Elon Musk praises Mythos/Fable, promises not to ‘cut off’ Anthropic
Key Takeaways
- Elon Musk publicly reversed his earlier skepticism about Anthropic, calling it the current leader in AI and praising its Mythos/Fable model, according to TechCrunch.
- Anthropic agreed in May to buy the entire 300-megawatt output of xAI’s Colossus 1 data center near Memphis, Tennessee, paying $1.25 billion per month through May 2029 — a deal worth about $40 billion in revenue for SpaceX’s xAI unit, TechCrunch reported.
- Google separately signed a deal to rent SpaceX infrastructure through June 2029 for $920 million per month, per TechCrunch, underscoring how central SpaceX has become to rival AI labs’ compute needs.
From Rival Sniping to Reluctant Praise
The backdrop to this story is a notable about-face. TechCrunch reports that Musk wrote on Thursday, “I was clearly wrong about Anthropic,” a reference to his September 2025 post on X in which he claimed that winning was never a realistic outcome for the company. That dismissal looked shaky even at the time, since Anthropic was already reported to hold the largest AI market share among enterprise customers, according to TechCrunch.
The renewed praise came after users on X raised a pointed question: could Musk, given his history of friction with AI competitors, simply cut Anthropic off from SpaceX’s infrastructure if he chose to compete more aggressively? Musk rejected the idea outright, telling users that shutting out a rival “is not my style,” as quoted by TechCrunch. He went further, calling Anthropic the obvious current leader in AI and stating that no company has released a model as good as Mythos/Fable, adding that Anthropic will “undoubtedly” have a follow-up, Mythos 2, ready soon.
To back up his claim of not squeezing competitors, Musk pointed to Tesla’s 2014 patent pledge — a commitment, originally detailed in a company blog post and now housed under Tesla’s formal patent pledge, not to sue anyone using its technology in good faith. He also cited Tesla’s decision to open its Supercharger network and charging-port design to other automakers, and noted that SpaceX launches competing satellite systems without raising prices or imposing unfair terms on rivals, according to TechCrunch.
Why the Anthropic-SpaceX Deal Matters Beyond the Rhetoric
The reassurances matter because Anthropic’s dependence on SpaceX is not trivial. TechCrunch confirms that as of July 2026, Anthropic ranks among SpaceX’s largest customers, built on the May agreement to buy the full 300-megawatt capacity of the Colossus 1 data center. That commitment, running through May 2029 at $1.25 billion per month, adds up to roughly $40 billion in revenue for SpaceX’s xAI division — figures that make clear why both sides have strong incentives to keep the arrangement stable, regardless of Musk’s personal statements.
Google’s parallel deal, paying $920 million per month for SpaceX infrastructure through June 2029 as reported by TechCrunch, reinforces that this is now a broader pattern: major AI labs are leaning on SpaceX’s compute capacity at scale, even when those labs compete directly with Musk’s own AI ambitions through xAI. For readers tracking the AI economy, this signals that infrastructure providers and AI developers are becoming intertwined in ways that blur traditional competitive lines — a dynamic that can ripple into how investors price risk across AI-linked equities and, by extension, into crypto markets that often move in sympathy with AI infrastructure sentiment.
TechCrunch also notes that Musk is not without a history of contentious tactics toward rivals — pointing to his lawsuit against OpenAI as one example. That context matters for assessing how much weight to put on personal assurances versus contractual and financial safeguards. As TechCrunch observes, Anthropic doesn’t need to rely solely on Musk’s stated principles: there would be real contractual consequences if SpaceX abruptly disrupted Anthropic’s infrastructure, and SpaceX has substantial financial motivation to preserve the relationship. There may also be a secondary benefit for SpaceX, as TechCrunch points out, since hosting Anthropic’s workloads could let SpaceX’s engineers gain experience supporting a fast-growing AI company’s infrastructure, similar to how Amazon’s engineers support AI clients.
The report also raises a more sensitive issue: proximity and visibility. TechCrunch recalls that during his trial against OpenAI, Musk acknowledged that AI “distilling” — where one company sets up numerous fake accounts to query a rival’s model and reverse-engineer its behavior — is a real practice, telling a lawyer, as reported by the New York Times, that “generally AI companies distill other AI companies.” Anthropic itself accused three Chinese model makers of this kind of behavior in February, according to TechCrunch. While Anthropic and Google likely assume they have protections in place while using SpaceX’s infrastructure, the arrangement inherently gives SpaceX more insight into how Anthropic operates than most competitors would ordinarily have.
Hype Check
Claim: Musk says he would never use SpaceX’s infrastructure to disadvantage Anthropic, framing the $40 billion, multi-year compute deal as evidence of goodwill toward a competitor. Reality: TechCrunch reports the deal carries binding contractual terms through May 2029, and SpaceX has clear financial incentive — the $1.25 billion monthly payments, plus a comparable $920 million monthly deal with Google — to keep the arrangement intact regardless of Musk’s personal stance; the source also notes Musk’s history of aggressive rival tactics, such as suing OpenAI, and flags that hosting Anthropic’s compute gives SpaceX unusual visibility into its operations. Verdict: Mixed. This is not financial advice.
Source
Researched with AI assistance, fact-checked and edited by a human. Not financial advice.