Elon Musk has dramatically reversed his stance on Anthropic, publicly declaring the company the "obvious current leader in AI" and committing to protect its access to SpaceX computing resources despite direct competition between their AI divisions. The reversal marks a significant shift from Musk's earlier dismissals of the company as "doomed" and accusations that its Claude models were biased and "misanthropic and evil."

What Changed Musk's Mind About Anthropic?

On Thursday, Musk responded to speculation about SpaceX potentially terminating Anthropic's lease on the Colossus 1 data center facility in Memphis. In his response on X, Musk acknowledged his previous criticism was unfounded, stating "I was clearly wrong about Anthropic." He specifically praised Anthropic's recent Mythos and Fable models as "unmatched" in their capabilities.

On Thursday, Musk

This public endorsement comes just days after SpaceXAI, Musk's AI division, launched Grok 4.5 on July 8 as a direct competitor to Anthropic's offerings. Despite this rivalry, Musk emphasized that principled competition should take precedence over retaliation. He pledged never to cut off Anthropic's access to SpaceX compute resources "in a way that hurt them badly," citing his companies' historical commitment to supporting competitors through open patents, Supercharger access, and fair satellite launches.

Musk

How Large Is Anthropic's Computing Deal With SpaceX?

The partnership between Anthropic and SpaceX represents one of the largest compute infrastructure agreements in the AI industry. In May 2026, Anthropic secured a lease for the full capacity of SpaceX's Colossus 1 facility, a massive data center containing over 220,000 Nvidia graphics processing units (GPUs) and 300 megawatts of power capacity. The financial commitment is substantial, with Anthropic paying approximately $1.25 billion monthly through May 2029, bringing the total potential value to over $40 billion.

The arrangement includes flexibility for both parties, with either Anthropic or SpaceX able to exit the agreement on 90 days' notice. Beyond supporting inference for Claude models, the deal also includes discussions about future orbital data centers, suggesting long-term collaboration on next-generation computing infrastructure.

How to Understand the Competitive Landscape Between SpaceX and Anthropic

• Recent Model Launches: Anthropic released Claude Fable 5 and a restricted version of Mythos 5 in June 2026, setting new performance benchmarks in coding and agentic tasks, which are AI systems designed to take autonomous actions toward specific goals.
• SpaceX's Counter-Move: SpaceXAI responded on July 8 with Grok 4.5, positioning it as faster and significantly cheaper than Anthropic's offerings, creating direct market competition.
• Organizational Structure: SpaceX merged with Musk's AI startup xAI, making the two companies direct rivals in the large language model (LLM) market, where companies build AI systems trained on vast amounts of text data.

Musk's previous criticism of Anthropic had been sharp and personal. In February 2026, after the company raised $30 billion at a $380 billion valuation, Musk accused Claude models of being biased against specific demographic groups and repeatedly mocked Anthropic's name by predicting it was "doomed" to become the most misanthropic AI firm. His public reversal suggests either a reassessment of Anthropic's technical capabilities or a strategic decision to prioritize business relationships over public feuding.

Anthropic is expected to go public later in 2026, which could further elevate its market position. SpaceX itself debuted on the Nasdaq in June 2026, giving both companies public market profiles. At the time of reporting, retail trader sentiment on platforms like Stocktwits remained bearish toward both SPCX (SpaceX) stock and Anthropic, with SPCX having fallen 5 percent since its Nasdaq debut.

Musk's commitment to maintain Anthropic's compute access despite competitive pressure signals that even in the intensely competitive AI market, infrastructure partnerships may transcend direct model rivalries. The decision to publicly pledge fair treatment suggests Musk views long-term industry stability and his companies' reputations as more valuable than short-term competitive advantages gained through sabotage.