Anthropic doubles Claude Code limits following strategic compute deal with SpaceX
At its Code with Claude conference, the company announced a partnership with SpaceX to utilise the entire Colossus 1 data centre in Memphis, enabling significant upgrades to Pro and Max plan usage windows.
Anthropic has announced a strategic partnership with SpaceX to utilise the entire compute capacity of the latter's data centre in Memphis, Tennessee. The agreement was revealed at the company's Code with Claude developer conference, where CEO Dario Amodei stated the primary objective is to increase usage limits for Pro and Max plan subscribers.
As a direct result of the deal, Anthropic has doubled the five-hour window limits for Claude Code and removed peak-hours limit reductions for these specific accounts. The company also confirmed it has raised API limits for its Opus model, aiming to alleviate the service disruptions and capacity constraints that have frustrated users on platforms such as Hacker News and Reddit.
The infrastructure underpinning this expansion is SpaceX's Colossus 1 supercomputer, which features over 220,000 NVIDIA GPUs. The firm claims the agreement grants access to more than 300 megawatts of new compute capacity, supported by dense deployments of H100, H200, and next-generation GB200 accelerators. SpaceX highlighted these specifications as central to the deal's ability to meet the surging demand for AI services.
This agreement marks a significant shift in the relationship between the two companies, following Elon Musk's previous public criticism of Anthropic. While Musk had previously declared on X that "Anthropic hates Western Civilization," he later noted that discussions with senior members of the Anthropic team left him impressed and that no one set off his "evil detector."
The surge in demand driving this partnership is attributed to users shifting away from OpenAI following military controversy and a broader industry move toward multi-agent workflows. Anthropic has also expressed interest in collaborating with SpaceX to develop gigawatts of orbital compute capacity, linking this to the challenge of terrestrial power and cooling limitations for future systems.
The announcement follows similar deals with Microsoft, Amazon, and others to address capacity issues, though gains from some of those arrangements will take time to materialise. The infrastructure is expected to come online within a month to immediately resolve the current service disruptions affecting developers relying on the platform's advanced capabilities.
