This is quite the unexpected alliance: Anthropic has signed a deal with SpaceX to use the entire compute capacity of Colossus 1 — the massive data center in Memphis, Tennessee, originally built for Elon Musk’s xAI.
What’s in the deal?
Colossus 1 brings over 300 megawatts of compute power and more than 220,000 Nvidia GPUs — including H100, H200, and the brand-new GB200 accelerators. This gives Anthropic a massive infrastructure boost that flows directly into Claude products.
What changes for you
The impact is immediate and concrete:
- Claude Code: The five-hour limit for Pro, Max, and Enterprise customers is being doubled
- Peak-hour throttling: Gone entirely for Pro and Max accounts
- API limits: Rate limits for Claude Opus models are increasing significantly
These aren’t vague promises for ‘someday’ — the changes take effect now.
And then there’s space
Almost casually, the deal mentions that Anthropic has ‘expressed interest’ in working with SpaceX on orbital AI compute capacity — multiple gigawatts in space. The logic: the next generation of AI systems will need more power than terrestrial data centers can deliver. SpaceX is the only company with the launch cadence and constellation operations experience to make this a near-term engineering program rather than a research concept.
Sounds like science fiction? Maybe. But if anyone can take the orbit seriously as a compute location, it’s SpaceX.
What it all means
For SpaceX, Anthropic is a marquee customer ahead of its planned IPO. For Anthropic, the deal solves an acute problem: demand for Claude has grown faster than infrastructure. And for us users, it simply means less waiting, more tokens.
The irony of Musk’s infrastructure now powering Claude while he’s simultaneously suing OpenAI is something you just have to appreciate.