The NetApp Platform is built for the Agentic Enterprise

Ask enterprise teams running AI today what keeps them up at night, and the answer has changed. A year ago it was pilots and experiments, proving AI could work, and getting something into production. Today it’s scale: pushing AI into production workloads and moving the needle on efficiency, competitiveness, and speed to market. Agentic AI is accelerating that shift, changing how work gets done, and increasing the pressure on the data and infrastructure underneath it.

Enterprises already faced data challenges. AI has made those challenges impossible to defer.

• The first is scale. AI Factories are pushing traditional storage architectures to a scale they were never designed for. Conventional designs can’t feed GPUs fast enough to keep them busy, and utilization can fall into the single digits. At gigawatt scale, the cost of that idle time runs into the millions.
• The second is activation. Enterprise data is fragmented across on-premises systems, public clouds, and edge sites, governed inconsistently and largely unclassified. Making it usable has meant standing up custom pipelines and an engineering effort for every new project. Engineering becomes the bottleneck, and activation stays operationally complex and costly. Enterprises need a simple, repeatable way to activate data, not work that gets rebuilt each time.
• The third is control. Your infrastructure is borderless across data centers, colos, edge sites, and public clouds, some of it under sovereignty requirements that dictate where data can live. Agents act on that data at machine speed while the controls around it remain human-paced and applied by hand, one environment at a time. An estate that takes days to inventory and audit cannot govern activity that happens in milliseconds.
• The fourth is ROI. Budgets for AI were approved on faith; now leaders need proof. Every GPU cycle, every gigawatt of accelerated computing, every dollar of storage, and every hour of engineering time has to translate into a business result. When scale, activation, and control aren’t solved, compute sits idle and ROI is the first casualty. Bringing acceleration to the data, instead of moving data to the compute, is what turns that equation around.

NetApp is solving for all four challenges. Today at NetApp INSIGHT, we unveil new capabilities built into the NetApp Platform that let enterprises move at the speed agentic AI demands while cutting complexity and lowering costs:

An AI Factory revolution

Gigawatt-scale AI Factories are the result of AI extending into virtually every workload. They’re the next major wave in data infrastructure innovation, and we’re going to lead that wave with NetApp Novus. This industry-first architecture, built on NetApp ONTAP, enables operators to maximize the efficiency, utilization, and profitability of their GPU investments. Built to support a zettabyte-scale file system, Novus is designed to decouple metadata and data to provide 100 TB/s throughput that can keep millions of GPUs productive. Scale beyond boundaries with the fastest storage on the planet.

Read about it

Secure, zero-copy data activation for the agentic enterprise

NetApp AI Data Services makes enterprise data usable where it already lives, instead of requiring teams to build a pipeline to reach it. Organizations can now discover, understand, govern, and operationalize data with secure, in-place, zero-copy data activation. This brings accelerated computing close to where data already lives, with zero code changes required on the customer side. AI Data Services turns dark, unstructured enterprise data into trusted context for analytics, assistants, and AI agents. And it now extends across the full enterprise data estate, including NetApp ONTAP^®, StorageGRID, and non-NetApp storage.

Read about it

Enterprises empowered to build anything, anywhere

We’re extending the NetApp Platform in several ways: with NetApp Console we’re accelerating data estate management through Autonomous Operations inside customer-defined guardrails, and we’re delivering unified visibility and control across environments via Fleet Management. With Keystone Sovereign, we’re adding clearer regional governance controls in sovereignty-sensitive markets. AI ChatOps brings conversational automation to the control plane, so teams describe what they want and the platform works out how to deliver it. Throughout, enterprises get consistent operations and policy that follow the data wherever teams choose to build.

Read about it

AI-ready data, on every cloud you run

The NetApp Platform now extends natively across every major hyperscaler, making NetApp the only platform with that breadth of reach. Whether your AI workloads run on-premises, in a single cloud, or across several, the same data services, governance, and activation capabilities travel with your data. Enterprises don't have to standardize on one cloud to get consistent AI-readiness — or rebuild their data foundation every time they add a new environment.

What’s next

What links these three announcements is a premise I laid out a few weeks ago: the real bottleneck of the AI era isn’t compute, it’s data readiness. The data you need is already yours. It sits in your datacenters, your clouds, your edge sites. Data readiness means that data is always reachable in place, governed and secure, and fast enough to matter in the moment a model or an agent asks for it.

That’s what the NetApp Platform is built to do, from edge to cloud. With native integrations into every major public cloud, the NetApp Platform gives you unmatched flexibility to make all your data AI-ready, wherever it lives. We’re extending those capabilities today — at the scale AI Factories demand, at the speed agents move, and across every environment you run in. Agents are what enterprises are building this year, and something else will follow. A data foundation this capable means you’re ready for anything.

This announcement includes forward-looking statements regarding anticipated product features and functionality. These statements reflect our current plans and are subject to change. Actual product features, functionality and timing may differ from those described, and we undertake no obligation to deliver any specific feature or functionality except as required by applicable law.