(Credit: Anthropic)
Elon Musk’s ambitions for orbiting data centers might have a major partner in Anthropic, which said today that it's interested in working with SpaceX on the technology.
Anthropic brought up the topic while announcing a deal to use the Memphis Colossus 1 supercomputer from xAI, which was absorbed by SpaceX in February. “We’ve signed an agreement with SpaceX to use all of the compute capacity at their Colossus 1 data center. This gives us access to more than 300 megawatts of new capacity (over 220,000 Nvidia GPUs) within the month,” Anthropic said.
For now, Anthropic’s interest in orbiting data centers is vague. In a blog post, the company merely noted: “As part of this agreement, we have also expressed interest in partnering with SpaceX to develop multiple gigawatts of orbital AI compute capacity.”
The interest is notable, considering OpenAI CEO Sam Altman (and Musk nemesis) has been dismissive of orbiting data centers, an unproven technology. However, SpaceX is betting that the concept will win out in the long term, pointing to the benefits of harnessing solar power directly in Earth's orbit. Nvidia is also working on a chip for orbiting data centers, but experts say the industry will face major technical and environmental hurdles.
(Credit: xAI)In the meantime, the bigger, more immediate development is the Colossus 1 deal. The move is surprising considering Colossus was originally built to realize xAI's own ambitions for cutting-edge generative AI development, including taking on Microsoft. In February, Musk also slammed Anthropic, alleging that the company's Claude AI was racist. "Frankly, I don’t think there is anything you can do to escape the inevitable irony of Anthropic ending up being Misanthropic," he tweeted at the time.
However, The Information reports that xAI has struggled to use the 500,000 Nvidia GPUs it acquired, resulting in GPU utilization of only 11%. “To be fair to xAI, everyone struggles with GPU utilization, and a researcher at a rival firm said cracking 40% was difficult for most of xAI’s competitors. But a rate of 11% is appallingly low,” The Information notes.
It’s also no secret that xAI has faced some turmoil; all of its original co-founders have now left the company, except for Musk. In March, the SpaceX CEO argued that, "xAI was not built right first time around, so [it's] being rebuilt from the foundations up. Same thing happened with Tesla."
The partnership with Anthropic suggests that xAI can effectively lease GPUs, even though it risks supporting a competitor. And xAI also has additional GPUs it can rely on for its own business through the Colossus 2 supercomputer, also based in Memphis.
On why he made the deal, Musk tweeted: "I spent a lot of time last week with senior members of the Anthropic team to understand what they do to ensure Claude is good for humanity and was impressed. Everyone I met was highly competent and cared a great deal about doing the right thing. No one set off my evil detector. So long as they engage in critical self-examination, Claude will probably be good."
The deal also occurs when SpaceX is preparing to IPO in an effort to raise tens of billions to develop orbiting data centers. Musk added: "We reserve the right to reclaim the compute if their AI engages in actions that harm humanity."
In another tweet, the SpaceX CEO said, "xAI will be dissolved as a separate company, so it will just be SpaceXAI, the AI products from SpaceX." A press release also uses the term SpaceXAI.
For Anthropic, the deal adds some badly needed compute capacity. “This additional capacity will directly improve capacity for Claude Pro and Claude Max subscribers,” it says. As part of the deal, Anthropic is also raising usage limits for Claude Code, a popular coding agent for developers.