Anthropic has announced a significant increase in usage limits for its Claude Code AI platform, attributing the expansion to a new agreement with SpaceX that grants access to extensive compute power at SpaceX's Memphis data center.
- SpaceX partnership unlocks over 300 megawatts of compute power for Anthropic.
- Claude Code usage limits doubled for Pro and Max subscribers with peak-hour caps removed.
- Anthropic pursuing orbital data centers to meet future AI compute demands.
What happened
At the Code with Claude conference, Anthropic revealed a major new deal with SpaceX that provides access to the full compute capacity of the Colossus 1 supercomputer at SpaceX's Memphis facility. This infrastructure includes over 220,000 NVIDIA GPUs such as H100 and H200 models, dramatically boosting Anthropic's processing capabilities.
Following this arrangement, Anthropic doubled usage limits on its Claude Code platform for Pro and Max subscribers, eliminated previous peak-hour usage reductions, and raised API limits for its Opus AI model. These changes respond to growing user demand and have been accompanied by similar recent agreements with Microsoft, Amazon, Google, and Nvidia.
Why it matters
Demand for Anthropic’s AI products has surged amid professional developer migration from competitors and growing interest in multi-agent AI applications that require significant compute resources. Prior limitations on usage during peak times had sparked frustration within user communities on platforms like Reddit and Hacker News.
By expanding compute access and raising user limits, Anthropic is positioning itself to capture greater market share and support more complex and intensive workflows. The SpaceX partnership also signals a forward-looking effort to explore innovative compute solutions, including deploying significant AI processing capacity in orbit to overcome terrestrial power and cooling constraints.
What to watch next
Observers should monitor how quickly increased compute availability translates into improved user experience and scalability of Claude Code services, especially as usage surges among software development professionals. Additional outcomes from Anthropic’s other compute partnerships may materialize over coming months, further shaping service capacity.
Anthropic’s expressed interest in building large-scale orbital data centers with SpaceX introduces a potentially transformative approach to AI infrastructure. Future developments in this area could redefine limitations associated with compute power availability, offering new opportunities for next-generation AI training and operation.