Dell & NVIDIA boost AI data orchestration & storage

Dell Technologies has expanded its Dell AI Data Platform with NVIDIA, adding software and infrastructure focused on data preparation, orchestration, and storage performance for AI workloads and autonomous agents.

The update has two main parts. New data engines and orchestration features prepare structured and unstructured information for AI use, while new storage software and performance options target large-scale training and inference.

Data lifecycle

A key addition is the Dell Data Orchestration Engine, built on technology from Dell's Dataloop acquisition. It uses no-code and low-code tools to run workflows across the AI data lifecycle, covering discovery, labelling, enrichment, and transformation into governed datasets.

The engine also supports active learning and human-in-the-loop workflows, enabling organisations to improve dataset quality while maintaining governance and control.

Dell has also introduced a marketplace for the orchestration engine, with a curated library spanning NVIDIA NIM microservices, NVIDIA AI Blueprints, and more than 200 other models, applications, and templates. The marketplace is intended to provide production-ready data workflows for deployment.

Alongside orchestration, Dell added an AI Assistant within the Dell Data Analytics Engine. The assistant provides a conversational natural-language interface inside SQL analytics, allowing business users to query, visualise, and collaborate on governed data products without specialist SQL skills.

NVIDIA blueprints

The platform now supports infrastructure for NVIDIA's AI-Q 2.0 blueprint, which Dell said provides a path to building custom AI agents that operate across enterprise data.

It also includes NVIDIA-accelerated integrations for data preparation, retrieval, and reasoning pipelines across structured and unstructured information. Dell said customers gain access to a growing library of NVIDIA blueprints and NIM microservices.

Dell also said the NVIDIA Nemotron 3 Super model is available through Dell Enterprise Hub on Hugging Face.

Separately, Dell said it will support NVIDIA STX, a modular reference design based on the next-generation NVIDIA Vera Rubin NVL72 platform, NVIDIA BlueField-4 DPUs, and NVIDIA Spectrum-X Ethernet networking.

GPU acceleration

Within the data platform layer, Dell said it will introduce NVIDIA RTX PRO Blackwell Server Edition GPUs, bringing acceleration closer to the data platform instead of reserving it for training and inference.

Dell also pointed to NVIDIA CUDA-X libraries in the stack, including NVIDIA cuDF for structured data processing and NVIDIA cuVS for vector indexing and search for unstructured data. It cited internal testing showing up to 3x faster SQL queries and up to 12x faster vector indexing under certain configurations and comparisons.

Storage scale

Dell is also expanding its storage software for AI environments where GPU clusters can be constrained by data throughput and access patterns. It introduced Dell Lightning File System, a parallel file system designed for AI training and inference.

Dell said Lightning File System can deliver up to 150GB per second per rack unit and up to 20x higher performance than flash-only scale-out file competitors in some scenarios. It also cited up to 2x higher throughput per rack unit than competing parallel file systems. Dell said the file system integrates with NVIDIA-based AI infrastructure.

Another addition is Dell Exascale Storage, a "3-in-1" approach for extreme-scale AI and high-performance computing. Dell said IT teams can deploy file, object, and parallel file system storage software on Dell PowerEdge servers.

According to Dell, Exascale Storage supports Dell PowerScale, Dell ObjectScale, and Dell Lightning File System on a shared hardware platform. It also supports NVIDIA CX-8 and CX-9 SuperNICs, with network connectivity planned up to 800GbE. Dell cited internal analysis showing read performance up to 6TB per second per rack.

Dell also highlighted support for the NVIDIA CMX context memory storage platform and inference acceleration with KV cache on shared storage across PowerScale, ObjectScale, and Lightning File System. This approach is designed to let organisations move KV cache from GPU memory to shared storage, depending on performance requirements.

In addition, Dell pointed to PowerScale performance testing for its software-driven Parallel Network File System architecture. Dell said internal testing shows up to 6x faster performance with large files in enterprise AI environments compared with NFSv3 under certain test conditions.

Market context

The announcement comes as large organisations try to move AI projects from pilots to production. Fragmented data estates, governance constraints, and infrastructure bottlenecks are frequently cited as barriers to deploying generative AI systems and agent-style applications.

Dell linked the updates to its broader Dell AI Factory with NVIDIA portfolio. It said more than 4,000 customers are deploying the Dell AI Factory, and cited an Enterprise Strategy Group paper commissioned by Dell reporting early adopters seeing up to 2.6x ROI within the first year under a modelled approach.

Travis Vigil, Senior Vice President, ISG Product Management, Dell Technologies, said the industry's core challenge is data readiness rather than model selection.

"The number one problem enterprises face when moving AI pilots to production is curating the data they already have and putting it to work. The Dell AI Data Platform with NVIDIA automates the entire data lifecycle and delivers the speed and scale AI workloads demand. We've done the integration work, so customers deploy faster, scale with confidence and see real returns. Together with NVIDIA, we're defining what enterprise AI infrastructure needs to be."

Jason Hardy, Vice President, Storage Technologies, NVIDIA, said agentic AI changes infrastructure priorities for enterprise deployments.

"The shift to autonomous agents requires a fundamentally different approach to data infrastructure, with automated orchestration, AI-native storage and GPU-optimised performance architected to work together. Dell's enterprise expertise, combined with full-stack NVIDIA AI infrastructure, creates the foundation organisations need to deploy AI at scale," Hardy said.