NVIDIA and Oracle Expand NVIDIA Vera Rubin to Multi-Gigawatt Scale and Bring NVIDIA Vera CPU to Oracle Cloud Infrastructure
nvidia ·
NVIDIA and Oracle have announced the general availability of NVIDIA Vera Rubin NVL72 systems and NVIDIA Vera CPUs on Oracle Cloud Infrastructure. This collaboration expands OCI Superclusters to multi-gigawatt scale, supporting the growing demand for frontier AI training, inference, and agentic applications. The infrastructure combines GPU-intensive capabilities with CPU resources for tasks like tool execution and code testing, allowing customers to scale workloads independently. Additionally, plans are in place to introduce HGX Rubin NVL8 systems, which offer a flexible eight-GPU configuration using Intel Xeon processors. OpenAI is already utilizing this infrastructure to train and serve models such as GPT-6 Astra on large-scale superclusters. The integration aims to reduce GPU idle time by accelerating CPU-bound agent operations through high single-thread performance and memory bandwidth. Industry leaders from both companies emphasize that this partnership helps organizations accelerate AI development while maintaining efficient and secure infrastructure.
nvidia