NVIDIA Vera CPU moves into production for agentic AI workloads

NVIDIA has announced Vera, a high-performance CPU built for AI agents and the wider set of workloads now running inside AI factories, including reinforcement learning, data processing and orchestration.

The processor is now in full production and is designed to complete tasks 1.8x faster than x86 CPUs, according to NVIDIA. Vera follows the company’s Grace CPUs, which NVIDIA said have reached nearly 2.5 million shipments to date, and extends its CPU strategy into systems where AI agents are expected to run code, use tools, evaluate results and handle more complex workflows.

Vera will serve as the CPU behind standalone Vera servers, NVIDIA Vera Rubin systems and Vera BlueField-4 STX AI storage platforms. NVIDIA said customers exploring or planning to adopt the CPU include NYSE, Anthropic, OpenAI, SpaceXAI, ByteDance, CoreWeave, Lambda, Nebius, Nscale and Oracle Cloud Infrastructure.

Vera is designed for CPU-heavy agentic workloads

NVIDIA said Vera is powered by Olympus, a custom CPU core built for the CPU-side work required by AI agents and data infrastructure. These tasks include Python runtimes, sandboxed code execution, orchestration logic and analytics pipelines.

The processor includes 88 Olympus cores, Spatial Multithreading and an LPDDR5X memory subsystem with up to 1.2TB/s of bandwidth. NVIDIA said the design helps reduce the time agents spend waiting on CPU-bound steps, allowing AI factories to keep accelerators running more efficiently across training, inference and agentic execution.

“AI agents will be the largest users of computing,” said Jensen Huang, founder and CEO of NVIDIA. “Vera is the first CPU designed for that future — built to run agentic AI at hyperscale with extraordinary performance, efficiency and programmability.”

NVIDIA also cited Phoronix benchmark results, saying Vera delivered the fastest overall performance across agentic workloads, including code compilation, Python, Java and database processing. These workloads are described as important to modern AI factories, especially where agents need tool use and sandbox execution.

NVIDIA links Vera to AI factory economics

NVIDIA is positioning Vera around the economics of AI factories, where the company said the focus is moving from cores per dollar to tokens per dollar. In that model, CPU performance affects how quickly agentic, data-processing and orchestration work can be completed before workloads move through the rest of the AI system.

Vera will act as the host CPU for NVIDIA Vera Rubin platforms through second-generation NVIDIA NVLink-C2C interconnect technology, which provides up to 1.8TB/s of coherent bandwidth between CPU and GPU. NVIDIA said Vera also extends NVIDIA Confidential Computing at rack scale for agentic workloads.

The NVIDIA Vera BlueField-4 STX processor combines Vera with high-performance networking, storage acceleration and in-silicon security, creating AI-native data platforms designed for secure deployment.

NYSE is among the early customers named by NVIDIA. The exchange operator said it plans to use Vera CPUs in collaboration with Redpanda and HPE to scale capacity and optimise latency across its infrastructure.

“At the NYSE, our focus is to optimise the latency, throughput and reliability of the systems underpinning our unrivalled infrastructure,” said Lynn Martin, president of NYSE Group. “The NYSE processes more than 1.1 trillion messages per day, and in collaboration with Redpanda and HPE, using NVIDIA Vera CPUs, we will be scaling our capacity while further optimising latency to power a high-performance, resilient and AI-ready market infrastructure.”

System builders and cloud providers prepare Vera deployments

NVIDIA said Vera CPUs will be available in dense, liquid-cooled racks for large-scale agentic AI and reinforcement learning environments, as well as two-socket air-cooled systems for enterprise, cloud, data processing and AI factory deployments.

Infrastructure providers offering Vera CPU-based systems include Aivres, ASRock Rack, ASUS, Compal, Dell, Foxconn, GIGABYTE, HPE, Hyve Solutions, Inventec, Lenovo, MiTAC Computing, MSI, Pegatron, Quanta Cloud Technology, Supermicro, Wistron and Wiwynn.

Dell Technologies, HPE, Lenovo and Supermicro will offer Vera in standalone CPU server configurations. NVIDIA described this as the first standard CPU option beyond x86.

Cloud service providers planning to deploy Vera CPUs include Akamai, ByteDance, Cloudflare, CoreWeave, Crusoe, Lambda, Nebius, Nscale, Oracle Cloud Infrastructure, Redpanda, Starburst, Together AI and Vultr.

Vera systems will be available from system builders and cloud partners starting this autumn.