Elon Musk’s SpaceX has been secured as a pivotal partner by Anthropic, enhancing its AI infrastructure just before the company goes public. On Wednesday, it was announced that Anthropic had agreed to utilize all computational resources at SpaceX’s Colossus 1 data center, thereby gaining access to over 220,000 NVIDIA GPUs within the month.
This deal is set to enhance services for Claude Pro and Claude Max subscribers by significantly increasing API rate limits for Claude Opus and doubling five-hour rate limits for Claude Code on Pro, Max, Team, and Enterprise plans, all effective immediately.
The agreement with SpaceX builds on Anthropic’s existing compute partnerships, including a 5 gigawatt deal with Amazon that promises nearly 1 gigawatt of new capacity by year-end; another 5 gigawatt partnership with Google and Broadcom slated for 2027; a $30 billion Microsoft-NVIDIA collaboration covering Azure capacity; and a $50 billion investment in U.S. AI infrastructure with Fluidstack.
Anthropic has also expressed interest in extending its partnership with SpaceX to include orbital AI compute capabilities, going beyond terrestrial data centers. This strategic timing is crucial as SpaceX prepares for an IPO, having filed confidentially with the SEC on April 1. The company aims for a valuation between $1.75 trillion and $2 trillion, with the public S-1 expected by late May and roadshow activities planned for the week of June 8.
By adding Anthropic as a named compute client before its listing, SpaceX strengthens its business proposition beyond launch services and Starlink, now highlighting AI infrastructure as a new revenue stream. This development coincides with Anthropic’s efforts to expand internationally in compliance with data residency requirements for regulated industries.