Anthropic has raised usage limits for paid Claude users as it scales up its computing infrastructure through a partnership with SpaceX. The company said this expanded capacity will benefit subscribers on Claude Pro and Claude Max plans, while free users will not receive any changes. As part of the deal, Anthropic stated it will gain access to the full computing power of SpaceX’s Colossus 1 data center, adding over 300 megawatts of capacity and more than 220,000 NVIDIA GPUs within a month.

The company introduced three key updates to usage limits for paid customers. It has doubled Claude Code’s five-hour rate limits across Pro, Max, Team, and Enterprise seat-based plans, removed peak-hour restrictions for Pro and Max users, and significantly increased API rate limits for Claude Opus models.

Under the new API structure, Tier-I users now receive up to 500,000 input tokens per minute (previously 30,000) and 80,000 output tokens per minute (up from 8,000). Tier-II limits have risen to 2 million input and 200,000 output tokens per minute, while Tier-III users get up to 5 million input and 400,000 output tokens. Tier-IV users now have limits of 10 million input tokens and 800,000 output tokens per minute. Rate limits control how much data users can process in a given time, while tokens represent chunks of text used by AI systems to interpret inputs and generate responses. Higher limits allow more complex and larger workloads.

Anthropic also outlined broader infrastructure expansion plans, including large-scale compute partnerships with Amazon, Google, Broadcom, Microsoft, and NVIDIA. These efforts include multi-gigawatt capacity expansions and billions of dollars in investments across global AI infrastructure. The company said future growth will also focus on international deployment, especially in regulated sectors like healthcare, finance, and government, with an emphasis on compliance and data residency requirements.