Your Data ONTAP: NetApp & WWT's One Foundation for AI, Ransomware Resilience, and Hybrid Cloud

"Every proposal on my desk made sense on its own. Security needed better ransomware recovery. The AI team needed faster storage. Finance needed to get the cloud bill under control. What nobody could show me was how the three fit together, or why we were about to pay to store, protect, and move the same data three different ways." 

The executive isn't describing a planning failure. Each team did exactly what it was asked to do, and each was asked a different question. 

Operationalizing AI has overtaken cybersecurity as CIOs' top priority, according to the Gartner C-level Communities 2026 Leadership Perspective Survey. Ransomware remains the top cyber risk concern for CISOs, according to the World Economic Forum's Global Cybersecurity Outlook 2026. Finance, meanwhile, is watching the cloud bill. No one was asked to look at what the three had in common: the same data. 

That shared data is where the problem lives. AI readiness, ransomware resilience, and cloud cost are symptoms of one condition: data that is poorly governed, unevenly protected, and scattered across silos. Solve that once, at the data layer, and the fix pays off three times.  

That is why WWT and NetApp organize our work around three pillars: AI, Security, and Hybrid Cloud. Each pillar addresses one of those symptoms, and all three rest on the same NetApp ONTAP foundation. ONTAP is the data management software that runs underneath NetApp storage and inside the major public clouds, so the same capabilities follow your data wherever it goes. Because every environment is different, we validate that approach on real workloads before you make a commitment.  

Same Data, Three Different Pressures 

The data at the center of all three is mostly unstructured: contracts, drawings, research files, images, recordings, and logs that sit in file shares and object stores rather than in the neat rows of a database. It is often the most valuable information an organization owns. It is also the hardest to keep track of, and each leader feels the strain differently. 

• The CISO's focus is recovery. If we are hit tomorrow, is our last good copy clean, and can we prove it? In a 2024 Sophos study, organizations whose backups were compromised during an attack paid a median of $3 million to recover. That is eight times what organizations with intact backups paid.
• The Chief Data Officer or head of AI has a pilot that works. It can't reach production until governance teams know exactly what is in the data and who can see it. In a 2025 S&P Global Market Intelligence survey, 42% of organizations said they had abandoned most of their AI initiatives, up from 17% a year earlier.
• The CIO and CFO are probably watching the cloud storage bill climb, partly because different teams keep copying the same data into different clouds. Respondents to the Flexera 2026 State of the Cloud Report estimated that 29% of their cloud spending is wasted or unnecessary.

Each of those pressures traces back to the same data. Data that AI can't find is usually data that security can't see. Data nobody can see tends to get copied instead of managed. Every copy adds to the bill and creates one more thing to protect. Left unaddressed, the organization pays to store, protect, and move the same data three different ways, which is exactly the bill the executive saw coming. 

Where Your Data Sits Shapes Everything  

Breaking that cycle starts with the platform that holds the data. That platform matters more than most organizations realize, because of a principle NetApp has long designed around: data gravity. The idea is simple. Large volumes of data are slow, costly, and risky to move, so applications, analytics, and AI tend to go to the data rather than the other way around. The bigger the dataset, the stronger the pull. That makes the platform under your data one of your most consequential technology decisions. Yet it usually comes last. By then, the CISO has talked to security vendors, the AI leader to GPU providers, and the infrastructure team to cloud providers. 

Choosing the platform last is costly, because the three problems are tied together. AI can't leave the pilot stage until someone can vouch for the data it learns from. Security tools can't protect what they can't see. And if your protections change every time data moves to a cloud, you don't have one set of rules. You have several, each with its own gaps. 

Every storage refresh, edge build, or cloud migration is a chance to get the foundation right. Before you approve the next project, ask your team three questions. 

1. Is the platform AI-ready? Can AI tools use the data where it already lives, with the same access rules in place? Or will someone have to copy it somewhere new first?
2. Does the platform protect the data itself? Will the platform spot an attack while it's happening? And afterward, will we be able to prove our recovery copy is clean?
3. Does it work everywhere? Will the same protections, policies, and team skills carry over to our data center, our edge sites, and every cloud we use?

A "no" to any of these isn't a failure. It's a gap, and gaps have a way of turning into separate projects with separate budgets. 

NetApp ONTAP was built to answer yes to all three questions, and each question maps to one of the three pillars. 

Pillar One: AI Needs Trusted Data, Delivered at the Speed GPUs Demand 

AI readiness is a property of the data, not the model. Yet most AI programs choose the model first, then the GPUs, and the storage last. When storage comes last, so do the hard questions about the data itself. In early 2025, Gartner predicted that through 2026, organizations would abandon 60% of AI projects that aren't supported by AI-ready data. 

