Nvidia and Oracle plan to expand Oracle Cloud Infrastructure (OCI) Superclusters with Vera Rubin NVL72 systems to multi-gigawatt scale. The companies also announced general availability of Vera Rubin NVL72 systems and Vera CPUs on OCI. The systems are intended for AI model training and inference, while the CPUs are aimed at tasks AI agents perform between model calls.
Each NVL72 rack connects 72 GPUs through Nvidia NVLink, according to Nvidia. OCI Superclusters also use Oracle Acceleron networking, Nvidia Spectrum-X Ethernet networking, Spectrum-6 switches and BlueField-4 DPUs. The companies say OCI’s network can scale across hundreds of thousands of GPUs.
Vera CPUs are designed for agent tasks such as running tools, testing code and retrieving data. Nvidia says OCI Kubernetes Engine lets customers scale CPU and GPU resources independently, while Oracle Acceleron provides hardware-based isolation for workloads running AI-generated code.
OpenAI trains models and runs tools on OCI
Nvidia says OpenAI has trained and served GPT-6 Astra on an OCI Supercluster in Abilene with more than 100,000 Nvidia Blackwell GPUs. The company also identifies OpenAI among customers with early access to Vera CPUs on OCI. OpenAI says it is bringing Vera into production on OCI for workloads that execute code and run tools.
OCI also plans to introduce HGX Rubin NVL8, an eight-GPU system for x86-based servers. Nvidia says the planned configuration has two 64-core Intel Xeon 6 CPUs, 4TB of DDR5 memory and 6.4 Tbps of RDMA networking. It is intended for AI training, fine-tuning and inference.
The companies describe Vera Rubin NVL72 systems and Vera CPUs as generally available on OCI. The multi-gigawatt supercluster expansion and HGX Rubin NVL8 are planned additions.
