Anthropic strikes SpaceX compute deal and lifts Claude usage limits
TL;DR:
- Anthropic has signed an agreement with SpaceX to use the entire Colossus 1 data centre — over 300 megawatts of capacity and more than 220,000 NVIDIA GPUs — coming online within the month.
- Effective today, Anthropic has doubled Claude Code’s five-hour rate limits for Pro, Max, Team and seat-based Enterprise plans, removed peak-hours throttling on Pro and Max accounts, and raised API rate limits for Claude Opus models.
- Resultsense view: the SpaceX angle is the eyebrow-raiser, but the more substantive signal for UK customers is the international expansion language — Anthropic explicitly cites in-region infrastructure for regulated industries including financial services and government, where data residency is a live procurement question.
The SpaceX deal joins a string of recent Anthropic compute commitments and underlines how aggressively frontier labs are buying ahead. The pattern of locking in capacity through unconventional partners — SpaceX in this case — also reflects the supply-side pressure that has pushed AI infrastructure into nuclear, gas, and now satellite-adjacent territory.
The compute stack
Alongside the SpaceX agreement, Anthropic lists four other significant commitments: an up-to-five-gigawatt Amazon agreement (with nearly 1GW of new capacity by the end of 2026), a five-gigawatt Google and Broadcom deal that begins coming online in 2027, a strategic partnership with Microsoft and NVIDIA that includes $30 billion of Azure capacity, and a $50 billion American AI infrastructure investment with Fluidstack. The company says it trains and runs Claude across AWS Trainium, Google TPUs and NVIDIA GPUs, and notes interest in eventually developing multiple gigawatts of “orbital AI compute capacity” with SpaceX.
What changes for customers today
For Claude Code subscribers, the headline practical change is doubled five-hour rate limits across paid tiers, and the removal of the peak-hours reduction on Pro and Max plans. API customers using Claude Opus will see noticeably higher rate limits. Anthropic also reiterated a recent commitment to cover any consumer electricity price increases caused by its US data centres — and said it is exploring how to extend that pledge to new jurisdictions as it adds international capacity.
UK and EU framing
Anthropic’s emphasis on partnering with “democratic countries whose legal and regulatory frameworks support investments of this scale” is, in context, a deliberate signal. The line is published in the same week the EU diluted its AI Act and the US Commerce Department tightened pre-deployment evaluations under CASI. For UK enterprise customers, particularly in banking and government — where ICO automated-decision-making rules under Article 22C are now in force and the AI Safety Institute holds bilateral testing arrangements — Anthropic’s flagged international expansion in Asia and Europe through its Amazon collaboration is the practically important detail.
Looking forward
The wider compute build-out also lands the same week Reuters reports Microsoft is considering shelving its 2030 clean-energy target as AI lifts power use. Anthropic’s pricing pledge on consumer electricity is the most visible attempt by a frontier lab to address that backlash. Whether it extends to the UK as in-region capacity comes online, and on what terms, is a useful procurement question for any UK CIO assessing supplier risk over the next 18 months.