Elon Musk says he will not cut off Anthropic from SpaceX servers

In short: Elon Musk publicly said he would not shut Anthropic off from SpaceX’s computing systems, even though the companies compete in AI.

What happened

Elon Musk responded on X after people suggested he could remove Anthropic from SpaceX-run servers without warning. He said doing that was “not my style,” and he posted praise for Anthropic’s AI models, called Mythos and Fable.

This matters because Anthropic is now a major customer of SpaceX’s AI and computing unit. Anthropic agreed in May to rent a large amount of “compute,” which is basically the raw computer power needed to run and train AI (like renting a huge fleet of engines to do heavy work). The deal is reported to be worth about $1.25 billion per month through May 2029, or roughly $40 billion total.

TechCrunch also noted that Google has signed a separate deal to rent SpaceX infrastructure through June 2029 for $920 million per month. These contracts became possible after Musk’s xAI merged with SpaceX earlier in 2026.

Why it matters

For everyday users, this is a reminder that popular AI tools depend on a small number of companies that own the computers behind the scenes. If a provider and a customer fall out, service could be disrupted, although contracts usually include penalties if one side breaks the agreement. It also shows how competitors can still rely on each other when the cost of computing power is extremely high.

Source: TechCrunch AI