NetApp INSIGHT 2026: Built for the Agentic Era

AI has changed the mission.

Generating answers is no longer enough. AI now takes action. Agents operate at machine speed. Every decision depends on data.

If that data is fragmented, outdated, ungoverned, or unavailable, the consequences are no longer limited to a bad report. The wrong action can happen instantly—and at scale.

At INSIGHT 2026, NetApp introduced a clear response.

One platform.
Autonomous operations.
Data activated in place.
Infrastructure built for AI-factory scale.

One Platform. One Control Plane.

The NetApp platform brings storage, data services, resilience, and operations together across on-premises environments, public clouds, sovereign clouds, and neoclouds.

The NetApp Console is the front door.

Now, it is becoming more autonomous.

New agentic capabilities bring intent-based control to the Console. Customers can define the outcome. The platform can execute the work within established guardrails.

This is more than another chatbot.

Organizations can bring their own large language models. They can interact through conversational operations or connect agents through an MCP server. Closed-loop automation can provision resources, optimize the environment, and remediate issues—with human oversight where it matters.

Less manual work.
Faster decisions.
Controlled execution.

The Console will also support local deployment. It can run inside a customer’s own data center, including disconnected and air-gapped environments.

The operating model stays consistent. The deployment model becomes a choice.

SaaS or local.
Connected or isolated.
One Console.

Sovereignty Without Compromise

Data sovereignty is now an operational requirement.

NetApp is addressing it with Keystone Sovereign, a new service class designed for organizations that must keep their data, operations, and support within defined geographic and jurisdictional boundaries.

The initial focus is Europe, including Germany and France.

Keystone Sovereign keeps the operating model local. Data stays within the required boundary. Administration and support remain within the approved jurisdiction.

Control is not implied.
It is designed into the service.

Activate Data Where It Lives

Traditional AI pipelines move too much data.

They copy it. Transform it. Duplicate it. Move it again.

That model is slow, expensive, and difficult to govern. It cannot keep pace with agents that need current enterprise data now.

NetApp is advancing a different model: zero-copy data activation.

The goal is direct access to data in place, with discovery, classification, governance, and security built in. Enterprises can use the data they already have without first forcing it through another complex pipeline.

No unnecessary copies.
No stale datasets.
No loss of control.

The result is a governed data foundation for retrieval, analytics, and agentic AI.

Introducing NetApp Novus

The largest AI environments face a different problem.

GPUs consume data at extreme speed. Traditional storage was not designed for tens of thousands—or hundreds of thousands—of them operating together.

Metadata operations surge. Sequential reads expand. Idle GPUs destroy the economics.

NetApp Novus is a new storage architecture built for that scale.

Novus separates metadata operations from data traffic. Each layer can scale independently. Using pNFS and NFS 4.2, GPU nodes can reach data directly with GPU direct connect without requiring a proprietary client or traditional parallel file system.

 It is designed for API-driven, multi-tenant environments where operators may add hundreds of storage nodes or thousands of GPUs at a time.

The target is uncompromising:

• Hundreds of GPUs.

• Multi-exabyte namespaces.

• Massive throughput.

• Native NFS simplicity.

• Core ONTAP capabilities at hardware-line-rate speed.

• Fastest storage on earth with 100TB/s in a single namespace

• Multiple clusters within a single namespace

Novus is not built for ordinary infrastructure.

It is built for AI factories, hyperscalers, neoclouds, and the largest enterprise AI environments in the world.

NetApp and Oracle to Launch Fully Managed Cloud Storage Service for Demanding AI and Enterprise Workloads

Another major INSIGHT 2026 announcement extends the NetApp data platform to Oracle Cloud Infrastructure. NetApp and Oracle plan to introduce Oracle Cloud Infrastructure NetApp Storage Service, a fully managed, OCI-native service that brings enterprise-grade NetApp ONTAP data management capabilities directly to OCI.

The service is designed to help organizations migrate, run, protect, and modernize databases, enterprise applications, virtualized environments, EDA and HPC workloads, regulated applications, and AI data pipelines while preserving familiar storage operations.

Customers will be able to manage the service through the OCI Console and SDKs, ONTAP APIs, and established operational workflows. That native integration is intended to reduce application redesign and storage-management changes when moving existing ONTAP workloads to OCI.

For mission-critical and regulated environments, the planned service combines consistent data management, high availability, security, multiprotocol access, and predictable performance across on-premises and cloud infrastructure.

General availability is planned within the next 12 months. The announcement completes an important hyperscaler chapter for NetApp by adding a native Oracle Cloud service alongside its established offerings with AWS, Microsoft Azure, and Google Cloud.

Source: Oracle press release, September 29, 2026

The New Measure of AI Infrastructure

For years, infrastructure economics focused on cost per terabyte or cost per virtual machine.

AI introduces a new measure: cost per token.

That cost depends on more than compute. It depends on keeping expensive GPUs supplied with trusted data. It depends on removing bottlenecks. It depends on operating at machine speed without surrendering governance.

That is the direction set at INSIGHT 2026.

Autonomous operations.
Sovereign control.
Zero-copy data activation.
AI-factory performance.

AI does not wait for infrastructure.

Now, the infrastructure is AI ready.