Anthropic Partners with SpaceX for AI Compute Capacity, Boosts Claude API Rate Limits
SpaceX's Colossus 1 data center powers Anthropic with 300 megawatts of AI compute
Anthropic has announced a strategic partnership with Elon Musk's SpaceX to tap into over 300 megawatts of processing power from SpaceX's Colossus 1 data center. This collaboration gives Anthropic a massive compute boost, essential for training and running its advanced Claude models. The deal marks one of the largest dedicated AI compute allocations from a non-cloud provider, signaling Anthropic's aggressive push to scale infrastructure independently of major cloud platforms like AWS or Azure.
Alongside the compute expansion, Anthropic is significantly increasing rate limits for its Claude models. Claude Code's rate limits are being doubled, while Claude Opus API users will see maximum input tokens per minute rise from 2 million to 10 million — a 5x jump. This means developers can now process longer documents, run more concurrent requests, and handle heavier workloads without throttling. The combined effect positions Anthropic to compete more directly with OpenAI and Google DeepMind in high-throughput enterprise AI applications.
- Anthropic gains access to 300+ megawatts from SpaceX's Colossus 1 data center for AI compute
- Claude API input token limit jumps from 2M to 10M per minute, a 5x increase
- Claude Code rate limits are being doubled to support more frequent developer interactions
Why It Matters
Massive compute boost and higher API limits enable enterprises to run larger AI workloads at scale, challenging cloud hyperscalers.