“Vera is arriving at a turning point for AI. As intelligence becomes agentic — capable of reasoning and acting — the importance of the systems orchestrating that work is elevated,” Huang said in a statement. “The CPU is no longer simply supporting the model; it’s driving it. With breakthrough performance and energy efficiency, Vera unlocks AI systems that think faster and scale further.”
The new chip expands Nvidia’s CPU lineup following the earlier Grace processor and is intended to support what the company calls “AI factories” — large-scale computing environments built to run AI services and agents. Nvidia says Vera delivers higher throughput and responsiveness for workloads such as coding assistants, enterprise agents and consumer AI tools.
Several hyperscale cloud providers and infrastructure companies are already collaborating with Nvidia on deployments. Partners include Alibaba, CoreWeave, Meta and Oracle Cloud Infrastructure, alongside hardware manufacturers such as Dell Technologies, HPE, Lenovo and Supermicro.
The processor is part of the Vera Rubin NVL72 platform, which pairs Vera CPUs with Nvidia GPUs using the company’s NVLink-C2C interconnect. Nvidia says the connection enables 1.8 TB/s of coherent bandwidth, allowing faster data sharing between CPUs and GPUs for AI workloads.
Nvidia also introduced a rack system built around the new processor. A single rack integrates 256 liquid-cooled Vera CPUs and can support more than 22,500 concurrent CPU environments running at full performance. The design uses Nvidia’s MGX modular architecture and is supported by dozens of ecosystem partners.
The chip includes 88 custom Nvidia-designed Olympus cores built to handle AI infrastructure tasks such as orchestration, analytics pipelines and runtime systems. Each core can run two tasks simultaneously through Nvidia Spatial Multithreading to support multi-tenant environments where many AI jobs run at once.
To support demanding workloads, the processor also features a new memory system based on LPDDR5X, delivering up to 1.2 TB/s of bandwidth. Nvidia says the design provides twice the bandwidth while using half the power compared with general-purpose CPUs.
Early partners have begun testing the platform. Alex Gallego, CEO of streaming data company Redpanda, said the company saw improved performance when benchmarking workloads on Vera. “Redpanda recently tested Nvidia Vera running Apache Kafka-compatible workloads and saw dramatically better performance than other systems we’ve benchmarked, delivering up to 5.5 times lower latency,” he said in a statement.
Research institutions are also preparing to deploy the new processor. Planned users include the Leibniz Supercomputing Centre, Los Alamos National Laboratory, Lawrence Berkeley National Laboratory’s National Energy Research Scientific Computing Center and the Texas Advanced Computing Center.
“At TACC, we recently tested Nvidia’s Vera CPU platform as we prepare for deployment in our upcoming Horizon system — and running six of our scientific applications, we saw impressive early results,” said John Cazes, director of high-performance computing at TACC. “Vera’s per-core performance and memory bandwidth represent a giant step forward for scientific computing, and we look forward to bringing Vera-based nodes to our CPU users on Horizon later this year.”
Nvidia said Vera systems will be available from partners later this year as the company expands its hardware platform designed to support large-scale AI deployments.
This analysis is based on reporting from GamesBeat.
Image courtesy of Nvidia.
This article was generated with AI assistance and reviewed for accuracy and quality.
