In a significant move to secure the massive computational power required for its ambitious artificial intelligence goals, Elon Musk’s xAI has signed a major deal to build a supercomputer on Oracle’s cloud infrastructure. The partnership will see xAI lease tens of thousands of Nvidia H100 GPUs from Oracle, creating one of the world’s largest AI training clusters to power the development of its next-generation Grok model.
The deal represents a monumental win for Oracle, positioning its Oracle Cloud Infrastructure (OCI) as a formidable competitor to established leaders like Amazon Web Services, Microsoft Azure, and Google Cloud in the lucrative AI infrastructure market. While specific financial terms were not disclosed, the agreement underscores the intense demand for high-performance computing resources as the AI arms race escalates.
This collaboration is a critical step in realizing xAI’s plan for a “Gigafactory of Compute,” a term Musk has used to describe the unprecedented scale of processing power needed to achieve artificial general intelligence. By leveraging Oracle’s established infrastructure, xAI can rapidly deploy the necessary hardware to train increasingly sophisticated models designed to rival those from OpenAI, Google, and Anthropic. The partnership is also strengthened by the long-standing relationship between Musk and Oracle Chairman Larry Ellison, who was a significant investor in Musk’s acquisition of Twitter (now X).
For the broader tech industry, this move highlights the critical role of cloud providers as kingmakers in the AI revolution. Access to vast fleets of GPUs is the primary bottleneck for AI development, and companies that can supply this compute are becoming central to the entire ecosystem. As xAI prepares to train Grok 3 and beyond, its reliance on Oracle’s hardware will be a key factor in its quest to build what Musk describes as “a maximally curious AI.”
English 
Deutsch