Google has opened a new path to expand its artificial intelligence infrastructure with its multi-year cloud agreement with SpaceX. Company, $920 million each month between October 2026 and June 2029 paying about 110,000 Nvidia GPUsIt will access the CPU, memory and related components. This capacity is especially important for Google, which wants to meet the demand for Gemini Enterprise.
The deal comes on the heels of SpaceX’s similar collaboration with Anthropic. Anthropic announced that SpaceX has increased Claude Code limits and increased Claude Opus API limits with capacity in the Colossus 1 data center. Thanks to this partnership, the company more than 300 megawatts new capacity and More than 220,000 Nvidia GPUs indicates that you have gained access.
New balance of power in artificial intelligence infrastructure
On Google’s side, the deal will go to full scale in October, but the company retains the right to terminate the contract if capacity targets are not met. In fact, this article shows that AI companies are now competing not only on model quality but also on timely access to hardware.
With these agreements, SpaceX turns its high-cost transaction infrastructure into a source of income. For Google, the result means higher Gemini capacity, and for developers, it means more stable enterprise AI service. Long story short, data center owners are now becoming as decisive as model manufacturers in the GPU war.