Key Highlights
- Partnership established between Anthropic and SpaceXAI grants access to Colossus 1 supercomputer infrastructure
- Claude Code users across Pro, Max, Team, and Enterprise tiers will see doubled five-hour rate limits
- Computing agreement provides Anthropic with 220,000+ Nvidia GPUs plus 300MW of additional capacity in 30 days
- Financial analysts at Wells Fargo predict revenue acceleration for Anthropic, currently at $44B ARR
- Dario Amodei, Anthropic’s CEO, disclosed Q1 2025 revenue surged 80-fold, creating infrastructure challenges and outages
In a strategic computing partnership, Anthropic has finalized an agreement with SpaceXAI to utilize the Colossus 1 AI supercomputer, providing the company behind Claude with substantial computational resources.
SpaceXAI, previously operating as xAI and responsible for creating the Grok conversational AI, completed its merger with SpaceX in recent months. Elon Musk oversees the unified entity alongside his Tesla responsibilities.
The partnership grants Anthropic immediate access to an impressive arsenal of over 220,000 Nvidia GPUs, accompanied by more than 300 megawatts of fresh capacity scheduled for deployment within 30 days.
Anthropic intends to leverage this computational expansion to enhance service capabilities for Claude users. Specifically, Claude Code will see its five-hour rate limits doubled across all paid subscription tiers, including Pro, Max, Team, and Enterprise plans.
Additionally, the company is eliminating peak-hour throttling restrictions on Claude Code for both Pro and Max subscribers. API rate limits for Opus model variants will similarly receive upward adjustments.
According to Wells Fargo analysts, this infrastructure deal positions Anthropic for continued revenue expansion. The firm highlighted that Anthropic’s annualized recurring revenue recently climbed to $44 billion, representing significant growth from $30 billion at April’s conclusion and $14 billion recorded in February.
Infrastructure Strain From Explosive Expansion
Dario Amodei, CEO of Anthropic, issued an apology earlier this week following service disruptions that affected users across multiple model offerings.
During remarks at Anthropic’s developer conference, Amodei revealed that first-quarter 2025 revenue multiplied by a factor of 80. The company had initially projected only 10x growth for the period.
“That is the reason we have had difficulties with compute,” Amodei explained. He emphasized that Anthropic is aggressively pursuing additional capacity expansion.
Beyond immediate infrastructure needs, the SpaceXAI collaboration may unlock ambitious future initiatives. Anthropic has signaled interest in collaborating with SpaceX on orbital AI computing infrastructure capable of delivering multiple gigawatts of processing power.
Comprehensive Capacity Strategy
The SpaceXAI arrangement represents just one component of Anthropic’s multi-pronged infrastructure expansion strategy.
The AI company previously secured an agreement with Amazon for up to 5 gigawatts of computing capacity. Initial gigawatt delivery is projected for late 2026.
Anthropic has additionally established capacity commitments with Google and Broadcom, with infrastructure scheduled for 2027 activation. A separate collaboration with Microsoft and Nvidia provides computing resources through the Azure cloud platform.
The company has also invested in Fluidstack to support domestic AI infrastructure development throughout the United States.
Among all these strategic partnerships, the Colossus 1 arrangement with SpaceXAI delivers the most immediate impact, with GPU resources becoming available within weeks.


