At a glance
- NVIDIA and Oracle today announced plans to expand Oracle Cloud Infrastructure (OCI) Superclusters with NVIDIA Vera Rubin NVL72 systems to multi-gigawatt scale, supporting growing customer demand for AI training, inference and agentic applications
NVIDIA and Oracle Expand Vera Rubin to Multi-Gigawatt Scale on OCI
NVIDIA and Oracle have announced an expanded collaboration to bring NVIDIA Vera Rubin NVL72 systems to Oracle Cloud Infrastructure (OCI) Superclusters at multi-gigawatt scale. The move addresses growing customer demand for AI training, inference, and agentic applications, with both NVIDIA Vera Rubin NVL72 and NVIDIA Vera CPUs now generally available on OCI.
The companies also plan to introduce NVIDIA HGX Rubin NVL8 systems to OCI, providing organizations with additional options for frontier model training, inference efficiency, and agentic workloads.
OCI Superclusters and Networking Architecture
OCI Superclusters combine NVIDIA NVL72 rack-scale systems with Oracle Acceleron networking, NVIDIA Spectrum-X Ethernet networking featuring NVIDIA Spectrum-6 switches, and NVIDIA BlueField-4 DPUs. The infrastructure includes gigawatt-class power and cooling designed to train and serve the largest AI models.
Each rack connects 72 GPUs through NVIDIA NVLink scale-up networking. 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, which includes NVIDIA GB200 and GB300 NVL72 systems. NVIDIA’s common CUDA software foundation and continued software optimization are intended to help customers support evolving models and workloads across hardware generations.
HGX Rubin NVL8 Configuration Details
OCI 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.
This configuration offers a flexible deployment option for AI training, fine-tuning, and inference. It enables customers to adopt next-generation Rubin GPU technology in a smaller eight-GPU configuration while maintaining compatibility with x86-based applications. The flexibility allows customers to adapt AI infrastructure to different models and workloads as needs evolve, using accelerated computing capacity across both training and inference to make better use of infrastructure investments.
Vera CPUs Target Agentic Workloads
NVIDIA Vera CPUs on OCI are designed to accelerate the CPU-intensive tasks that AI agents perform between model calls, including executing tools, testing code, and retrieving data. Slow execution of these 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 to help agents complete CPU-intensive tasks faster. High single-thread performance accelerates individual tool calls and code execution, while high memory bandwidth keeps CPU cores supplied with data as agents run concurrent tasks. Together, these capabilities help reduce GPU idle time and maintain consistent performance as workloads scale.
OCI Kubernetes Engine enables customers to scale CPU and GPU resources independently, helping reduce idle capacity and put more infrastructure to productive use. Oracle Acceleron provides hardware-based isolation for workloads running untrusted, AI-generated code.
OpenAI Deployment and Operator Implications
OpenAI has trained and served GPT-6 Astra on an OCI Supercluster with more than 100,000 NVIDIA Blackwell GPUs in Abilene. Customers with early access to NVIDIA Vera CPUs on OCI, including OpenAI, are exploring their potential across a range of AI workloads. OpenAI is bringing Vera into production on OCI to support agentic sandbox workloads at scale.
For data center operators, the expansion implies infrastructure must support multi-gigawatt power and cooling densities while maintaining workload isolation and security. The ability to scale CPU and GPU resources independently addresses a common operational challenge: idle capacity during mixed training and inference cycles.
Market Context
The collaboration extends Oracle and NVIDIA’s multigenerational partnership across training, inference, and agentic AI. By combining NVIDIA Vera Rubin GPUs and Vera CPUs with Oracle Acceleron networking and AI-optimized data centers, the companies aim to help customers keep GPUs productive, secure, and scalable as demand grows.
NVIDIA emphasized that AI factories must be productive, durable, and fungible—delivering efficient computing, remaining useful through continued software advances, and serving different models and workloads as demand changes.
MegaWatt Intelligence Desk
Hyperscale & Grid Analyst. `More stories by $MegaWatt Intelligence Desk →`
