Skip to content
IFAI

AI models

Anthropic taps SpaceX compute to double Claude usage limits

Published Reach 64

Anthropic just struck a deal with SpaceX to take over all capacity at the Colossus 1 data center, securing over 220,000 NVIDIA GPUs. The massive compute boost allows Anthropic to instantly double Claude Code's rate limits for paid plans, eliminate peak-hour restrictions, and raise Claude Opus API caps, while also exploring future orbital AI compute.

Anthropic taps SpaceX compute to double Claude usage limits
Image: therundown.ai

Just months ago, Elon Musk was publicly posting that Anthropic "hates Western Civilization" and ought to be renamed "Misanthropic." Now, he is handing them the keys to his biggest supercluster.

Anthropic has signed an agreement with SpaceX to take over all compute capacity at the Colossus 1 data center in Memphis. The deal injects over 300 megawatts of fresh capacity—translating to more than 220,000 NVIDIA GPUs—within the month.

Claude gets room to breathe

For power users who live inside Claude Code, the relief is landing immediately. Effective today, Anthropic is doubling Claude Code's five-hour usage limits across Pro, Max, Team, and seat-based Enterprise plans.

Pro and Max accounts will no longer face rate reductions during peak operating hours. Developers calling Claude Opus through the API are also seeing significant rate limit increases.

This cluster joins an already staggering stack of compute commitments. Anthropic has an up to 5 GW deal with Amazon that brings nearly 1 GW online by late 2026, alongside a 5 GW Google and Broadcom agreement that The Information reported as a $200B deal over five years.

Throw in $30B of Azure capacity via Microsoft and NVIDIA, plus a $50B buildout with Fluidstack, and Anthropic is pulling levers across every major cloud. The two companies even floated an early interest in building multiple gigawatts of orbital AI compute in space.

The enemy of my enemy

Musk brushed off the irony on X, stating that SpaceX will rent compute to "AI companies that are taking the right steps to ensure it is good for humanity."

Claude already runs across a mixed fleet of AWS Trainium chips, Google TPUs, and NVIDIA GPUs. To soothe pushback over utility bills, Anthropic previously promised to cover local consumer power price spikes caused by its US facilities, a pledge it is now exploring extending abroad.

Still, the broader dynamic is hard to miss. Musk gains a massive revenue tenant for SpaceXAI's real estate while arming OpenAI's most dangerous competitor with serious raw horsepower.

Why it matters

For developers and engineering teams, the daily friction of hitting hard caps inside Claude Code just dropped dramatically. If peak-hour throttling previously kept you from treating Claude as a primary pair programmer, this update makes the Pro and Max tiers much easier to justify right now.

We think Anthropic's multi-cloud balancing act is brilliant execution. Rather than locking its destiny to AWS or GCP alone, the company is hoovering up spare gigawatts from every deep-pocketed infrastructure player in the world, including Musk. That makes Anthropic virtually impossible for any single tech giant to corner.

For the broader ecosystem, it signals that the compute land rush is superseding personal feuds. SpaceXAI is quietly stepping into the role of a frontier compute landlord, proving that when hundreds of thousands of GPUs are on the table, pragmatism beats ideology every time.

Sources (2)

  1. [1]
    anthropic.com(opens in a new tab)Originalanthropic.com
  2. [2]
    theverge.com(opens in a new tab)Reporttheverge.com

Best of AI, daily