Cloud and Data Center Solutions

NVIDIA CPU Platforms

Built for the AI factory.

Stay Informed

Introduction
Benefits
Products
Resources
Next Steps

Introduction
Benefits
Products
Resources
Next Steps

Stay Informed

Introduction

The Engine of the Next-Generation Data Center

NVIDIA CPU platforms provide the foundation for modern AI infrastructure and AI factories. NVIDIA Vera is purpose-built for agentic AI and reinforcement learning, running the tool calls, sandboxed code execution, data pipelines, analytics, and orchestration that keep large-scale AI factories moving. NVIDIA Grace™ extends NVIDIA’s CPU platform across accelerated computing, hyperscale cloud, edge, and telco environments—together delivering the CPU foundation for AI factories.

NVIDIA Vera CPU

NVIDIA Vera is purpose-built for agentic AI and the CPU-intensive work of the AI factory—tool calls, sandboxed code execution, orchestration, analytics, cloud services, and storage. Vera pairs with NVIDIA GPUs as the host CPU for accelerated systems or runs independently across the data center to keep AI pipelines moving.

Learn More

Benefits

Technological Breakthroughs

Accelerate CPU-to-GPU Connections With NVIDIA NVLink-C2C

NVIDIA Vera uses the second-generation NVIDIA NVLink™ Chip-to-Chip (C2C) to connect CPUs and GPUs with 1.8 terabytes per second (TB/s) of bidirectional bandwidth, 7x faster than PCIe Gen 6, enabling efficient data transfer and reducing memory management complexity.

Tap Into Energy-Efficient CPU Memory With LPDDR5X

NVIDIA CPUs use LPDDR5X memory with error-correction code, combining server-class reliability with energy-efficient data movement. In NVIDIA Vera, LPDDR5X delivers up to 1.2 terabytes per second of memory bandwidth, helping feed agentic AI, analytics, and orchestration workloads while reducing memory power across AI factory infrastructure.

Boost Performance and Efficiency With Advanced CPU Architecture

NVIDIA Vera features custom NVIDIA Olympus Arm®-compatible CPU cores and second-generation NVIDIA Scalable Coherency Fabric (SCF) to deliver high single-thread performance, massive data movement, and predictable throughput under load. NVIDIA Grace uses high-performance, Arm-based CPU cores and NVIDIA SCF to deliver efficient performance for accelerated computing, cloud, edge, and data center workloads.

Diverse Configurations for Accelerated and CPU Workloads

NVIDIA CPU platforms support flexible AI factory designs, from tightly integrated CPU-GPU architectures like NVIDIA Vera Rubin and NVIDIA Grace Blackwell to standalone single- and dual-socket CPU servers. Vera extends the platform across agentic AI, reinforcement learning, analytics, orchestration, and storage, while Grace provides the proven CPU foundation for accelerated computing and power-efficient computing.

Products

Explore the CPU Lineup

NVIDIA Grace CPU

Breakthrough CPU performance and efficiency for the modern data center.

NVIDIA Grace CPU Superchip

The Grace CPU Superchip is composed of two Grace CPUs connected coherently over NVIDIA NVLink-C2C at 900 GB/s. It packs 144 Arm Neoverse V2 cores into a single module, with server-class LPDDR5X memory that delivers up to 1 terabyte per second (TB/s) of memory bandwidth.

Learn More About the Grace CPU Superchip

NVIDIA Grace CPU C1

The NVIDIA Grace CPU C1 is a single-socket, high-performance server platform that doesn't compromise performance or bandwidth. It’s optimized for scalable and edge platforms, including hyperscale cloud, content delivery network (CDN), storage, telco, and other high-performance edge platforms.

Read the Grace CPU Superchip’s Datasheet

NVIDIA CPU and GPU

Connecting the NVIDIA CPU and GPU for large-scale AI and HPC applications.

NVIDIA GH200 Grace Hopper Superchip

The NVIDIA Grace Hopper™ Superchip combines Grace CPU and Hopper GPU architectures using NVLink-C2C to enable a unified memory space for accelerated AI and HPC workloads.

Learn More About the Grace Hopper Superchip

NVIDIA GB200 NVL4

NVIDIA GB200 NVL4 unlocks the future of converged HPC and AI, delivering revolutionary performance through four NVIDIA Blackwell GPUs connected with an NVLink bridge and two NVIDIA Grace CPUs connected over NVLink-C2C. Compatible with liquid-cooled, NVIDIA MGX™ modular servers, it boosts performance for scientific computing, AI for science training, and inference applications by 2x over the prior generation.

Read the GB200 NVL4’s Datasheet