Anthropic taps SpaceX Colossus 1 to boost Claude capacity

Anthropic began routing capacity from SpaceX’s Colossus 1 into its systems this month, accessing more than 300 megawatts connected to over 220,000 Nvidia GPUs to increase capacity for Claude.

The Colossus 1 supply has been directed into Claude Pro and Claude Max. Claude Code’s five‑hour session limits have been doubled for Pro, Max, Team and seat‑based Enterprise plans. Pro and Max accounts will no longer face the prior peak‑hour cap. API customers will see higher rate limits for Claude Opus models.

Anthropic prioritizes Claude Code for developers who run long coding sessions for debugging, refactoring, file review, agent-style tasks and API‑heavy automation. The company expects the added Colossus capacity to ease constraints on long, uninterrupted sessions and increase available inference capacity during busy periods.

The SpaceX agreement is part of a wider infrastructure expansion that includes cloud and chip partners. Anthropic has an agreement with Amazon for up to 5 gigawatts of capacity, with nearly 1 gigawatt expected online by the end of 2026. A separate 5‑gigawatt arrangement with Google and Broadcom begins delivering capacity in 2027. Anthropic has a strategic partnership tied to Microsoft and Nvidia that includes $30 billion of Azure capacity, and it has outlined a $50 billion U.S. AI infrastructure plan with Fluidstack.

Anthropic trains and serves models on a mix of hardware, including AWS Trainium chips, Google TPUs and Nvidia GPUs. The SpaceX deal includes a potential future option to develop multiple gigawatts of orbital AI compute capacity, though the current Colossus 1 allocation is ground‑based and intended to address immediate demand for higher‑capacity tiers.

The company is expanding capacity outside the United States to meet compliance and data‑residency requirements for enterprise clients in financial services, healthcare and government. The Amazon agreement includes extra inference capacity in Asia and Europe. Anthropic said it is targeting new sites in democracies with legal and regulatory frameworks suited to very large AI deployments and secure supply chains for chips and networking equipment.

A Bank of America strategist warned that possible public listings for SpaceX and Anthropic could add a large supply of shares to the market and exert downward pressure on prices. The strategist pointed to recent index‑rule changes and faster index entry as factors that could force passive funds to free capital for new issues.

Anthropic expects the SpaceX capacity to benefit its heaviest users immediately and to form one element of a multi‑partner infrastructure strategy to scale Claude for consumer and enterprise customers.