Anthropic signed a deal with Elon Musk’s SpaceX to use the full compute capacity at xAI’s Colossus 1 data center in Memphis, Tennessee, the companies announced on May 6. The agreement gives Anthropic access to more than 300 megawatts of new capacity, equivalent to over 220,000 NVIDIA GPUs, coming online within the month.

Product Changes Effective Immediately

Three changes took effect the same day, according to Anthropic’s blog post:

  • Claude Code’s five-hour rate limits doubled for Pro, Max, Team, and seat-based Enterprise plans
  • Peak-hours throttling removed for Claude Code on Pro and Max accounts
  • API rate limits for Claude Opus models increased “considerably”

The rate limit changes tie infrastructure capacity directly to agent product performance. Anthropic framed the improvements as a consequence of the compute expansion, not an independent product decision.

Compute Portfolio

Colossus 1 is one piece of a broader capacity push. Anthropic listed its other recent compute agreements in the announcement: an up to 5 GW deal with Amazon (nearly 1 GW by end of 2026), a 5 GW agreement with Google and Broadcom (coming online in 2027), a $30 billion Microsoft and NVIDIA partnership for Azure capacity, and a $50 billion investment in American AI infrastructure with Fluidstack.

The company also “expressed interest in partnering with SpaceX to develop multiple gigawatts of orbital AI compute capacity,” according to the announcement. SpaceX recently filed FCC paperwork to launch a constellation of up to one million satellites for AI compute purposes, as Engadget reported.

The Musk Factor

The partnership follows months of public hostility from Musk toward Anthropic. He previously called the company “misanthropic” and said it “hates Western Civilization,” according to CNBC. The reversal came abruptly: Musk posted on X on May 6 that he had spent time with senior Anthropic team members and was “impressed,” adding that “no one set off my evil detector.”

Musk also announced in a separate post that xAI “will be dissolved as a separate company” and rebranded as SpaceXAI, CNBC reported. The timing coincided with the ongoing Musk v. Altman trial in Oakland, where Musk testified over three days last week.

“We reserve the right to reclaim the compute if their AI engages in actions that harm humanity,” Musk wrote on X.

The Compute Bottleneck Is Now Visible

Anthropic disclosed last month that demand for Claude had created “inevitable strain on our infrastructure,” impacting reliability and performance during peak hours, according to CNBC. The company is also in talks to raise funds at a $900 billion valuation.

The deal makes the compute-to-product pipeline explicit: more GPUs translate directly into higher rate limits for agent tools like Claude Code. For teams running autonomous coding agents against Anthropic’s API, the capacity expansion removes a constraint that was throttling production workloads. Whether 300 megawatts closes the gap between demand and supply, or simply buys time until the next bottleneck, depends on how fast agent adoption accelerates through the rest of 2026.