Elon Musk once dismissed Anthropic as a lost cause in the race for artificial intelligence supremacy. No longer. On July 9, the chief executive of SpaceX and xAI took to X with words that mark a sharp turn. “I was clearly wrong about Anthropic,” he posted. “They are obviously currently the leader in AI. No company has released a model as good as Mythos/Fable and they will undoubtedly have Mythos 2 ready soon.”
The post came in response to concerns that Musk might abruptly deny Anthropic access to the massive compute resources it rents from his companies. He pushed back hard. “And I would never cut them off in a way that hurt them badly, even as a competitor. That’s not my style.” Short. Direct. And loaded with implications for an industry where infrastructure often matters more than models.
Trust cuts both ways here. Anthropic pays SpaceX roughly $1.25 billion a month for 300 megawatts of compute from the Colossus 1 data center near Memphis. The arrangement, signed in May, runs through 2029 and could generate $40 billion for Musk’s operation. TechCrunch first reported the details of this unusual partnership. Google signed a similar but smaller deal for the same infrastructure. These arrangements turned xAI, after its February merger with SpaceX, into something of a specialized cloud provider for frontier AI labs.
But history between Musk and Anthropic runs deep and often bitter. In September 2025 Musk wrote that “Winning was never in the set of possible outcomes for Anthropic.” He has criticized the company’s safety focus, its ties to Google, and what he saw as hypocritical positioning. Dario Amodei’s firm emerged from OpenAI in 2021 partly over concerns about Musk’s influence. The two sides have traded barbs for years.
So why the change? Timing matters. Anthropic released its Mythos class models in June. Fable serves as the commercially available version wrapped in heavy guardrails. Mythos itself carries advanced capabilities in areas like cybersecurity that triggered regulatory scrutiny. The U.S. government issued an export control directive in mid-June that forced Anthropic to suspend access to both models for all foreign nationals, including its own employees. The company disabled them entirely to comply. Firstpost covered Musk’s reversal and the context of his prior criticism.
Independent benchmarks showed Mythos and Fable setting records. One coding evaluation delivered 80.3 percent on SWE-Bench Pro, far ahead of competitors. Andrej Karpathy called the jump a “major-version step change.” Developers at Stripe reportedly ported 50 million lines of code in a single day. No other lab had shown that level of practical performance. Musk’s admission aligns with observable reality. His xAI unit launched Grok 4.5 around the same period, yet he conceded the lead.
The compute deal adds another layer. Anthropic ranks among SpaceX’s largest customers as of July 2026. Terminating service would trigger contractual penalties and damage the revenue stream. It would also signal to every potential cloud customer that Musk’s companies cannot be relied upon. He pointed to past behavior as proof of character. Tesla opened its patents in 2014 and later shared Supercharger designs. SpaceX competes in satellite systems without predatory pricing. “Even my worst enemies can attack me on this platform,” he added.
Yet skepticism lingers. Musk sued OpenAI and lost. During that litigation he acknowledged the practice of model distillation, in which one lab queries another’s system at scale to extract capabilities. Anthropic itself accused three Chinese developers of distillation attacks against Claude earlier this year. Hosting a rival’s inference workloads gives xAI engineers unusual visibility into performance patterns, latency characteristics, and optimization tricks. Business Insider examined the shift from Musk’s earlier disparagement to current cooperation.
Proximity creates opportunity. Amazon’s engineers gained deep knowledge by supporting Anthropic and OpenAI workloads on its chips. SpaceX staff could do the same. The arrangement benefits both sides today. Anthropic secures reliable, high-performance infrastructure during a period of explosive demand. Musk’s companies collect substantial payments and gather operational intelligence that could inform future systems.
Still, the relationship carries risk. Contracts expire. Competitive pressures mount. Government intervention, as seen with the June export controls on Mythos and Fable, can reshape access overnight. Some observers on X speculated the regulatory action itself might have been influenced by rivals. Others viewed Musk’s praise as strategic positioning ahead of larger battles over standards, regulation, and talent.
Anthropic has grown fast. The company reached a $965 billion valuation after raising $65 billion. Enterprise adoption surveys from 2025 already showed it leading in market share for business use cases. Its models emphasize constitutional principles and safety testing that appeal to regulated industries. Mythos pushed those capabilities further, even if the public version required heavy filtering.
Musk’s post did more than praise a competitor. It signaled acceptance of a multipolar AI field where no single player dominates every dimension. xAI focuses on truth-seeking and maximum curiosity. Anthropic stresses safety and alignment. OpenAI pursues general intelligence with commercial scale. Each has strengths. Each rents infrastructure from the others when needed.
The statement also reassures developers. Many worried that reliance on a Musk-controlled cloud could end suddenly if xAI released a superior model. His explicit promise reduces that fear, at least for now. “That’s not my style” carries weight when backed by billions in monthly revenue and a public track record of opening certain technologies.
But style can evolve. Musk has reversed positions before. The AI sector moves at breakneck speed. Mythos 2 could arrive before year-end. Grok iterations continue. New regulatory actions may target frontier capabilities regardless of ownership. Contracts get renegotiated. And personal rivalries never fully disappear in an industry this concentrated.
For now the message is clear. Musk sees Anthropic at the front of the pack. He intends to keep supplying the power that keeps its models running. The rest of the industry will watch closely to see whether that commitment holds when the next breakthrough shifts the balance again. In AI, today’s leader can become tomorrow’s challenger. Partnerships that look odd on paper sometimes prove the most durable.
Recent coverage reinforces the shift. Crypto Briefing detailed the GPU lease and valuation figures surrounding the praise. The episode highlights how compute scarcity and model performance now intertwine in ways few predicted even a year ago. Musk’s words may calm nerves in the short term. They also underscore the tangled dependencies that define the current state of frontier AI development.
WebProNews is an iEntry Publication