Google to Pay SpaceX $920M Monthly for AI Compute
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Google agreed to pay SpaceX $920 million per month for AI compute covering ~110,000 Nvidia GPUs, plus CPUs, memory, and other components, from October 2025 through June 2029, with capacity ramping up through September at a reduced fee.
- SpaceX must deliver the committed GPUs by September 30, 2026, or Google can immediately terminate the deal or accept reduced capacity after a one-month grace period; after 2025, either party can exit with 90 days' notice.
- Google Cloud framed the deal as "bridge capacity" to meet surging demand for Gemini Enterprise, the business-tier subscriptions Google introduced in October.
- Anthropic announced a deal last month to use all of SpaceX's Colossus 1 data center capacity in Memphis—SpaceX's second major compute pact since its February merger with xAI, which valued the combined entity at $1.25 trillion.
- SpaceX's AI segment posted a $2.5 billion operating loss on just $818 million in revenue in Q1, with $7.7 billion of the company's $10.1 billion in Q1 capex committed to AI infrastructure.
- Alphabet—which invested in SpaceX at a $12 billion valuation in 2015—is selling $85 billion in stock, including a $10 billion investment from Berkshire Hathaway, and raised its 2025 capex guidance to $180-190 billion.
Why it matters: SpaceX lands a $920 million-per-month customer with a roughly 44-month revenue floor at a moment when its AI unit is burning $2.5 billion per quarter. The deal gives Google bridge capacity for Gemini Enterprise demand while Alphabet raises $85 billion in stock to fund its own $180-190 billion 2025 capex push, and neocloud competitors like CoreWeave and Nebius saw their stocks whipsaw on the news.
