Anthropic landed $517 billion in computing power contracts over 11 months, though its scale may still fall short of OpenAI’s.

From Beating AI Express: A report by The Information reveals Anthropic has secured at least 14.8GW of new compute capacity over the past 11 months, with deployment scheduled over the coming years. Based on current public contract estimates, the potential total cost could reach a maximum of $517 billion, with most expenditures set to occur over the next decade. Adding the 1–2GW of capacity it had locked in prior to last October, Anthropic’s total contracted compute capacity stands at roughly 16GW. This expansion is primarily driven by demand for its Claude AI models. This year, growth in Claude Code and Claude Cowork exceeded Anthropic’s expectations, prompting the company to aggressively secure compute resources. Under new agreements, Amazon will provide up to 5GW, while Google and Broadcom will supply another 5GW; Microsoft and NVIDIA will contribute up to around 1GW. Anthropic also leased all compute capacity of SpaceX’s Colossus 1, gaining access to more than 220,000 NVIDIA GPUs at once, including H100, H200, and GB200 models. Even with this massive investment, Anthropic’s compute plans may still lag behind OpenAI’s. Anthropic has locked in approximately 16GW, with many contracts extending beyond 2030. OpenAI has set a target of reaching 30GW by 2030, and projects cumulative compute investments of roughly $750 billion by that year. The $517 billion figure is The Information’s estimated maximum potential cost based on existing cloud, chip, and data center contracts. Some compute capacity may be delayed or unused, and certain contracts allow for early cancellation.