‘We transform NetApp from being your data storage provider to your data platform provider to help you transform your data into knowledge,’ says NetApp CEO George Kurian.
NetApp this week accelerated its push to be at the center of making data available and ready for use with AI.
The San Jose, Calif.-based company, more known for its storage prowess, has over the past couple of years made its mark as a provider of technology and services focused on data intelligence. Nowhere was this more evident than at this week’s NetApp Insight 2025 conference where customers and channel partners were introduced to new ways the company is helping them discover, migrate and transform data for use with AI.
NetApp CEO George Kurian (pictured), during his keynote presentation, made it clear that the opportunity provided by AI is contingent on a company’s data, and that traditional methods of transforming raw data into AI-ready data are highly inefficient and were built for another era.
[Related: NetApp CEO: ‘Building On A Position Of Strength’ With AI]
Kurian also reinforced NetApp’s message that a unified data platform built on a foundation of unified data storage with an intelligent active metadata fabric is key to transforming data into knowledge.
“We stand in a long line of innovators and stewards of data,” he said. “Data is the foundation of knowledge, the knowledge to understand where we’ve come from, to share our stories, our traditions, to be able to understand ourselves and to understand each other, and by doing so, to build a better world, to make ourselves better, to make our teams better, to make our organizations better, to make our communities better and to make the world better.”
NetApp makes sure customer data is transformed into the way customers want to use it.
“In turn, we transform NetApp from being your data storage provider to your data platform provider to help you transform your data into knowledge,” he said.
Sandeep Singh, NetApp’s senior vice president and general manager for enterprise storage, said during a press briefing that customers’ underlying data infrastructure is currently siloed but needs to be unified across on-premises and cloud and across workloads, and that NetApp is helping with that modernization with new technologies for making data ready for AI and then protecting that data.
“We have built the NetApp Data Platform to help [customers] have a data platform that is tied to any app, any data, any cloud,” he said. “What that fundamentally means for customers is that they can seamlessly deploy their application workloads and have it optimized and accelerated [and] seamlessly have their data stitched together across on-prem and cloud.”
There’s a lot going on with NetApp’s push to be the center of both making data AI-ready and then protecting it. Read on to learn about the company’s new technologies.
NetApp AFX Disaggregated Storage Platform
Singh (pictured) introduced NetApp AFX, which he described as the “first enterprise-grade disaggregated storage architecture.”
“NetApp AFX gives customers the levels of performance that they need to be able to keep their GPUs fed, wherever the GPUs live,” he said. “It is a disaggregated architecture so that customers can scale the performance and capacity independently. It gives them linear scaling as well as efficient scaling.”
NetApp AFX is an exascale architecture, Singh said.
“It can scale into terabytes per second of bandwidth,” he said. “That is critically important not only for inferencing use cases, but also for the training and checkpointing use cases there. And it basically can deliver all the way into 128-node scalability.”
Tied to that performance and scale are NetApp’s enterprise-grade data management capabilities, Singh said.
“AFX is built on Ontap, this time disaggregated Ontap,” he said. “What that means is AFX, first of all, is plug-and-play. One of our early access customers, a very large enterprise, was able to install AFX and, with all of the automation with Ontap that they had in place, they were able to just go and manage AFX with it. Secondly, because it’s Ontap, it brings enterprise-proven reliability, availability, multitenancy. When you think about the centers of excellence that customers are creating for AI, ultimately it requires multiple data science and data engineering teams and eventually agents to have access to the right datasets. And this is where you need that secure multitenancy and [quality of service] built in.”
Data mobility also gets a boost with NetApp AFX, Singh said.
“Think about the use cases where customers’ data might be in one data center and their GPU cloud might be in a different data center, or it might be in the cloud, and customers need to be able to have a zero-copy access to that data while preserving all of the security permissions,” he said. “NetApp with Ontap, it is just built in and available to them.”
AFX is also cloud-connected because Ontap is the backbone that unifies on-prem and cloud, providing seamless accessibility to data using AI tools and frameworks in the public cloud, Singh said.
“We have been investing in integration, for example, of [Google] Vertex AI and Azure AI, fundamentally enabling customers to have hybrid cloud workflows,” he said. “AFX can be seamlessly accessible using our FlexCache technology that gives that zero-copy access in the cloud, and then we seamlessly integrate that into Vertex AI and Azure AI.”
AFX With Nvidia
NetApp AFX was built with Nvidia’s AI technology in mind. AI applications can run on NetApp’s new DX-50 Nvidia L40 GPU-based storage controllers, which attach directly to the cluster over Ethernet, Singh said.
“It is a standard Ethernet network that we’re using across the cluster,” he said. “Your controllers plug into it. The storage shelves plug into it. The DX 50 plugs into it.”
