Anthropic has signed an agreement with SpaceX for access to the full computing power of the Colossus 1 data center in Memphis. This gives the AI company access to more than 220,000 Nvidia GPUs and 300 megawatts of new capacity. The additional computing power will immediately increase usage limits for Claude Code and the Claude API.
The additional capacity made available by SpaceX is expected within a month. Colossus 1 was originally built by xAI, Elon Musk’s AI startup and a competitor of Anthropic.
The expansion of computing capacity directly translates to higher usage limits. Starting today, Claude Code’s five-hour rate limits are being doubled for Pro, Max, Team, and Enterprise subscriptions. At the same time, the peak-time restriction for Pro and Max accounts is being lifted. The increased input and output token limits are listed in the table below.
It became clear earlier this year that users were encountering limitations when developers raised concerns about Claude’s usage limits. The API rate limits for Claude Opus models are also being significantly increased.
Part of a broader infrastructure strategy
Anthropic has been significantly scaling up its computing capacity recently. For instance, Amazon announced a $25 billion investment in Anthropic, centered on an agreement for up to 5 gigawatts of computing capacity. This was followed by a 5 GW deal with Google and Broadcom, a partnership with Microsoft and Nvidia worth $30 billion in Azure capacity, and a $50 billion investment in U.S. AI infrastructure via Fluidstack.
In its latest announcement, Anthropic also expressed interest in partnering with SpaceX to develop several gigawatts of AI computing capacity in space. It is not yet clear when that will become a reality.
Finally, Anthropic is working to expand its infrastructure internationally. Enterprise customers in sectors such as financial services, healthcare, and the public sector are increasingly requesting regional infrastructure due to compliance and data residency requirements.