Anthropic Secures SpaceX Colossus 1 Deal, Doubles Claude Code Rate Limits

Anthropic announced on May 6 a major infrastructure deal with SpaceX, securing dedicated training and inference capacity on the Colossus 1 supercomputing cluster, with the immediate effect of doubling rate limits for all Claude Code users globally.

Key points:

• Colossus 1 provides Anthropic with 200,000 H100-equivalent GPUs of dedicated capacity, removing the compute ceiling that had constrained Claude Code's per-user rate limits.

• Claude Code rate limits were doubled same-day: Pro users now receive 4,000 agentic turns per day, up from 2,000.

• The partnership also includes a long-term training reservation for Claude 5, with Colossus providing the primary training cluster.

Compute access is the silent governor of AI capability delivery. Anthropic's ability to double rate limits overnight demonstrates that their previous constraints were infrastructure-driven, not product-driven. SpaceX entering the AI compute market as a provider to Anthropic creates an interesting dynamic: Musk's infrastructure supporting Anthropic while the Musk v. Altman legal battle plays out simultaneously.

If you have been hitting Claude Code rate limits during intensive agentic tasks, the doubling of limits removes a significant daily workflow constraint. Update your automation scripts to take full advantage of the new headroom. The Colossus partnership signals that Claude 5 training is actively underway. Watch for capability announcements in Q3-Q4 2026.

Why It Matters: The overnight doubling of rate limits demonstrates that compute access—not product decisions—was the binding constraint, while the Claude 5 training reservation signals Anthropic's next major capability push is underway.