Making Data AI-Ready: 13 Storage Vendors Bring Latest Tech To Nvidia GTC

Access to data that has been prepared for use in AI training, inference and other tasks is key to ensuring that storage systems don’t act as bottlenecks to full performance of GPUs and AI applications. CRN looks at 13 storage vendors that brought their best AI offerings to Nvidia’s GTC conference.

GPUs are the devices on which AI runs, and the importance of increased GPU performance was a key theme of this year’s Nvidia GTC 2026 event.

But no matter how much performance those GPUs offer, they come to a standstill while waiting for data. That has elevated the availability and performance of data to the same level of importance as GPUs.

To ensure data does not serve as the bottleneck for GPUs and the AI applications running on them, storage vendors are approaching the AI industry on multiple fronts. These include increased performance of storage systems to feed data to the GPUs and improved management of data to ensure it is ready and in a form the GPUs can use.

[Related: Analysis: Nvidia’s AI Dominance Expands To Networking As It Makes Bigger CPU Push]

Several of the IT industry’s top data storage and data management companies were at Nvidia GTC 2026 showing their latest entries for AI. This included new storage systems from companies like Dell Technologies, NetApp, Everpure, Hammerspace, DDN, Vast Data and Hitachi Vantara, along with new management schemes from several other vendors.

CRN is highlighting new AI-focused storage technologies from 13 top storage vendors. Read on for all the details.

ScaleFlux, FarmGPU and Lightbits Labs took to GTC to unveil a collaborative architecture they said was designed to overcome the memory and I/O limitations of long-context AI inference. By integrating ScaleFlux’s high-performance NVMe storage, FarmGPU’s managed environment and Lightbits’ LightInferra software, this new platform optimizes KV-cache persistence and streaming. The companies said the architecture eliminates GPU stalls by prefetching essential data over high-speed RDMA, significantly reducing Time-to-First-Token and increasing GPU utilization by up to 3X. Featuring AI-native security and tenant isolation, the platform enables enterprises to serve larger models and longer conversations more efficiently, delivering a responsive, scalable and cost-effective infrastructure for next-generation AI workload

Hitachi iQ from Hitachi Vantara is an enterprise AI infrastructure platform that integrates compute, networking and storage into a validated infrastructure stack. Built on Hitachi Vantara’s VSP One data platform, Hitachi iQ now supports new Nvidia Blackwell and Blackwell Ultra GPUs across air-cooled and liquid-cooled configurations, targeting workloads ranging from model development and fine-tuning to inference and agentic applications. Hitachi iQ Studio, the AI software component, provides tools for deploying and managing AI agents within secure enterprise environments, including new blueprints for multi-agent orchestration with supervisor and worker roles. Its expanded capabilities with Hammerspace enable distributed data access via MCP, reducing the need to move data between environments.

At GTC 2026, Vdura unveiled three updates to its GPU-native AI storage platform: immediate availability of RDMA; a preview of Context-Aware Tiering; and optimized infrastructure built on AMD EPYC Turin processors and Nvidia ConnectX-7 networking. RDMA enables direct GPU-to-storage data transfers that bypass the CPU, reducing latency for AI training and inference workloads. Context-Aware Tiering, planned for later in 2026, will intelligently move data across storage tiers, including local NVMe SSD, DRAM and durable storage, based on workload patterns. RDMA is currently available on V5000 and V7000-class systems.

Everpure Evergreen//One for FlashBlade//EXA

Evergreen//One (EG1) for FlashBlade//EXA from Everpure, formerly known as Pure Storage, is a flexible, consumption-based storage service designed for large-scale AI training and inference. It replaces traditional hardware purchasing with a subscription model, aligning costs with usage and enabling seamless growth as AI workloads and data complexity evolve. Everpure said that as enterprises historically struggle with a “performance collapse” when scaling AI from a few nodes to hundreds, this service provides the high-throughput and linear scalability required for massive AI workloads from a single dashboard. By delivering a unified subscription for global deployment, Everpure aims to move the market from rigid hardware silos toward flexible, global AI factories that scale on demand.

DDN Horizon And AI Factory Experience

DDN Horizon from DDN is a full-stack AI orchestration control plane designed to operationalize GPU and data infrastructure as AI-as-a-Service platforms. The software orchestrates compute, storage and AI pipelines across on-premises, hybrid and cloud environments to help organizations provision AI workspaces, training environments and inference services through a self-service model. The company also showed its AI Factory Experience under which DDN, Supermicro and Nvidia collaborated on a fully operational, turnkey AI infrastructure stack. The environment, hosted inside a custom-built platform, runs live enterprise AI pipelines including retrieval-augmented generation, financial risk analytics, genomics workflows and video analytics on DDN Enterprise AI HyperPOD systems powered by Nidia GPUs.

Nutanix took to GTC to unveil its Agentic AI offering, a full software stack purpose-built to help customers accelerate adoption of Agentic AI for business transformation. The offering integrates with Nvidia AI Enterprise at the Agent Builder layer and orchestrates the Nvidia-certified ecosystem of AI factories. It aims to deliver performance, compliance and security for Agentic AI applications and help minimize aggregate token costs. It also empowers enterprise infrastructure and platform teams to simply build, scale and operate AI factories and empowers developer teams with a rich set of AI PaaS services integrated with Nvidia AI Enterprise to accelerate deployment of Agentic AI workloads, the company said.

Vast Data Vast Foundation Stacks

The new Vast Foundation Stacks from Vast Data are an open-source library that extends Nvidia Blueprints into production-ready AI pipeline implementations designed to run natively on the Vast AI Operating System. By packaging proven architectural patterns into deployable templates, Foundation Stacks helps organizations move from experimentation to production faster. Each stack integrates data access, database services, compute orchestration, eventing and pipeline execution within a unified environment to help eliminate the need to assemble complex infrastructure. This also helps developers focus on the business logic connecting AI to enterprise data and workflows while consistently deploying scalable pipelines across cloud and on-premises environments where the Vast AI OS run