Anthropic just closed a $65 billion Series H round at a $965 billion post-money valuation. Run-rate revenue crossed $47 billion earlier this month. For builders already on Claude, this changes the infrastructure calculus in a meaningful way.
The round was led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, with co-leads including Capital Group, Coatue, D1 Capital Partners, GIC, ICONIQ, and XN. The $65B figure includes $15 billion of previously committed investments from hyperscalers, among them $5 billion from Amazon.
Three compute deals stand out. Anthropic signed agreements with Amazon for up to five gigawatts of new capacity. It also signed with Google and Broadcom for five gigawatts of next-generation TPU capacity, and with SpaceX for GPU capacity in Colossus 1 and Colossus 2. That is a significant expansion in a short window, and the company says this compute growth happened in recent weeks.
Strategic hardware partners are now part of the picture too. Micron, Samsung, and SK hynix joined as infrastructure partners. Their role covers memory, storage, and logic chips. The framing here is supply-chain reliability, not just raw capacity. If you have hit rate limits or seen latency variance under load, this is the stated fix.
For multi-cloud teams, the availability story is now complete. Claude is the first frontier model available on all three of the largest cloud platforms: AWS, Google Cloud, and Microsoft Azure. AWS remains the primary cloud and training partner. That means you can deploy Claude wherever your data and compliance requirements already point you, without a separate negotiation.
On the product side, CFO Krishna Rao called out Claude Code and Cowork specifically as tools the company is working to make more helpful, more powerful, and more adaptable. Those are the two products getting the most direct investment attention from this round, alongside safety and interpretability research.
The investor list spans sovereign wealth funds (GIC, Temasek, MGX), large asset managers (Fidelity, T. Rowe Price, Baillie Gifford), and infrastructure-adjacent firms (Blackstone, Brookfield). That breadth signals this is not a bet on a single use case. It is a bet on Claude becoming general-purpose enterprise infrastructure.
What to do now: If your team is on a single-cloud deployment and you have been deferring a multi-cloud Claude evaluation, the AWS plus Google Cloud plus Azure availability removes the last logistical excuse. Run a latency and availability benchmark across the three platforms against your actual workload. The compute expansion also means this is a reasonable time to re-test throughput limits that may have constrained you earlier this year.