The Biggest AI Infrastructure Deal Yet? Google Commits Billions to SpaceX Ahead of Record IPO
Artificial Intelligence & Machine Learning News
Just days before its highly anticipated public debut, SpaceX has secured another massive artificial intelligence infrastructure agreement — this time with Google.
According to a recent regulatory filing, Google agreed to spend approximately $920 million per month beginning in October 2026 and continuing through June 2029. In return, the tech giant will gain access to an enormous pool of high-performance AI hardware, including roughly 110,000 NVIDIA GPUs, CPUs, memory systems, and related computing infrastructure.
The agreement marks yet another sign that demand for AI computing capacity has reached historic levels. As companies race to deploy increasingly sophisticated artificial intelligence systems, access to large-scale compute infrastructure has become one of the most valuable resources in technology.
The arrangement closely resembles the blockbuster compute partnership SpaceX signed earlier this year with Anthropic. Under that deal, Anthropic committed to paying approximately $1.25 billion per month through 2029 for dedicated access to the Colossus 1 supercomputing campus near Memphis, Tennessee. That facility was initially constructed by xAI, now integrated into SpaceX’s broader AI ambitions.
Industry analysts believe Google’s agreement represents access to roughly half the computing power allocated under the Anthropic partnership. While SpaceX did not disclose which data center would support Google’s workloads, Elon Musk previously indicated that future AI development for xAI would likely rely heavily on the upcoming Colossus 2 facility.
Unlike Anthropic, which reportedly faced significant compute limitations before partnering with SpaceX, Google already ranks among the world’s largest owners of AI infrastructure. Even so, soaring adoption of generative AI tools appears to be forcing the company to secure additional capacity faster than it can build internally.
Google stated that demand for its AI products — particularly its Gemini Enterprise platform and enterprise AI agent services — has exceeded internal expectations. The company described the partnership as a temporary but strategic solution designed to bridge the gap while its own infrastructure expansion accelerates.
The agreement also underscores the extraordinary scale of spending currently taking place across the artificial intelligence sector. Alphabet, Google’s parent company, has already committed more than $180 billion in capital expenditures this year alone and recently warned investors that spending could rise substantially again in 2027. To help finance that expansion, Alphabet announced plans for an $80 billion equity offering.
Like the Anthropic agreement, the Google contract includes provisions allowing either company to exit the partnership after December 2026 with 90 days’ notice. The filing also notes that Google’s access to the infrastructure will gradually scale up through September at reduced pricing before reaching full deployment levels.
SpaceX must deliver the agreed compute capacity by September 30, 2026. Failure to do so would allow Google to either terminate the agreement after a short grace period or renegotiate lower monthly payments based on the amount of infrastructure successfully delivered.
The timing of the announcement is especially notable because it arrives just one week before SpaceX is expected to begin trading publicly on the Nasdaq. Securities filings indicate the company aims to raise approximately $75 billion at a valuation near $1.75 trillion, potentially making it the largest IPO in financial history.
Google has maintained a long-standing investment relationship with SpaceX for years. Following the IPO, analysts estimate Google’s stake in the company could exceed $100 billion in value.
The two firms are also reportedly exploring ambitious long-term collaborations involving orbital data centers, a concept many believe could become a critical component of SpaceX’s next-generation AI and cloud computing strategy after going public.
As the global AI arms race intensifies, the battle is no longer just about algorithms or software. Increasingly, victory may depend on who controls the world’s largest pools of computing power.
