Google has entered into a major long-term agreement with SpaceX to secure additional artificial intelligence computing power, committing to payments of $920 million per month through June 2029. The deal, disclosed in regulatory filings ahead of SpaceX’s anticipated IPO, reflects the growing demand for AI infrastructure as enterprise adoption accelerates.
Under the agreement, Google will gain access to approximately 110,000 Nvidia GPUs, along with supporting processors, memory, and related data centre infrastructure hosted by SpaceX. The full capacity is expected to be available by September 2026, with a phased ramp-up beginning earlier at reduced pricing.
The contract includes performance safeguards. If SpaceX fails to deliver the promised GPU capacity by the agreed deadline, Google can either terminate the arrangement immediately or continue under revised payment terms after a brief grace period. Starting next year, both companies will also have the option to exit the agreement with 90 days’ notice.
Google said the additional infrastructure is intended to support rising demand for Gemini Enterprise, its AI platform designed for large organisations. According to the company, customer adoption has exceeded initial expectations, creating the need for expanded computing resources.
The agreement follows SpaceX’s merger with xAI, Elon Musk’s artificial intelligence venture, creating a combined business valued at $1.25 trillion. It also highlights SpaceX’s growing focus on monetising its AI infrastructure investments ahead of its public market debut.
Despite significant AI-related spending, SpaceX continues to expand its data centre business, positioning itself as a major provider of large-scale computing capacity for leading AI companies.
