Elon Musk praises Mythos/Fable, vows not to cut off Anthropic

Can Anthropic rely on Elon Musk to host its models? After users on X suggested that Musk might arbitrarily remove the AI lab from SpaceX’s servers to cripple a competitor, Musk responded with effusive praise, stating that such a move is “not my style.”

“I was clearly wrong about Anthropic,” Musk wrote on Thursday, referencing his September 2025 X post where he claimed, “Winning was never in the set of possible outcomes for Anthropic.” Of course, even then, Anthropic was already a winner, reportedly holding the largest enterprise AI market share.

Those anti-Anthropic days appear to be behind Musk, and not just on X. By July 2026, Anthropic had become one of SpaceX’s largest customers.

To recap: In May, Anthropic signed a deal to purchase 300 megawatts of compute — the full output of xAI’s Colossus 1 data center near Memphis, Tennessee. (Musk’s xAI merged with SpaceX in February.) Anthropic agreed to pay $1.25 billion per month through May 2029, generating roughly $40 billion in revenue for SpaceX’s xAI unit. For context, Google also signed a deal to lease SpaceX infrastructure through June 2029 at $920 million per month.

Musk insists that Anthropic’s decision wasn’t dangerous and that he holds great admiration for his rival.

“They are clearly the current leader in AI. No company has released a model as strong as Mythos/Fable, and they will undoubtedly unveil Mythos 2 soon. I would never cut them off in a way that severely harms them, even as a competitor. That’s not my style,” he wrote.

As evidence of his non-aggressive approach, he cited Tesla’s 2014 decision (detailed in a now-deleted blog post, preserved under its patent pledge) to refrain from filing patent lawsuits against anyone using its technology in good faith. He also noted that Tesla opened its Supercharger network and charging port design to competitors.

“SpaceX launches competing satellite systems without raising prices or imposing unfair terms. Even my worst enemies can criticize me on this platform,” he wrote, offering another example.

Of course, Musk is not above using tactics against rivals, particularly those with whom he has a history. For instance, he sued OpenAI.

Anthropic does not have to rely on Musk adhering to his “style,” however. Abruptly shutting down Anthropic’s infrastructure would trigger contractual penalties, not to mention the substantial benefits SpaceX gains from maintaining the deal. Anthropic pays generously, and SpaceX’s engineers can learn to build and support Anthropic’s rapidly expanding AI, much like Amazon’s engineers do.

That proximity may offer other advantages. During his trial against OpenAI, Musk acknowledged the reality of AI “distilling” — a process where a model maker creates numerous fake accounts to send prompts to a competitor and learn its workings. As the New York Times reported, when asked whether xAI had ever distilled technology from OpenAI, Musk replied: “Generally AI companies distill other AI companies.”

In February, Anthropic accused three Chinese model makers of doing this to Claude. Presumably, Anthropic and Google believe they have safeguards against SpaceX engaging in similar practices while using its infrastructure. However, hosting Anthropic’s compute could still grant SpaceX greater insight into the company’s operations than most competitors would ever obtain.

At present, this partnership appears to offer nothing but upside for Musk’s company. As for tomorrow, once the three-year contract matures, who knows?