Elon Musk has publicly praised Anthropic and pledged never to abruptly "cut off" the AI lab from SpaceX's servers, after users on X suggested he could disable a rival that relies on his infrastructure. Musk went so far as to admit he was "clearly wrong about Anthropic," reversing his September 2025 claim that "winning was never in the set of possible outcomes" for the company. The remarks matter because Anthropic is now one of SpaceX's largest customers, raising questions about whether a competitor should trust Musk to host its models.

The comments follow a May 2026 deal in which Anthropic agreed to buy 300 megawatts of compute — the entire output of xAI's Colossus 1 data centre near Memphis — paying $1.25 billion a month through May 2029, worth around $40 billion to SpaceX's xAI unit (xAI merged with SpaceX in February). Google signed a comparable arrangement at $920 million a month through June 2029. Musk cited Tesla's patent pledge, its opened Supercharger network and SpaceX's satellite practices as evidence of his non-predatory "style," though the article notes he has previously sued OpenAI and that hosting Anthropic's compute could give SpaceX unusual visibility into a rival's operations, including the risk of "distilling."