California, 6 June 2026 – Alphabet Inc’s Google has signed a major cloud computing agreement with Elon Musk’s SpaceX, committing to monthly payments of US$920 million for artificial intelligence (AI) computing capacity through June 2029 as competition intensifies among technology companies racing to secure access to high-performance AI infrastructure.
The agreement, which begins in October 2026, could be worth approximately US$30 billion over its duration, underscoring the growing strategic value of computing power in the global AI industry.
Under the arrangement, Google will gain access to approximately 110,000 Nvidia graphics processing units (GPUs), alongside central processing units, memory chips and related infrastructure required to support large-scale AI model training and deployment.
The deal comes as demand for generative AI services continues to outpace available computing resources across the industry. Google said the agreement is intended to provide additional capacity for its AI offerings, particularly its Gemini Enterprise platform, which has experienced stronger-than-expected customer adoption.
The technology giant has been investing aggressively in AI infrastructure as enterprises increasingly deploy AI-powered tools across customer service, software development, research and business operations.
The latest agreement also reflects the emergence of a new competitive dynamic within the AI ecosystem, where companies that traditionally compete are increasingly collaborating to secure access to critical infrastructure.
SpaceX, through its AI operations and related computing assets, has been positioning itself as a large-scale provider of AI infrastructure. The company has invested heavily in data centre capacity, including facilities in Memphis, Tennessee, while expanding operations into Mississippi as it seeks to establish a stronger foothold in the rapidly growing AI compute market.
The agreement follows a similar infrastructure arrangement previously signed between SpaceX and AI developer Anthropic, signalling rising demand from technology firms seeking alternative sources of computing power amid persistent shortages of advanced AI chips.
While Google and SpaceX maintain a competitive relationship in several technology sectors, the two companies also share strategic ties. Earlier this year, SpaceX disclosed that Google held a stake of slightly more than 6% in the company at the end of 2025. Following the merger of SpaceX and xAI earlier this year, analysts estimate Google’s ownership interest now stands at approximately 5%.
Beyond AI infrastructure, the companies have reportedly explored potential collaborations involving orbital data centres and space-based computing projects, highlighting how next-generation technology development is increasingly blurring the lines between aerospace, cloud computing and artificial intelligence.
The agreement includes provisions allowing Google to terminate the contract if SpaceX fails to deliver the required Nvidia chip capacity by the end of September 2026. Both parties also retain the right to end the arrangement with 90 days’ notice.
The Ledger Asia Insights
This deal is less about cloud services and more about control of AI infrastructure.
As AI adoption accelerates globally, computing power is becoming the industry’s most valuable resource, creating a new class of technology providers focused on supplying GPUs, data centres and energy-intensive computing capacity. The willingness of Google to commit up to US$30 billion illustrates how critical infrastructure availability has become to maintaining competitiveness in AI.
For SpaceX and xAI, the agreement validates a broader strategy of monetising infrastructure assets beyond developing AI models. Rather than competing solely on software capabilities, the group is positioning itself as a foundational provider of AI compute resources, potentially opening a significant new revenue stream ahead of any future public market ambitions.
The deal also highlights a growing trend in the AI economy: competitors increasingly becoming customers. As demand continues to exceed supply, access to infrastructure may prove more important than ownership of the most advanced AI models themselves.
