# NVIDIA STX

The foundation for AI-native data platforms.

[Read Announcement](https://nvidianews.nvidia.com/news/nvidia-launches-bluefield-4-stx-storage-architecture-with-broad-industry-adoption)

## Overview

## The Next Generation of Accelerated Data Infrastructure

NVIDIA STX is a modular reference architecture for an AI-native data platform specifically engineered to solve architectural challenges. It serves as the critical layer that bridges the gap between massive AI compute and data storage by centralizing intelligent data handling. By leveraging the [NVIDIA Vera Rubin](https://www.nvidia.com/en-us/data-center/technologies/rubin.md) architecture, [BlueField®-4](https://www.nvidia.com/en-us/networking/products/data-processing-unit.md), and [Spectrum-X™ networking](https://www.nvidia.com/en-us/networking/spectrumx.md), STX provides the foundation for partners building a universal data engine that accelerates the full AI lifecycle, from training and analytics to real-time agentic inference.

### NVIDIA BlueField-4 STX Storage Platform Launches With Broad Industry Adoption

NVIDIA STX reference architecture, co-designed with leading storage partners, enables enterprises, cloud and AI providers to easily deploy accelerated storage infrastructure capable of the long-context reasoning required for agentic AI.

[Read Press Release](https://nvidianews.nvidia.com/news/nvidia-launches-bluefield-4-stx-storage-architecture-with-broad-industry-adoption)

### Storage Foundation for AI-Native Data Platforms

Learn how NVIDIA STX redefines storage infrastructure by bringing acceleration to every data layer and eliminating the biggest bottlenecks of agentic AI.

[Read Solution Overview](https://resources.nvidia.com/en-us-ai-storage/stx-solution-overview)

Technology

## What Is NVIDIA STX?

NVIDIA STX is a new modular reference architecture for an AI-native storage and data platform, co-designed with leading partners. Built on NVIDIA accelerated compute, networking, and AI software, NVIDIA STX provides optimized storage and context memory for enterprises and large-scale AI leaders.

NVIDIA STX is co-designed with partners to build and deploy AI storage at scale and accelerate every storage AI workload. STX includes new system designs built on NVIDIA accelerated computing, networking, and AI software.

Features

## Advancing AI-Ready Data Infrastructure for AI Factories

NVIDIA STX serves as the blueprint for high-speed data access and seamless acceleration across the [AI factory](https://www.nvidia.com/en-us/solutions/ai-factories.md).

### Next-Gen, AI-Enabled Data Infrastructure

This reference architecture unifies NVIDIA Vera, Rubin, and BlueField-4 to upgrade traditional storage. It provides a blueprint for storage leaders to build foundational data engines, overcoming legacy limitations to power the future of enterprise intelligence.

### Acceleration at Every Data Layer

This reference architecture bridges hardware and software to boost data throughout the AI lifecycle by bringing compute to the data. NVIDIA STX provides new system designs to accelerate high performance storage, enterprise AI data, and context memory.

### Optimized Data Infrastructure for AI Factories

STX redefines performance by bringing compute directly to the data. By offloading key-value (KV) cache and vector tasks to BlueField-4, it accelerates context storage and indexing. Optimized for retrieval-augmented generation (RAG) and [agentic AI workflows](https://www.nvidia.com/en-us/solutions/ai/agentic-ai.md), STX increases tokens per second and efficiency to sustain massive GPU cluster throughput.

Product Ecosystem

## A Comprehensive Accelerated Data Platform

Powered by data-optimized NVIDIA hardware and software, STX introduces a purpose-built design for AI-ready data.

### NVIDIA Vera CPU

The NVIDIA Vera CPU is purpose-built for agentic reasoning, coordinating data movement, memory, and workflows across accelerated systems.

[Learn About NVIDIA Vera](https://www.nvidia.com/en-us/data-center/vera-cpu.md)

### NVIDIA Rubin Platform

The NVIDIA 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.

[Learn About the NVIDIA Rubin Platform](https://www.nvidia.com/en-us/data-center/technologies/rubin.md)

### NVIDIA BlueField DPUs

NVIDIA BlueField-4 powers the NVIDIA Inference Context Memory Storage Platform, an AI-native data storage infrastructure designed to accelerate and scale gigascale agentic inference.

[Learn About NVIDIA BlueField-4](https://www.nvidia.com/en-us/networking/products/data-processing-unit.md)

### NVIDIA Spectrum-X Ethernet Networking

The world’s first Ethernet networking platform, NVIDIA Spectrum-X is designed for training and deploying gigascale AI.

[Learn About NVIDIA Spectrum-X Ethernet](https://www.nvidia.com/en-us/networking/spectrumx.md)

### NVIDIA CMX Context Memory Storage

This AI-native storage infrastructure uses BlueField to eliminate inference GPU stalls, deliver power efficiency, and enable high-speed KV-sharing to build inference infrastructure scale.

[Learn About NVIDIA CMX](https://www.nvidia.com/en-us/data-center/ai-storage/cmx.md)

### NVIDIA AI Data Platform

The NVIDIA AI Data Platform is a customizable reference design that integrates accelerated computing into enterprise storage.

[Learn About the NVIDIA AI Data Platform](https://www.nvidia.com/en-us/data-center/ai-data-platform.md)

Adopters

## Meet Our Partners

STX is co-designed with leading storage partners to deliver a data platform solution with acceleration at every layer.

## Get Started

### Stay Informed About NVIDIA STX

Learn how STX supports AI-native data infrastructure designed for the unique demands of the age of AI.

[Notify Me](#notify-me)

### Redefining Storage Infrastructure for AI

Discover how NVIDIA STX accelerates every data layer, eliminating the primary bottlenecks associated with agentic AI, to create a robust storage foundation for AI-native data platforms.

[Learn More](https://resources.nvidia.com/en-us-ai-storage/stx-solution-overview)

## Email Me When Available

Welcome back.
Not you? Log Out

Welcome
back. Not you? Clear form