Nvidia and Oracle announced on October 27, 2026, that Vera CPUs and Vera Rubin NVL72 systems were generally available on Oracle Cloud Infrastructure (OCI), with OpenAI among customers having early access, according to Nvidia's announcement. Nvidia CEO Jensen Huang said, "We are bringing Vera into production on OCI to support these workloads at scale." Nvidia says its Olympus cores deliver 1.8x faster single-threaded performance, while AMD says its own tests show 1.2x the per-core performance and 2.24x the platform-level performance of a Vera-based platform. Neither company has published a price, and the independent data is thin. Oracle also committed to expanding its Vera Rubin superclusters to multi-gigawatt scale, a plan rather than delivered capacity, and described HGX Rubin NVL8, an x86-based server with eight Rubin GPUs connected through NVLink, as a configuration it intends to introduce, with no shipping date or price. Earlier, on October 21 and 22, Nvidia said Vera Rubin was in full production and that partners were shipping systems to AI labs.