Oracle Expands Nvidia AI Server Deal to Multi-Gigawatt Scale, Adds Google TPUs

On October 28, 2026, Oracle announced an expanded cloud AI chip agreement with Nvidia, scaling Nvidia's latest Vera Rubin AI servers to multi-gigawatt capacity. Oracle also said it will begin offering Google's tensor processing unit (TPU) chips for the first time, and plans to add inference chips from SambaNova Systems and Cerebras Systems as part of its effort to diversify cloud chip suppliers.

Oracle said that even in Nvidia-based systems, it will move away from some of Nvidia's networking chips in favor of its own Acceleron networking chips. SambaNova and Cerebras both make chips designed for AI inference, the process by which AI systems respond to user requests.