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
Nvidia’s Vera CPU is now generally available on Oracle’s OCI Superclusters, a win for Nvidia as it moves past NVL72 to a standalone Vera deployment at a major hyperscaler.
Nvidia and Oracle have officially released Nvidia’s Vera CPUs on Oracle Cloud Infrastructure (OCI) Superclusters, according to a recent announcement at Oracle AI World. This marks Nvidia’s first standalone CPU push into the hyperscale data center market, challenging x86 competition on Oracle’s platform.
Independent benchmarking on a local server has revealed that Vera’s memory bandwidth outperforms that of dual Xeon 6 servers. Nvidia’s memory system boasts a bandwidth of 1.2 TB/s, which surpasses the capabilities of Intel’s Xeon 6980P system. This advantage is particularly noticeable in workloads involving extensive database queries or large amounts of documentation.
While Vera is not intended to directly compete with x86 processors, its strong performance in memory bandwidth and Python execution makes it an attractive option for agentic AI workloads and general data center applications alike. Oracle’s cautious approach to deploying multi-gigawatt Vera instances highlights the company’s cautious optimism regarding the chip’s potential in the data center ecosystem.
Written by urgent.news from Tom's Hardware's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- Oracle to offer Google TPUs on cloud infrastructure in 2027 investing.com
- Oracle expands chip supply deal with Nvidia, adds Google and Cerebras offerings channelnewsasia.com
