Anthropic plugs into SpaceX's 220,000-GPU Colossus — and doubles Claude's rate limits

Short summary

Anthropic secured exclusive access to SpaceX's Colossus data center (220K GPUs), increasing Claude API rate limits 10–16× across all tiers and doubling Claude Code limits within a month. The deal removes peak-hour throttling for Pro/Max users and eliminates a major constraint on agent development. Infrastructure partnerships are becoming a key competitive moat as Anthropic, OpenAI, and others race to scale compute.

• •Claude API rate limits increase 10–16× across all tiers (e.g., Opus Tier 1: 30K→500K input tokens/min)
• •Claude Code Pro/Max/Team/Enterprise limits double within a month; peak-hour throttling removed
• •Infrastructure scale is now a competitive differentiator; Anthropic has partnerships with Amazon (5GW), Google (5GW), Microsoft ($30B), Fluidstack ($50B), and SpaceX (Colossus)

Generated with AI, which can make mistakes.