One example from WWT's AI practice shows how that plays out. A customer moving quickly on an AI deployment chose its storage at the very end. Soon after, the team realized they needed multi-tenancy, which lets several teams share one system while keeping their data strictly separated. The storage they had chosen didn't support it. That single gap changed how they could place guardrails around the data, and the deployment had to be reshaped around it. 

The lesson goes beyond one missing feature. When AI is treated as a separate project, it gets its own storage, its own tools, and its own team to learn them. That's how new silos form. When AI is treated as a data problem, the first question changes from "which model?" to "can the platform we already run serve this data to AI safely?" For organizations running ONTAP, the answer is often yes.  

Multi-tenancy, the capability that customer was missing, is already there, along with snapshots and ransomware protection. Because those safeguards are in place, the AI project doesn't need a separate system that security and governance teams must vet from scratch. The pilot can move into everyday use on a platform the organization already trusts, without creating a new silo to staff and secure. 

"Safely" is the word security leaders focus on. In a 2026 Gigamon survey, 80% of more than 300 CISOs said weak governance of unsanctioned AI use (employees using AI tools the company hasn't approved) is their top challenge in securing data. At its core, that is a worry about not knowing what data AI can reach. Governance is one of three things AI-ready data needs: 

• Available. AI can reach the data where it lives, at the speed GPUs demand, without first copying it somewhere new.
• Consistent. Every team and every model works from the same trusted version of the data, with the same policies attached, on-premises and in the cloud.
• Secure and governed. Govern what feeds the model. Classification and tenant-level guardrails keep sensitive data out of retrieval-augmented generation (RAG) pipelines, the systems that feed company documents to an AI assistant.

Sometimes the platform you already run can't carry the load, and the organization needs new capacity built for AI. NetApp's AI systems scale from a single department up to gigawatt-scale AI factories, the massive GPU clusters run by cloud providers and the largest organizations. 

Capacity is only half of AI readiness. The other half is knowing what your data is and whether AI should touch it. That is the job of the NetApp AI Data Engine (AIDE). It builds a searchable catalog of your data across NetApp and other vendors' storage, then prepares it for AI where it already lives. AIDE follows the NVIDIA AI Data Platform model, which breaks AI readiness into four jobs: discovering data, managing it, securing and governing it, and transforming it into the metadata, embeddings, and vectors AI can use. 

In business terms: RAG lets an AI assistant answer questions from your company's own documents. Vectorization turns those documents into numerical fingerprints that capture meaning, so the assistant can find the right passage even when the exact words never appear. Both depend on storage that can find, serve, and protect data quickly. And because GPUs are among the most expensive assets in the data center, every minute they spend waiting on storage is investment sitting idle. 

Data valuable enough to train an AI model is also valuable enough to steal or hold for ransom. So, AI data has to be protected and recoverable, too. That brings us to the second pillar. 

Pillar Two: Security Belongs Where the Data Lives 

Detecting an attack matters. Recovering from one is what decides how long the business is down. Both depend on knowing where attackers aim. Increasingly, ransomware goes straight for unstructured data in file shares and object stores. Endpoint and network tools remain essential, but they aren't built to see thousands of files on a share being rewritten in a pattern that looks like encryption. The storage system is. 

That is why NetApp builds protection into ONTAP itself, rather than beside it. Many protection tools check data on a schedule, after it has been written. ONTAP watches the data as it changes. The difference comes down to timing. 

ONTAP organizes that protection around three jobs: 

Protect. Classification shows what data you hold and whether it is protected and in the right place. Encryption protects data at rest and in transit. For data in transit, that includes post-quantum cryptography (encryption designed to hold up against future quantum computers). Multi-admin verification requires a second administrator to approve risky actions, such as deleting snapshots. 

Detect and respond. Autonomous Ransomware Protection (ARP) uses machine learning to spot the randomness that encryption leaves behind, then takes a snapshot and alerts administrators. It is on by default for new volumes on current systems, and SE Labs gave its AI-powered version a AAA rating in independent testing. Workload Security watches user behavior and flags an account that suddenly opens files it has never touched. 

Recover. Immutable snapshots and SnapLock create copies no one can change or delete, not even an administrator. NetApp Cybervault keeps an isolated recovery copy. 

That last job answers the question that matters most after an attack: is my last good copy actually clean? Backups taken after an attacker got in carry the infection with them. In the Sophos study cited earlier, 94% of organizations hit by ransomware said attackers tried to compromise their backups. 

These capabilities are strongest together. NetApp Ransomware Resilience, a service in NetApp Console, connects ARP's detection to SnapLock and Cybervault copies. Its Clean Restore feature checks data before it goes back into everyday use. On eligible systems, NetApp backs that recovery with a ransomware recovery guarantee. 

For organizations with the most demanding requirements, there is an outside benchmark, too. ONTAP has been validated under the National Security Agency's (NSA's) Commercial Solutions for Classified (CSfC) program, a process that included a line-by-line review of its code. Most organizations will never store classified files. The validation still matters to them, because it shows how far ONTAP's encryption and controls have been independently tested. 

