NVIDIA Vera Rubin Platform
AI infrastructure for the era of agents.
Overview
Driving the Era of Agentic AI
The NVIDIA Vera Rubin platform is built for the age of agentic AI and reasoning, engineered to master multi-step problem-solving and massive long-context workflows at scale. Vera Rubin is a multi-rack POD-scale system that brings together five purpose-built rack-scale systems into one massive, coherent AI supercomputer. By eliminating critical bottlenecks in communication and memory movement, the platform supercharges inference, delivering more tokens per watt and lower cost per token compared to the NVIDIA Blackwell architecture.
Look Inside the Vera Rubin Platform
NVIDIA Vera Rubin NVL72
NVIDIA Vera Rubin NVL72 unifies leading-edge technologies from NVIDIA—72 Rubin GPUs, 36 Vera CPUs, ConnectX™-9 SuperNIC™s, and BlueField™-4 DPUs. It scales up intelligence in a third-generation rack-scale platform with the NVIDIA NVLink™ 6 switch and scales out with NVIDIA Quantum-X800 InfiniBand and Spectrum-X™ Ethernet to power the AI industrial revolution at scale.
Vera Rubin NVL72 features a new Transformer Engine with adaptive compression to boost NVFP4 inference performance, third-generation NVIDIA Confidential Computing that extends security across the full rack-scale platform, and a second-generation RAS engine that delivers rack-scale resiliency.
NVIDIA Vera CPU
The NVIDIA Vera CPU rack delivers dense, liquid-cooled CPU infrastructure purpose-built for reinforcement learning and agentic AI at scale. Built on the NVIDIA MGX™ modular reference architecture, each rack integrates 256 NVIDIA Vera CPUs and supports more than 22,500 concurrent sandbox environments, giving AI factories scalable, energy-efficient CPU capacity for tool calls, evaluation, data processing, and orchestration.
NVIDIA Groq 3 LPX
NVIDIA Groq 3 LPX is the inference accelerator for NVIDIA Vera Rubin, designed to meet the low-latency and large-context demands of agentic systems. By combining Rubin GPUs for high-bandwidth memory (HBM) and LPUs for static random-access memory (SRAM), NVIDIA Vera Rubin with LPX delivers a new class of inference performance for trillion-parameter models and million-token contexts.
NVIDIA Vera BlueField-4 STX
NVIDIA Vera BlueField-4 STX is a modular foundation for rack-scale AI-native storage solutions. By integrating NVIDIA Vera Rubin, BlueField-4 STX storage processor, Spectrum-X networking, and NVIDIA AI software, it optimizes the entire data lifecycle from data analytics to model training and full agentic AI workflows at scale.
NVIDIA Spectrum-6 SPX Ethernet
Spectrum-6 SPX Ethernet is engineered to accelerate networking across AI factories. Configurable with either NVIDIA Spectrum-X™ Ethernet or NVIDIA Quantum-X800 InfiniBand switches, it delivers low-latency, high-throughput rack-to-rack connectivity at scale.
Explore NVIDIA Vera Rubin Products
Inside the NVIDIA Vera Rubin Platform
Read this technical deep dive to learn how NVIDIA Vera Rubin treats the data center as the unit of compute, not the chip, establishing a new foundation for producing intelligence efficiently, securely, and predictably at scale.