Anthropic signed a SpaceX compute deal for 220,000+ NVIDIA GPUs at Colossus 1, and immediately doubled Claude Code rate limits across paid plans.
Key Takeaways
SpaceX’s Colossus 1 adds 300+ MW (220,000+ NVIDIA GPUs) to Anthropic’s capacity within the month.
Claude Code five-hour rate limits doubled for Pro, Max, Team, and seat-based Enterprise; peak-hours throttling removed for Pro and Max.
API rate limits raised considerably for Claude Opus models; specific figures in an updated table on the announcement page.
Total compute pipeline now spans Amazon (up to 5 GW), Google/Broadcom (5 GW), Microsoft/NVIDIA ($30B Azure), and Fluidstack ($50B).
Anthropic flagged interest in orbital AI compute capacity with SpaceX, and is expanding internationally with in-region infrastructure for regulated industries.
Hacker News Comment Review
Commenters flagged that doubling the five-hour limit without raising the weekly cap means users hit the same ceiling faster, not later, reducing the practical benefit for heavy users.
The SpaceX angle drew skepticism about whether the orbital compute mention was genuine strategy or a contractual PR concession to close the deal.
Several commenters noted SpaceX is monetizing Colossus 1 capacity originally built for Grok, framing it as a smart pre-IPO revenue move regardless of xAI’s direction.
Notable Comments
@gpugreg: Questions whether the orbital compute line was serious strategy or inserted to get the deal done.
@mirzap: “It simply means that you can reach the weekly limits in three days instead of five.”
