Claude Usage Limits Raised, Compute Capacity Deal with SpaceX

Anthropic has partnered with SpaceX to significantly expand compute capacity, and along with it has raised usage limits for Claude Code and the Claude API.

Three changes take effect starting today.

• Double the 5-hour usage limits for Claude Code across Pro, Max, Team, seat-based Enterprise.
• Remove peak-hour limit reductions for Pro and Max accounts.
• Significantly raise API request limits for Claude Opus models.

Through the deal with SpaceX, Anthropic will use all of the compute capacity at the Colossus 1 data center, securing over 300MW of new capacity and more than 220,000 NVIDIA GPUs within a month. This expansion directly broadens available capacity for Claude Pro and Claude Max subscribers, and also leaves open the possibility of collaborating with SpaceX on developing multi-GW-scale orbital AI compute capacity.

The company also reiterated other commitments: a partnership with Amazon for up to 5GW, a 5GW deal with Google-Broadcom, a strategic partnership with Microsoft-NVIDIA, and a $50 billion investment plan with Fluidstack. Claude currently uses AWS Trainium, Google TPUs, and NVIDIA GPUs together, and the company plans to add further inference capacity in Asia and Europe, particularly for customers in regulated industries such as finance, healthcare, and government.

Anthropic is also considering extending overseas its pledge to cover consumer electricity price increases caused by domestic U.S. data centers. Countries where infrastructure is deployed will be limited to those with democratic institutions and secure supply chains, alongside parallel community reinvestment.