Anthropic taps SpaceX AI data center to ease AI crunch, doubles Claude Code limits as Musk shifts stance

Anthropic just secured a massive shot of computing power, and it could reshape the pace of its AI push.

SpaceX said Wednesday it will lease its Colossus 1 data center to Anthropic, linking Elon Musk’s company with a rival he had publicly criticized months ago. The deal gives the Claude maker access to one of the largest AI compute clusters in the country, located in Memphis, Tennessee, and packed with more than 220,000 Nvidia processors. Within weeks, Anthropic expects to add 300 megawatts of capacity.

“We’ve agreed to a partnership with SpaceX that will substantially increase our compute capacity. This, along with our other recent compute deals, means that we’ve been able to increase our usage limits for Claude Code and the Claude API.” Anthropic said in a statement.

As part of the agreement, Anthropic will gain access to more than 300 megawatts of compute capacity and said it has “expressed interest” in working with SpaceX on future space-based data centers that could reach multiple gigawatts. The company added that the new capacity will directly benefit its paid Claude Pro and Claude Max subscribers.

A compute surge unlocks faster Claude and bigger ambitions

That kind of scale arrives at a critical moment. Demand for Anthropic’s coding tools has surged, putting pressure on its infrastructure and forcing limits on how much developers can use Claude Code. With new capacity coming online, the company is lifting those constraints. Paid users will see rate limits double, and peak-hour caps for Pro and Max plans are being removed. Developers will be able to send far more requests to its Claude Opus models without hitting ceilings that previously slowed work.

The timing lines up with Anthropic’s developer event in San Francisco, where the company rolled out a feature it calls “dreaming.” The system reviews past interactions across sessions, identifies patterns, and updates files that store user preferences and context. The goal is simple: make the AI feel less like a stateless tool and more like a system that builds continuity over time.

The SpaceX partnership carries weight beyond raw compute. SpaceX is preparing for an eventual IPO and is pitching a broader role in AI infrastructure. Landing Anthropic as a customer gives that effort credibility at a moment when demand for data center capacity is stretching supply across the industry.

Musk’s stance on Anthropic has shifted over time. In a post on X, he said he decided to move forward with the deal after meeting the company’s leadership last week. Their focus on building AI that is “good for humanity” stood out to him. “No one set off my evil detector,” he wrote.

“By way of background for those who care, I spent a lot of time last week with senior members of the Anthropic team to understand what they do to ensure Claude is good for humanity and was impressed,” Musk said.

Same here.

By way of background for those who care, I spent a lot of time last week with senior members of the Anthropic team to understand what they do to ensure Claude is good for humanity and was impressed.

Everyone I met was highly competent and cared a great deal about…

— Elon Musk (@elonmusk) May 6, 2026

That tone marks a clear break from earlier comments. In February, Musk had criticized Anthropic’s direction, writing that it risked becoming misaligned with its stated mission. The new partnership suggests a more pragmatic approach, at least for now. Musk added that SpaceX has moved its own AI training work to a separate facility, Colossus 2, and plans to offer computing capacity to other AI companies under similar terms

Anthropic is already thinking past this deal. The company said it is interested in working with SpaceX on orbital data centers, a long-term idea Musk has floated as part of his broader vision for space-based infrastructure, Reuters reported. The concept would require significant capital and engineering breakthroughs, yet it signals how far the race for compute could go.

At the same time, Anthropic is sharpening its focus on enterprise customers. Its latest releases point in that direction. The “dreaming” feature is tied to tools that manage AI agents, programs that can carry out tasks with limited human input. The company has been expanding those capabilities, including systems that can break down work and assign it to specialized agents.

Earlier this week in New York, Anthropic introduced a set of finance-focused AI agents, highlighting how much of its revenue now comes from business users. The tech sector leads, followed by financial firms that are testing automation across analysis, reporting, and operations.

On stage in San Francisco, Claude Code head Boris Cherny showed how developers can set up routines that trigger the system to act autonomously. “The default isn’t, ‘I’m going to prompt Claude Code.’ The default is now, ‘I will have Claude prompt Claude Code,’” he said.

He closed with a simple message about where things stand. “The capability is already here. The gap left is how fast we put it to work.”