Anthropic commits about $40 billion to SpaceX for compute

Anthropic has become one of SpaceX’s largest customers as of July 2026, after signing a deal in May to buy 300 megawatts of compute — the entire output of xAI’s Colossus 1 data center, located near Memphis, Tennessee. Anthropic agreed to pay $1.25 billion per month through May 2029, a deal worth about $40 billion in revenue for SpaceX’s xAI unit, which merged with SpaceX in February.

Musk has reversed his earlier skepticism about the AI lab. Referring to a September 2025 post on X in which he said winning was never in the set of possible outcomes for Anthropic, Musk acknowledged on Thursday that he had been clearly wrong about the company. Even at that time, Anthropic was reported to have the biggest AI market share with enterprises. Praising Anthropic’s Mythos/Fable model — an AI model the company released — Musk said: “They are obviously currently the leader in AI. No company has released a model as good as Mythos/Fable and they will undoubtedly have Mythos 2 ready soon. And I would never cut them off in a way that hurt them badly, even as a competitor. That’s not my style.” As evidence of that approach, he pointed to Tesla’s 2014 decision not to pursue patent lawsuits against good-faith users of its technology, Tesla’s decision to open its Supercharger network to rivals, and SpaceX’s practice of launching competing satellite systems without raising prices or imposing unfair terms.

The arrangement sits alongside SpaceX’s other major infrastructure deals:

• Anthropic: $1.25 billion per month through May 2029, for 300 megawatts of compute from Colossus 1.
• Google: $920 million per month through June 2029, for SpaceX infrastructure.

Even as Musk avoids friction with Google and Anthropic, his record with other rivals differs — he sued OpenAI.

Anthropic doesn’t have to rely solely on Musk’s stated style: the multi-year contract carries its own consequences if SpaceX were to cut service, and SpaceX has clear financial incentive to keep the deal intact. But hosting Anthropic’s compute could still give SpaceX greater visibility into how the company operates than most competitors would ever have. Musk has acknowledged that AI companies generally distill each other’s models — a practice in which one model maker uses fake accounts to prompt a competitor’s model to learn how it works. Anthropic in February accused three Chinese model makers of doing this to Claude.

Why it matters

Anthropic’s about $40 billion compute deal with SpaceX’s xAI unit makes it one of SpaceX’s largest customers, raising questions about competitor dependency and the potential for data visibility despite Musk’s public assurances.