UK Claude subscribers get doubled limits in SpaceX deal

UK Claude Pro and Max subscribers will see doubled Claude Code rate limits from Thursday (7 May) after Anthropic signed an agreement to use all compute capacity at SpaceX’s Colossus 1 data centre.

The deal gives Anthropic access to more than 300 megawatts of capacity across over 220,000 NVIDIA GPUs at the Memphis facility, with the new compute coming online within the month.

Anthropic will use the additional capacity to directly improve service for Claude Pro and Claude Max subscribers, the two paid consumer tiers most popular among UK developers and writers.

The Memphis site originally housed Elon Musk’s xAI training infrastructure, and according to Data Center Dynamics the Anthropic lease covers a little under half of the company’s roughly 500,000 GPU fleet.

What are the changes?

Anthropic confirmed three immediate changes effective today.

Claude Code’s five-hour rate limits double for Pro, Max and Team plans.

The peak hours limit reduction disappears for Pro and Max accounts.

API rate limits also rise substantially for Opus models, the highest performance tier in the Claude family.

UK developers running Claude Code during London working hours stand to benefit most from the peak hours change.

The previous reduction tightened limits during the highest traffic window for European users, often pushing UK based teams onto smaller models or forcing them to wait for usage windows to reset.

Removing that cap means a Pro subscriber paying around £16 a month gets the full advertised capacity through the working day rather than a throttled version of it.

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.

— Claude (@claudeai) May 6, 2026

Musk’s reversal on Anthropic

The agreement is notable given Musk’s previous public criticism of Anthropic.

CNBC reports that Musk wrote on X in February that the company “hates Western civilization”, and he has previously labelled Anthropic “woke” and “misanthropic”.

On Wednesday, Musk said he had spent the past week with senior Anthropic staff and that no one had set off his “evil detector”.

Musk also confirmed plans to dissolve xAI as a separate company and rename it SpaceXAI, with model training now running on the newer Colossus 2 facility. He added that SpaceXAI reserves the right to reclaim compute if Claude takes actions that harm humanity.

The importance of compute

The SpaceX agreement is one of several major commitments Anthropic has announced in recent months.

The list includes a 5 gigawatt agreement with Amazon, a 5 gigawatt deal with Google and Broadcom, a $30 billion Azure capacity arrangement with Microsoft and NVIDIA, and a $50 billion American AI infrastructure investment with Fluidstack.

For UK subscribers the practical takeaway is that the AI services they pay for are now gated by raw access to power and chips.

Each new compute announcement translates fairly directly into higher usage caps, fewer throttling events and faster API response times.

It also signals that the cost of running frontier AI models is unlikely to fall in the short term, which makes the doubled limits a real consumer gain rather than a marketing gesture.

Anthropic also said it has expressed interest in working with SpaceX to develop multiple gigawatts of orbital AI compute capacity.

No timeline has been published for any space-based deployment, and the near-term impact for UK Claude subscribers stops at the Memphis facility coming online this month.

Now read: JD Sports brings AI checkout to UK shoppers this year