SpaceX discloses $920 million-a-month cloud compute agreement with Google

Google has agreed to pay SpaceX $920 million per month for cloud computing capacity, according to a SpaceX filing with the U.S. Securities and Exchange Commission.

SpaceX outlines the Google compute agreement

SpaceX disclosed a Cloud Service Agreement with Google LLC in a free writing prospectus filed with the SEC. The document says the agreement, dated June 5, 2026, covers access to approximately 110,000 NVIDIA GPUs and related components.

According to the SEC filing, Google’s monthly payments are set at $920 million from October 2026 through June 2029. AFP, in a report carried by RTL Today, also reported that full monthly payments are scheduled to begin in October 2026.

The Decoder reported that the deal gives Google access to around 110,000 Nvidia chips to help meet demand for its Gemini Enterprise platform. TechCrunch similarly reported that Google described the arrangement as bridge capacity for Gemini Enterprise demand.

A pre-IPO disclosure for SpaceX

TechCrunch reported that SpaceX lined up the Google compute deal ahead of its IPO. AFP also described the agreement as a pre-IPO deal under which SpaceX would provide AI computing capacity to Google.

The SEC filing is notable because it places specific figures on the transaction: the number of GPUs and related components, the monthly payment amount, and the expected payment period. Based on the filing’s stated monthly rate and term from October 2026 through June 2029, the contract represents a large recurring revenue commitment, though the filing excerpt does not specify every operational detail of the service.

Google turns to external capacity for AI demand

Google is one of the world’s largest cloud providers, but the reported agreement indicates that demand for AI infrastructure can exceed even major technology companies’ available capacity at certain points. TechCrunch reported that Google framed the deal as bridge capacity, suggesting the arrangement is intended to cover demand while other capacity becomes available.

The Decoder described the agreement as evidence of scarce AI infrastructure and of increasingly intertwined business relationships among major technology companies. The sources provided do not state whether Google will use the capacity exclusively for Gemini Enterprise, but both The Decoder and TechCrunch connect the deal to demand for Google’s enterprise AI offering.

What remains unclear

The SEC filing excerpt identifies the parties, the approximate GPU count, the monthly payment amount, and the payment window. It does not, in the provided excerpt, give a full technical breakdown of the deployment, data center locations, performance guarantees, or how Google will integrate the capacity into its broader cloud and AI operations.

For SpaceX, the agreement adds a large cloud-service revenue stream to its pre-IPO disclosures. For Google, the deal underscores the continuing pressure to secure high-end Nvidia GPU capacity for commercial AI services, even for companies that already operate major computing infrastructure.