admin The deal, announced on May 6, gives Anthropic use of all compute capacity at SpaceX’s Colossus 1 data centre within the month. The company said the additional infrastructure has enabled immediate increases in usage limits for Claude Code and the Claude API, easing constraints that have affected developers and enterprise customers as demand for AI coding and reasoning tools accelerates.
Claude Code users on Pro, Max, Team and seat-based Enterprise plans will see their five-hour rate limits doubled. Peak-hours limit reductions for Pro and Max accounts are also being removed, while rate limits for Claude Opus models on the API are being raised. The changes are aimed at Anthropic’s most active customers, including software teams using Claude for code generation, debugging, documentation and workflow automation.
The SpaceX agreement adds a new dimension to Anthropic’s infrastructure strategy. The company already has large-scale compute arrangements with Amazon, Google, Broadcom, Microsoft, Nvidia and Fluidstack. These include an agreement of up to 5 gigawatts with Amazon, with nearly 1 gigawatt expected by the end of 2026; a 5-gigawatt arrangement with Google and Broadcom beginning in 2027; $30 billion of Azure capacity through Microsoft and Nvidia; and a $50 billion investment in AI infrastructure with Fluidstack.
Anthropic’s move reflects a wider shift in the AI industry, where access to chips, power and data-centre capacity has become as important as model design. Leading AI developers are competing not only on benchmark performance but also on whether they can keep services available during periods of heavy demand. High-end Nvidia GPUs remain difficult to secure at scale, while electricity supply, cooling systems, networking equipment and construction timelines have become critical bottlenecks.
The agreement also carries strategic significance for SpaceX and Elon Musk’s broader technology empire. Colossus 1 has been associated with Musk’s AI infrastructure ambitions, and its use by Anthropic indicates that excess capacity can be monetised even while Musk-backed AI products compete in the same market. Claude is a direct rival to xAI’s Grok, as well as OpenAI’s ChatGPT, Google’s Gemini and Meta’s Llama-based systems.
Anthropic has framed the deal as part of a diversified approach to compute. Rather than relying on a single cloud provider or chip architecture, the company runs Claude across AWS Trainium, Google TPUs and Nvidia GPUs. This spread reduces dependence on any one supplier and gives the company flexibility as AI hardware markets evolve.
The company has also signalled interest in working with SpaceX on orbital AI compute capacity over the longer term. That idea remains at an early stage and faces significant technical and economic challenges, including launch costs, radiation protection, cooling in space and reliable power supply. For now, the commercial value lies in terrestrial capacity that can be brought online quickly.
Enterprise demand is a central driver of the expansion. Claude has gained traction among software developers, financial institutions, professional services firms and regulated industries that need high-capacity AI systems with clearer governance features. Anthropic has been building out tools for financial services, security, code modernisation and workplace integration, while also expanding support for Microsoft 365 and other enterprise environments.
The international dimension is also becoming more important. Anthropic has said some of its capacity growth will support customers in Asia and Europe, where data residency and compliance requirements can determine whether companies are able to adopt AI systems at scale. Regulated sectors such as finance, healthcare and government often require infrastructure located within specific jurisdictions and governed by familiar legal frameworks.
