize consumer usage. They raise to buy, reserve, and operate the compute that determines how quickly models can be trained and served. A valuation near $1 trillion is easier to understand when the company is also talking in gigawatts.
Anthropic says Claude is the first frontier model available on AWS, Google Cloud, and Microsoft Azure. That is a distribution advantage for enterprise buyers who do not want to move all AI workloads into one cloud. It also means Anthropic has to maintain relationships with the same hyperscalers that supply the infrastructure behind its growth.
Run-rate is useful but easy to overread
The $47 billion figure is run-rate revenue, not a full-year revenue statement. It is still an important signal. Run-rate tells readers the current commercial pace if recent revenue levels continue, which is why companies use it when growth is moving faster than annual reporting can capture.
The caution is simple: run-rate is a forward-looking shorthand. It does not prove what Anthropic will book over the next twelve months, and it says nothing by itself about margins. The real question is whether enterprise demand for Claude can keep scaling while training, inference, memory, and power costs rise with model ambition.
That is why the Series H announcement reads less like a normal private-market milestone and more like a statement about operating scale. Anthropic is telling the market that demand is large, compute is the bottleneck, and the company now has enough capital and infrastructure partnerships to keep pushing both.
