Anthropic Secures Major Compute Deal With SpaceXAI Ahead of IPO
The partnership grants access to over 300 megawatts of capacity to enhance Claude Pro and Max services, marking a strategic shift as SpaceXAI seeks to validate its infrastructure to investors.
Anthropic and the merged entity SpaceXAI have announced a significant computing agreement, allowing the former to utilise resources from xAI's Colossus 1 supercomputer located in Memphis, Tennessee. The deal was confirmed on Wednesday at Anthropic's annual developer conference in San Francisco, with further details released via a blog post from SpaceXAI. This collaboration provides Anthropic with access to more than 300 megawatts of new capacity, comprising approximately 220,000 Nvidia GPUs, including H100, H200, and GB200 chips.
The primary objective of the partnership is to directly improve capacity for Claude Pro and Claude Max subscribers, addressing recent physical constraints that have led to service disruptions for developers using Claude Code. The agreement comes at a critical juncture for SpaceXAI, which is seeking to go public as soon as next month. By securing a deal with a leading AI laboratory, the merged entity aims to bolster credibility with investors regarding the viability of its data centre projects and its potential expansion into orbital AI compute capacity.
The Colossus 1 facility, constructed on a former Electrolux site in 2024, was built in approximately 122 days and is claimed by the company to be one of the largest and fastest AI supercomputers in the world. Despite the technical alignment, the partnership follows a period of public tension between the two entities. Earlier this year, Elon Musk publicly criticised Anthropic's AI models, describing their policies as misanthropic and alleging unproven biases, though he has since indicated a more positive view of the team after spending time with senior members.
This strategic move highlights the intense competition for compute resources within the industry. Reports indicate that Anthropic has committed to spending $200 billion on Google's AI cloud services and over $100 billion on Amazon technologies over the next decade. These substantial contracts account for a significant portion of the backlog at major cloud providers, underscoring the scarcity of high-performance hardware as demand for generative AI tools continues to surge.
The announcement has not been without controversy regarding the environmental impact of the infrastructure. Emissions from gas turbines at the data centre have prompted complaints from local residents, and environmental protesters recently demonstrated at a nearby SpaceX investor gathering ahead of the IPO. Nevertheless, having a potential customer like Anthropic interested in developing orbital AI compute capacity could help SpaceXAI demonstrate to the market that there will be buyers for its expensive supercomputing projects.
