NVIDIA-Certified Storage

Type: Program Tags: NVIDIA, certified storage, AI factory, enterprise storage, GPUDirect Storage, AI Data Platform, storage partners Related: NVIDIA-Certified-Systems, NVIDIA-Enterprise-AI-Factory, NVIDIA-Enterprise-Reference-Architectures, Red-Hat-AI-Factory-with-NVIDIA, NVIDIA-RTX-PRO-AI-Factory, NVIDIA-HGX-AI-Factory, NVIDIA-NVL72-AI-Factory, NVIDIA-AI-Data-Platform, NVIDIA-STX, GPU-Direct-Storage, NVIDIA-DGX-SuperPOD, NVIDIA-DGX-Cloud Sources: https://www.nvidia.com/en-us/data-center/products/certified-storage/, https://docs.nvidia.com/ai-enterprise/planning-resource/ai-factory-white-paper/latest/ecosystem-architecture.html, https://docs.nvidia.com/ai-enterprise/deployment/red-hat-ai-factory/latest/prerequisites.html Last Updated: 2026-04-29

Summary

NVIDIA-Certified Storage is NVIDIA’s validation program for enterprise storage platforms that need to feed AI, HPC, RAG, KV-cache, and agentic AI workloads without becoming a bottleneck. It complements NVIDIA-Certified-Systems by validating the storage layer of AI factories and NVIDIA AI Data Platform deployments.

Detail

Purpose

AI factories need storage that can sustain large model training, high-volume inference, fine-tuning, retrieval, vector database access, and context-cache workloads. NVIDIA-Certified Storage gives enterprises and cloud providers a way to select partner storage that has been benchmarked and validated for performance, reliability, scalability, quality of service, security, and multi-tenancy.

Certification scope

  • General-purpose file and object storage for AI and accelerated computing workloads.
  • Scale-out performance across training, inference, fine-tuning, and KV-cache read/write patterns.
  • Reliability, availability, serviceability, multitenancy, security, and data services.
  • Special-purpose system certification for NVIDIA-AI-Data-Platform deployments, including KV-cache and document extraction validation.

AI factory context

The Enterprise AI Factory design guide treats certified storage as a core part of AI infrastructure. It helps feed compute nodes, support vector databases and knowledge bases, and protect large datasets and model artifacts. It is also related to GPU-Direct-Storage where partner systems use direct storage-to-GPU paths for throughput-sensitive workloads.

The Red-Hat-AI-Factory-with-NVIDIA prerequisites recommend NVIDIA-Certified Storage when OpenShift AI deployments need supported dynamic storage classes, object storage, model cache storage, and production data services.

Connections

Source Excerpts

  • NVIDIA’s certified storage page describes validated partner storage for performance, security, scale, and AI factory workloads.
  • The Enterprise AI Factory design guide says storage can become a bottleneck and calls out NVIDIA-Certified Storage as a key design component.
  • The Red Hat AI Factory prerequisites recommend NVIDIA Certified Storage for OpenShift AI deployment storage.

Resources