Bottom line: Anthropic doubles Claude limits through a SpaceX deal with 300 megawatts of compute capacity. Additionally, the company secures capacity from Amazon, Google, Microsoft, and NVIDIA for long-term growth.
Anthropic is doubling usage limits for Claude Code and increasing API rate limits for Claude Opus thanks to a new partnership with SpaceX. The company gains access to over 300 megawatts of compute capacity featuring more than 220,000 NVIDIA GPUs at the Colossus-1 data center.
The capacity expansion is part of a comprehensive infrastructure strategy by Anthropic. In addition to SpaceX, there are significant agreements with Amazon (up to 5 gigawatts, with 1 GW by the end of 2026), Google and Broadcom (5 GW from 2027), Microsoft and NVIDIA (30 billion dollars of Azure capacity), and a 50-billion-dollar investment with Fluidstack.
The company is using a diversified hardware portfolio consisting of AWS Trainium, Google TPUs, and NVIDIA GPUs. For enterprise customers in regulated industries such as financial services, healthcare, and government, Anthropic is strategically expanding infrastructure regionally. The Amazon partnership includes additional inference capacity in Asia and Europe. The choice of locations follows the principle: preferably in democratic countries with secure supply chains.
In the United States, Anthropic is absorbing electricity price increases through its own data centers. Similar models are currently being evaluated internationally to enable long-term independent and cost-efficient growth.