NVIDIA and Oracle make Vera Rubin systems and the Vera CPU available on Oracle Cloud

NVIDIA announced new steps with Oracle in a press release on 27 October 2026. NVIDIA's Vera Rubin NVL72 rack systems and Vera CPU are now generally available on Oracle Cloud Infrastructure (OCI). The two companies plan to expand OCI Superclusters with Vera Rubin NVL72 systems to multi-gigawatt scale; the release does not give a specific gigawatt or GPU count.

• In every NVL72 rack, 72 GPUs are linked together over NVLink.
• According to the release, Vera CPUs are meant for the processor-bound steps agents take between two model calls, for example calling tools, running code tests and fetching data.
• OCI Kubernetes Engine lets customers grow CPU and GPU capacity separately from each other.
• OCI also plans to offer HGX Rubin NVL8 servers with eight Rubin GPUs; no date was given.

Why it matters

According to the release, Oracle Acceleron keeps workloads that run untrusted, AI-written code separated at the hardware level, and OpenAI is among the customers with early access to the Vera CPU. For companies building AI agents, scaling CPU and GPU capacity separately could help cut idle capacity and the cost that comes with it. Teams in Türkiye planning agent projects on OCI or another cloud can measure early how much of a workload needs GPUs and how much needs CPUs, and choose server types accordingly.

• Vera Rubin NVL72 and the Vera CPU are generally available on OCI.
• Capacity is planned to grow to multi-gigawatt scale.
• CPU and GPU capacity can be scaled separately.