Anthropic Secures Massive Compute Deals With Google and SpaceX
Anthropic expanded its AI infrastructure through a $200 billion Google Cloud commitment and a partnership to utilize SpaceX's Colossus 1 data center in Memphis.
Anthropic rapidly expanded its global computing infrastructure through a series of massive financial and technical agreements in May 2026. The AI startup committed to spending $200 billion with Google Cloud over five years to acquire chips and computing capacity, an agreement that accounts for more than 40% of Google's recently disclosed revenue backlog. Parallel to this, Alphabet Inc. is investing up to $40 billion in Anthropic, with $10 billion provided immediately and $30 billion tied to performance milestones.
To further boost its Claude AI models, Anthropic partnered with SpaceX and xAI to utilize the full capacity of the Colossus 1 data center in Memphis, Tennessee. This facility provides over 300 megawatts of capacity and more than 220,000 Nvidia GPUs. The infrastructure boost allowed Anthropic to immediately double Claude Code rate limits for Pro, Max, Team, and Enterprise plans and increase API limits for Claude Opus models. The companies also expressed interest in developing orbital AI compute capacity.
Despite the technical gains, the Colossus 1 site remains under a congressional probe and faces legal challenges regarding Clean Air Act violations due to its use of gas turbines. The facility also continues to rely on Memphis drinking water for cooling after Elon Musk delayed the construction of a promised water reuse project. Beyond these deals, Anthropic has secured capacity through Amazon, Broadcom, and CoreWeave, and is pursuing further infrastructure agreements with Microsoft and Fluidstack.