Standalone Nvidia Vera CPUs land in Oracle Cloud's multi-gigawatt deployment — our independent benchmarking of a local server reveals Vera's massive memory bandwidth lead over dual Xeon 6 server

The announcement comes shortly after Google Cloud announced general availability for Vera Rubin last week, marking a significant achievement for Nvidia with two major hyperscalers now supporting the technology. Oracle specifically highlighted standalone Vera instances on OCI Superclusters, which represents a notable deployment in the GPU-dominant and AI-focused neocloud space.

Neocloud Expansion

Neocloud providers have been actively deploying Vera technology in recent weeks, with Nscale, CoreWeave, and Nebius among those offering instances. This widespread adoption in specialized cloud environments contrasts with Oracle's standalone deployment at a hyperscaler level, potentially signaling different strategic approaches to the technology.

Tom's Hardware has conducted benchmarks on a Vera server for an upcoming review, though these tests were not performed on OCI infrastructure. The performance characteristics observed may provide insights into how the technology could perform in Oracle's environment.