Anthropic x SpaceX: Claude Code doubles its limits

Anthropic has signed a deal with SpaceX to access the full capacity of the Colossus 1 data center in Memphis. Claude Code limits are doubled for all paid plans, and peak-hour restrictions disappear for Pro and Max. A structural response to growing pressure from developers.

Key Takeaways

• Claude Code 5-hour session quotas are doubled for Pro, Max, Team, and Enterprise plans
• Peak-hour usage restrictions are removed for Pro and Max accounts
• The SpaceX deal gives Anthropic access to over 300 megawatts and 220,000 Nvidia GPUs

An unexpected deal between two tech giants

Elon Musk has never hidden his hostility toward Anthropic. The xAI founder and major SpaceX shareholder has been among the most vocal critics of the company founded by Dario Amodei. That context makes this deal all the more striking.

Anthropic signed an agreement with SpaceX to use the full computing capacity of the Colossus 1 data center, gaining access to over 300 megawatts of new capacity and more than 220,000 Nvidia GPUs, including H100s, H200s, and next-generation GB200 accelerators. The full picture is available in Anthropic’s announcement on higher Claude limits and the SpaceX compute deal.

Ami Vora, Anthropic’s Chief Product Officer, announced the deal at Code for Claude, a developer event streamed live from San Francisco. The new capacity went live on May 7, 2026. For context, see our earlier piece on Horizon: Claude Code and AI Agents Are Blowing Up Dev Budgets.

What actually changes for subscribers

Anthropic is doubling Claude Code’s 5-hour session limits for Pro, Max, Team, and Enterprise plans. For Pro and Max accounts, the reduction of limits during peak hours is also lifted. Anthropic keeps expanding the tool, recently opening Claude Code Artifacts sharing.

The doubling only applies to 5-hour sessions. Weekly caps remain unchanged. Users can pack more work into a single session, but their total weekly volume stays the same.

The timing matters. On April 21, Anthropic quietly tried to remove Claude Code from its $20-per-month Pro plan to ease infrastructure pressure, only to reverse course 48 hours later after pushback from developers. The SpaceX deal is the structural answer to that problem. On the user side, strategies to stop wasting Claude tokens have also gained visibility. The tool draws mixed reactions, Alibaba having outright blocked it for its employees.

Also on Horizon:

• China cuts 12,000 university degrees made obsolete by AI
• 79% of parents already use AI to raise their kids
• GPT-5.5 Instant beats doctors on health answers

An infrastructure race that is accelerating

The SpaceX deal does not stand alone. Anthropic is building infrastructure partnerships at a pace rarely seen in the industry.

The company has already signed a deal worth up to 5 gigawatts with Amazon, including nearly 1 gigawatt of new capacity by end of 2026, a 5-gigawatt agreement with Google and Broadcom set to come online in 2027, a partnership with Microsoft and Nvidia covering $30 billion in Azure capacity, and a $50 billion investment in American AI infrastructure with Fluidstack.

Anthropic has also signaled interest in developing several gigawatts of orbital AI computing capacity with SpaceX, extending the collaboration beyond terrestrial data centers.

The race for computing infrastructure is now just as decisive as the race for better models. Whoever controls compute controls the pace of development. That race feeds Anthropic’s financial trajectory, closing in on a $900 billion valuation.

Follow the story on Horizon.