Anthropic taps Elon Musk’s SpaceX for more AI compute power
Andrew Harnik/Getty Images
- Anthropic and SpaceX have signed a new deal for 300 megawatts of compute capacity.
- The OpenAI rival will use Colossus One data center to boost Claude Code for developers.
- SpaceX, after acquiring xAI, is inking new deals to sell GPU access.
Anthropic is turning to Elon Musk for computing capacity.
As the company faces a surge in demand for its AI coding service, Claude Code, Anthropic will use SpaceX’s Colossus One data center to meet the moment. Chief Product Officer Ami Vora announced the blockbuster deal at Anthropic’s developer conference in San Francisco on Wednesday morning.
Vora said that Anthropic is doubling the current 5-hour rate limit for developers on Pro, Max, Team, and seat-based Enterprise plans. The company is also raising rate limits for developers using the advanced Claude Opus model’s API, allowing more input and output per minute. These moves will stretch the company’s already strained computing capacity.
Given this challenge, Anthropic has turned to SpaceX, which acquired Musk’s xAI lab back in February. SpaceX’s Memphis-based Colossus One data center is one of the world’s largest, housing tens of thousands of advanced Nvidia chips that run AI’s compute-intensive work. Musk’s rapid buildout of the site and the nearby Colossus 2 data center have provoked ire locally, with residents complaining of pollution from gas turbines.
Anthropic expects the deal to deliver more than 300 megawatts of computing capacity within the month, across more than 220,000 Nvidia GPUs. This builds on the company’s strategy of sourcing chip time and capacity from hyperscaling giants — work with Google, Amazon, and Microsoft is also underway.
The deal is another in SpaceX’s fledgling compute sales business. Business Insider reported in April that the coding startup Cursor would train its latest model on xAI’s GPUs. By inking computing deals with other companies, SpaceX and xAI can generate revenue from their own infrastructure, while also still developing AI models.
Musk, as recently as February, excoriated Anthropic’s Claude model as “misanthropic and evil.” But the billionaire’s mind has changed, he wrote Wednesday on X. Musk said he spent time last week with Anthropic leaders, 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.”
Same here.
By way of background for those who care, 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…
— Elon Musk (@elonmusk) May 6, 2026
There could be another deal between Anthropic and SpaceX in the works. Tucked low in the press release accompanying this announcement, Anthropic wrote that it has “expressed interest” in working with SpaceX to build data centers in space.
Have a tip? Contact this reporter via email at scouncil@insider.com or Signal, Telegram, or WhatsApp at 415-757-8198. Use a personal email address, a nonwork WiFi network, and a nonwork device; here’s our guide to sharing information securely.