Anthropic has entered into a significant agreement with SpaceX, committing to pay $1.25 billion per month until May 2029 for compute capacity across its COLOSSUS and COLOSSUS II platforms. This transaction highlights a broader trend where frontier AI companies secure long-term access to third-party compute resources to alleviate infrastructure bottlenecks and stabilize their model serving capabilities. As demand for high-performance GPU capacity continues to outstrip supply, partnerships with nontraditional compute providers have become increasingly common in the industry.

SpaceX: SpaceX is an aerospace company best known for rockets and satellite systems, but it also operates large-scale data center infrastructure in this context. Here, SpaceX is the provider of compute capacity to Anthropic under a long-term agreement across its Colossus data center assets.
Anthropic: Anthropic is an AI company that develops Claude, a large language model product used through chat and API access. In this news, Anthropic is the customer entering a compute agreement with SpaceX to expand its access to AI infrastructure and support product demand.

`json
{
“AI compute sourcing”: “Frontier AI companies are securing long-term access to third-party compute to address infrastructure constraints and maintain model serving capacity.”,
“Product capacity relief”: “When major labs acquire new compute resources, they aim to enhance usage capacity and reduce service limitations for paid users and API clients.”,
“Data center partnerships”: “AI infrastructure agreements now frequently involve nontraditional compute providers, highlighting the scarcity of high-performance GPU capacity in the industry.”
}
`