Key Takeaways:
Elon Musk reversed his stance on Anthropic, calling it the undisputed AI leader — a shift that coincides with a $40bn compute deal between the two rivals.
Elon Musk said he was "clearly wrong" about Anthropic, calling its Mythos and Fable models unmatched by any rival — a rare admission that also burnishes a $40bn business relationship between SpaceX and the AI startup.
"Anthropic is obviously currently the leader in AI," Musk wrote on X. "No other lab has shipped a model as good as Mythos or Fable."
The endorsement marks a sharp reversal from February, when Musk called Anthropic's models "misanthropic and evil" after it raised $30bn at a $380bn valuation. Five months later, the same company is his benchmark for the field. The shift comes as Anthropic pays Musk's xAI roughly $1.25bn a month through 2029 to lease the entire output of the Colossus 1 data center near Memphis — a deal worth close to $40bn.
The praise serves dual purposes. It is a genuine assessment of model quality, but it also tells the market that the AI company that matters is not OpenAI — the startup Musk co-founded, left in 2018, and is now suing. Both Anthropic and OpenAI filed confidentially for stock market listings in June, within days of each other. Anthropic, valued at about $965bn in private markets, is pushing for a Nasdaq debut as early as October.
The commercial ties between the two companies run deep. In May, Anthropic agreed to lease the full capacity of SpaceX's Colossus 1 facility — more than 220,000 Nvidia GPUs with 300 megawatts of power. The deal runs through May 2029 with either party able to exit on 90 days' notice. Musk responded to speculation that SpaceXAI could terminate the lease by pledging never to cut off Anthropic's access "in a way that hurt them badly," citing his companies' history of supporting competitors through open patents and Supercharger access.
The arrangement supports inference for Claude models and includes discussions on future orbital data centers. SpaceXAI countered Anthropic's June launch of Claude Fable 5 and Mythos 5 with Grok 4.5 on July 8, a model positioned as faster and cheaper.
Anthropic's ascent carries tangible upside for its two largest strategic investors. Amazon and Alphabet have each made substantial commitments to the startup, which trains and runs its models across both AWS and Google Cloud Platform using their custom silicon — Amazon's Trainium and Inferentia chips and Google's TPUs.
Greater adoption of Anthropic's models drives incremental compute demand, benefiting the cloud providers supplying the hardware. AWS revenue is growing 28% year over year with operating margins at 38%, while GCP revenue is accelerating 63% annually with margins above 30%. As Anthropic scales to next-generation models such as Mythos 2, the compute requirements will almost certainly increase, creating layered demand for cloud capacity.
For investors, the key question is whether the market has priced in Anthropic's trajectory. Amazon trades at a forward P/E that has compressed despite accelerating cloud revenue, suggesting the full upside from its Anthropic relationship is not yet reflected. Alphabet shows a similar pattern. With Anthropic targeting its first profitable quarter — around $559m in operating income on $10.9bn of revenue — and an IPO potentially weeks away, the financial stakes of Musk's endorsement extend far beyond model benchmarks.
This article is for informational purposes only and does not constitute investment advice.