In business terms: After an attack, leadership wants answers to two questions. How long will we be down? And when we restore, will we bring the attacker back in with the data? A recovery copy you can prove is clean answers both. Teams don't lose days testing one backup after another to find a safe one, so systems come back sooner. And because the copy has been checked, you avoid restoring the infection along with your files. That proof matters beyond IT, too. When the board, regulators, or the cyber insurer ask whether the business can recover, the CISO can answer with evidence instead of reassurance. 

But protection only counts if it travels with the data. And today, data goes a lot of places.

Pillar Three: Hybrid Cloud, One Data Fabric Everywhere  

Follow a single dataset for a week and the problem comes into focus. It starts in the data center. A copy goes to a public cloud for analytics. Another lands at an edge site, where a team needs it close at hand. Each stop often brings its own storage, its own protection settings, and its own management console. The data is the same everywhere. The rules around it are not. That is how policies drift and how cloud bills grow in ways no one planned. 

ONTAP runs on NetApp systems in the data center and at the edge. It also runs as a first-party native service in all three of the largest public clouds: Amazon FSx for NetApp ONTAP, Azure NetApp Files, and Google Cloud NetApp Volumes. First-party native means the cloud provider sells, bills, and supports it as its own service. That is what keeps the same policies and protection in place as data moves from the data center into whichever cloud needs it. NetApp calls the result a data fabric: one consistent way to store, protect, and manage data, wherever it sits. Four capabilities make that fabric practical: 

• FlexCache, data where it's needed: keeps the active part of a dataset close to the people using it, at another site or in a cloud, without a full copy.
• FabricPool, cold data costs less: files no one has opened in months move automatically to lower-cost storage and come back when needed.
• NetApp Keystone, pay as you grow: storage as a service that turns large capital purchases into a predictable operating expense.
• NetApp Console, one place to manage it all: a single control plane for on-premises and cloud environments.

Each pillar works because of the same thing underneath. 

The Connective Tissue: One ONTAP 

The foundation has a name, and the name has a story. As NetApp tells it, one of its founders wanted getting to your data to be as easy as getting a beer. Walk up to any tap. Pour the one you want. No fuss. For years the software was called Data ONTAP: data, on tap. 

In the 1990s, that sounded like a nice idea. Today, with data spread across data centers, edge sites, and more than one cloud, it's closer to a requirement. 

It also explains the payoff. Because all three pillars run on the same software, a decision made for one pays off in the other two. The easiest way to see it is in a single organization. 

Picture a financial services firm with years of research reports and client documents on an on-premises ONTAP system. 

• Protected. ARP watches the files as they change, and scheduled immutable snapshots give the security team a recovery copy it can trust.
• Mobile. An analytics team working in AWS needs the same files. FlexCache serves them in the cloud without a full copy, while the original stays under the same snapshot schedule and ransomware protection.
• AI-ready. The AI team wants a research assistant for analysts. AIDE catalogs and vectorizes the documents in place, with the original access controls intact, so the assistant cannot surface files an analyst isn't allowed to see.

Notice what didn't happen. No second storage platform was purchased for AI. No new security tool was bolted onto a cloud copy. No team had to learn three different consoles. One investment, one policy model, and one skill set produced three outcomes. 

In business terms: Learn once, use everywhere. Every additional platform brings its own licenses, training, patch cycle, and specialists. One platform lowers operating cost in ways that rarely show up in a price comparison. 

Remember the three proposals on that executive's desk? They were never three separate problems, and the next budget cycle doesn't have to treat them that way. It can be one conversation about one foundation, starting with evidence instead of promises. 

See It Work First: WWT's AI Proving Ground 

The smartest way to make an infrastructure decision is to see it work first. WWT's AI Proving Ground makes that possible. This production-grade lab sits inside WWT's Advanced Technology Center (ATC), which is backed by more than $1 billion in infrastructure and more than 200 OEMs. Here, teams get hands-on access to a live, eight-node NetApp AFX cluster and a dedicated AIPod Mini lab for departmental RAG. Both run alongside NVIDIA, Intel, and Cisco systems, so customers can test the full stack they plan to deploy, on real workloads, before it ever reaches production. 

Behind the lab is a partnership NetApp has recognized many times over. WWT has been named NetApp North America Partner of the Year seven times, and NetApp AI Innovation Partner of the Year in both 2025 and 2026.

Authors

Brian Bartell, Practice Manager, Primary Storage and Compute, WWT 

Steve Blanco, National Technical Partner Manager, NetApp 

Dale Darby, Technical Solutions Architect, WWT 

Ming Fang, National Channel Partner Manager, NetApp 

John Lochausen, Technical Solutions Architect III, WWT