Oracle and NVIDIA are expanding their collaboration by making NVIDIA’s latest Vera Rubin NVL72 systems and Vera CPUs generally available on Oracle Cloud Infrastructure (OCI). The infrastructure is designed to meet growing customer demand for AI training, inference and complex agentic applications, supporting both GPU- and CPU-intensive workloads.
The companies aim to scale OCI Superclusters to multi-gigawatt capacity with the new NVIDIA hardware. Customers can scale GPU and CPU resources independently to match specific workloads and reduce idle capacity.
Oracle also plans to introduce the NVIDIA HGX Rubin NVL8, a more flexible, smaller configuration with eight GPUs, to broaden its AI infrastructure offerings.