Anthropic turns to SpaceX compute deal as Claude usage limits rise

The agreement gives Anthropic access to more than 300 megawatts of capacity as demand for AI coding and API tools continues to test infrastructure limits.

Anthropic says its SpaceX compute partnership will add more than 300 megawatts of capacity as it increases Claude Code and Claude API usage limits.

Anthropic has agreed a compute partnership with SpaceX the space and satellite company founded by Elon Musk, that will give it access to all capacity at the Colossus 1 data center, adding more than 300 megawatts of infrastructure within the month.

The deal comes as Anthropic raises usage limits for Claude Code and the Claude API, including higher five-hour rate limits for Claude Code users and increased API rate limits for Claude Opus models. The move matters for developers, enterprise users, and education technology teams building with AI tools, where compute availability is increasingly tied to product reliability, speed, and access.

Anthropic announced the update on LinkedIn and in a company post. The company says the SpaceX agreement sits alongside other recent compute deals with Amazon, Google, Broadcom, Microsoft, NVIDIA, and Fluidstack.

Claude Code and API limits increase

Anthropic says three usage changes are effective immediately.

The company is doubling Claude Code’s five-hour rate limits for Pro, Max, Team, and seat-based Enterprise plans. It is also removing the peak hours limit reduction on Claude Code for Pro and Max accounts.

API rate limits for Claude Opus models are also being raised. Anthropic says the changes are designed to improve access for its most active customers, including developers using Claude for coding workflows and organizations building products on top of its API.

For EdTech companies working with generative AI, coding assistants, assessment tools, content generation, tutoring systems, or workflow automation, rate limits are not a minor product detail. They affect how often tools can be used, how reliably services perform under pressure, and whether AI features can scale beyond pilots.

SpaceX capacity joins wider infrastructure push

Anthropic says the SpaceX agreement gives it access to more than 300 megawatts of new capacity, including more than 220,000 NVIDIA GPUs, within the month. The company says this additional capacity will directly improve availability for Claude Pro and Claude Max subscribers.

The deal follows several other infrastructure commitments. Anthropic has announced an agreement with Amazon for up to five gigawatts of capacity, including nearly one gigawatt by the end of 2026. It has also agreed a five-gigawatt deal with Google and Broadcom, expected to begin coming online in 2027.

Anthropic has also announced a strategic partnership with Microsoft and NVIDIA that includes $30 billion of Azure capacity, alongside a $50 billion investment in American AI infrastructure with Fluidstack.

The company says it trains and runs Claude across a range of AI hardware, including AWS Trainium, Google TPUs, and NVIDIA GPUs. It has also expressed interest in working with SpaceX on multiple gigawatts of orbital AI compute capacity.

International expansion and data residency

Anthropic says some of its future capacity expansion will be international, with enterprise customers in financial services, healthcare, and government requiring in-region infrastructure for compliance and data residency.

The company says its collaboration with Amazon includes additional inference capacity in Asia and Europe. It also says it is being selective about where capacity is added, with a focus on democratic countries where legal, regulatory, supply chain, and infrastructure conditions can support large-scale AI investment.

Anthropic recently committed to cover consumer electricity price increases caused by its data centers in the US. As it expands internationally, the company says it is exploring how to extend that commitment into new jurisdictions and work with local leaders in communities hosting its facilities.

For AI companies, usage limits are becoming a visible measure of infrastructure strain. Anthropic’s latest changes show how compute access is moving from back-end procurement to a front-line product issue for developers, enterprises, and AI platforms serving education.