NVIDIA and Oracle Expand Vera Rubin to Multi-Gigawatt Scale on OCI
NVIDIA and Oracle announced on October 27 an expansion of their AI infrastructure collaboration. NVIDIA Vera Rubin NVL72 systems and NVIDIA Vera CPUs are now generally available on Oracle Cloud Infrastructure (OCI), and the two companies plan to expand OCI Superclusters with Vera Rubin to multi-gigawatt scale, supporting growing demand for AI training, inference and agentic applications.
- NVIDIA Vera Rubin NVL72 and NVIDIA Vera CPUs are now generally available on OCI
- Planned expansion takes OCI Superclusters with Vera Rubin to multi-gigawatt scale
- HGX Rubin NVL8 incoming: eight Rubin GPUs connected through NVLink in an x86-based server
- HGX Rubin NVL8 features two 64-core Intel Xeon 6 CPUs, 4TB DDR5 memory and 6.4 Tbps RDMA networking
- OpenAI has trained and served GPT-6 Astra on an OCI Supercluster with more than 100,000 NVIDIA Blackwell GPUs in Abilene
- Vera CPUs target CPU-intensive agent tasks, with OCI Kubernetes Engine scaling CPU and GPU resources independently
A Multi-Gigawatt Supercluster Blueprint
OCI Superclusters combine NVIDIA NVL72 rack-scale systems, Oracle Acceleron networking with NVIDIA Spectrum-X Ethernet featuring Spectrum-6 switches and BlueField-4 DPUs, and gigawatt-class power and cooling to train and serve the largest AI models. Each rack connects 72 GPUs through NVIDIA NVLink scale-up networking, while OCI’s resilient, multiplanar network scales across hundreds of thousands of GPUs. The planned Vera Rubin expansion builds on OCI’s existing NVIDIA-powered infrastructure, including GB200 and GB300 NVL72 systems, with NVIDIA’s common CUDA software foundation helping customers support evolving models and workloads across hardware generations.
HGX Rubin NVL8 Brings a Smaller Rubin Option
OCI also plans to introduce NVIDIA HGX Rubin NVL8, delivering eight Rubin GPUs connected through NVLink in an x86-based server configuration. The planned platform features two 64-core Intel Xeon 6 CPUs, 4TB of DDR5 memory and 6.4 Tbps of high-speed RDMA networking, providing a flexible deployment option for AI training, fine-tuning and inference. According to NVIDIA, HGX Rubin NVL8 enables customers to adopt next-generation Rubin GPU technology in a smaller eight-GPU configuration while maintaining compatibility with x86-based applications.
OpenAI Demonstrates the Infrastructure at Frontier Scale
OpenAI has demonstrated how this infrastructure supports frontier AI. In Abilene, OpenAI has trained and served GPT-6 Astra on an OCI Supercluster with more than 100,000 NVIDIA Blackwell GPUs. “As our frontier models advance, we need infrastructure that can keep pace,” said Sreekanth Pothanis, engineering leader, compute infra at OpenAI. “NVIDIA Vera Rubin on OCI gives us the computing performance and scale to accelerate model training and serve increasingly demanding AI workloads.”
“NVIDIA Vera CPUs on OCI deliver the world-class loaded performance needed to execute code and run tools across our rapidly growing agentic sandbox workloads,” said Uday Ruddarraju, chief technology officer of compute at OpenAI. “We are bringing Vera into production on OCI to support these workloads at scale.” Mahesh Thiagarajan, president of Oracle Cloud Infrastructure at Oracle, said the collaboration with NVIDIA extends those capabilities across training, inference and agentic AI, helping customers keep GPUs productive, secure and scaling efficiently as their needs grow. “With Oracle, we’re bringing Vera Rubin and Vera CPUs to OCI to help customers put more of their infrastructure to work today and sustain its value as AI evolves,” said Ian Buck, vice president of hyperscale and high-performance computing at NVIDIA.
Vera CPUs Address the Agent Bottleneck
NVIDIA noted that AI agents do more than generate responses. They call tools, execute code and retrieve information to complete tasks, and slow execution of those CPU-intensive operations can delay results and leave GPUs waiting between model calls. OCI Compute with NVIDIA Vera CPUs combines high single-thread performance with memory bandwidth: the former accelerates individual tool calls and code execution, while the latter keeps CPU cores supplied with data as agents run concurrent tasks, helping reduce GPU idle time. OCI Kubernetes Engine further enables customers to scale CPU and GPU resources independently, while Oracle Acceleron provides hardware-based isolation for workloads running untrusted, AI-generated code. Customers with early access to Vera CPUs on OCI, including OpenAI, are exploring their potential across a range of AI workloads.
Source: NVIDIA Newsroom
