Anthropic's compute stack: 5GW from Amazon, 5GW with Google and Broadcom, $30bn of Azure, $50bn via Fluidstack
Anthropic has assembled capacity across four partners rather than one, committing more than $100bn to AWS technologies over ten years for up to 5GW, a further 5GW with Google and Broadcom, $30bn of Microsoft Azure capacity, and $50bn of US infrastructure through Fluidstack.
Every one of Anthropic's major cloud competitors is also a supplier to it, and in several cases an investor. Deliberately depending on all of them is a strategy with a name: refusing to let any single provider hold the leverage that exclusivity would create.
The commitments are individually large and structurally different. More than $100bn to AWS over ten years for up to 5GW. A separate 5GW with Google and Broadcom, which pairs a cloud with a custom-silicon partner. $30bn of Azure capacity. $50bn of American infrastructure via Fluidstack, which is buildout rather than rental. And an agreement to take the entire capacity of SpaceX's Colossus 1 in Memphis — over 220,000 NVIDIA GPUs and more than 300MW.
Against roughly $30bn of run-rate revenue, up from about $9bn at the end of 2025, these are commitments that assume growth continues at a rate no company has sustained for long. That is the risk, and it is not hidden — it is the same bet the entire sector is making, sized to one balance sheet.
It also reframes the Decart discussions. Buying a team that improves chip utilisation is worth a great deal when capacity is contracted years in advance: efficiency gains land as recovered capacity you have already paid for.
Anthropic — Anthropic expands partnership with Google and Broadcom for multiple gigawatts of next-generation compute → · Anthropic — Anthropic and Amazon expand collaboration for up to 5 gigawatts of new compute →