NetApp AFX is also Nvidia DGX SuperPOD-validated, including with DGX GB300, to let customers seamlessly adopt it for their AI initiatives, Singh said.
AI Data Engine
NetApp also introduced the NetApp AI Data Engine, or AIDE, which the company describes as a secure, unified extension of Ontap integrated with the Nvidia AI Data Platform reference design aimed at helping organizations simplify and secure their entire AI data pipeline and manage it via a single, unified control plane.
AIDE unlocks a secure and efficient AI data pipeline so that customers can seamlessly leverage that as part of their AI initiatives, whether it’s for RAG or inferencing use cases, Singh said.
“Think about it as being able to, in real time, search and discover data rapidly, classify all of that data, and protect all of the data privacy,” he said. “Have it continuously connected so that as and when data changes, it is updated very rapidly. The vector embeddings are generated in a highly efficient manner.”
Singh said that customers typically find that data expands by 10X or 10X when vector embeddings are generated.
“With this AI Data Engine embedding the vector embeddings natively, we have a super-efficient mechanism for customers to have efficient vector embeddings available to them,” he said. “What that means is it’s simple, it’s secure, it’s efficient, and all of the security permissions are just natively protected there.”
Gagan Gulati (pictured), senior vice president and general manager for data services at NetApp, described four primary pillars on which AIDE is built during the conference’s press presentation.
The first is a metadata engine to help customers with data discovery, Gulati said. This, he said, is a major issue as data is rapidly growing across data silos in on-premises data centers and multiple clouds.
The second is data sync, he said.
“When data scientists create a dataset so they can train their model, they typically create a copy of the dataset,” he said. “They leave the original dataset alone, but the moment you create a copy of that dataset that copy is old because the original data keeps changing. So how do you now keep the datasets in sync because if you don’t keep them in sync, then the results that the AI will produce could be massively different from what the data is saying. And that’s where the data sync capability comes in. It uses the best of what NetApp has to offer in terms of our technologies with SnapDiff and SnapMirror to make sure the datasets and the metadata about them are always up to date.”
The third pillar is data guardrails, which address security and governance issues around AI data, Gulati said. Data guardrails in AIDE anonymize certain parts of a dataset, such as personally identifiable information or regulatory information to ensure the dataset is completely functional while addressing compliance and security issues, he said.
The fourth pillar of the AI Data Engine is the data curator, a module that NetApp built in conjunction with Nvidia to address issues around data bloat that happen as data changes lead to multiple copies of data becoming vector embeddings to go into the AI data pipelines, Gulati said.
To do that, NetApp worked with Nvidia to take advantage of that company’s vectorization APIs in Nvidia NIM or other APIs to reduce data bloat up to 5X in order to reduce storage costs and improve AI performance, he said.
AIDE And Keystone
All the benefits of disaggregated Ontap with AFX and the AI Data Engine will be available as part of NetApp’s Keystone Storage-as-a-Service offering, Singh said.
“A lot of the enterprises that are in POC [proof-of-concept] stages and that are just graduating into production use cases can seamlessly adopt this AI infrastructure as a service via Keystone with our industry-leading Storage as a Service, and be able to consume that and pay as they go and scale seamlessly as their AI initiatives grow,” he said. “We are super excited about not only unveiling AFX and AIDE but also making Keystone AI infrastructure as a service available to our customers.”
NetApp Ransomware Resilience Suite
NetApp is expanding the NetApp Ransomware Resilience autonomous ransomware protection suite, formerly known as Blue XP.
NetApp this week added two key new capabilities to NetApp Ransomware Resilience, Gulati (pictured) said.
The first is data breach detection, which utilizes the power of machine learning models to fully understand a user’s behavior at a storage layer to detect anomalies that show a data breach is happening, Gulati said.
“We work with companies like Cisco Splunk to be able to create an end-to-end pipeline of allowing security operators to be able to actually take action, like blocking users from doing any more damage across the NetApp estate,” he said. “And you can do that by being there in the Cisco Splunk console, or you can come to the NetApp console to do that.”
The second is an isolated recovery environment that gives operators the opportunity when recovering from a breach to upload the snapshot or backup copy and make sure it is clean from malware and viruses, Gulati said.
“This basically is to give our customers the ability and the confidence that they’re recovering from this attack and that after that, they’re not going to get reattacked, at least on the same data,” he said.
Expanded NetApp Shift Toolkit
Jeff Baxter, NetApp’s vice president of product marketing, used his keynote presentation to introduce changes to the NetApp Shift Toolkit.
The NetApp Shift Toolkit allows people to easily migrate between different hypervisors without having to recopy all of the data, Baxter said. So, for instance, if a customer wants to migrate a workload from Hyper-V to VMware or from VMware to Hyper-V, it can be done within minutes via NetApp’s Ontap operating system by simply refactoring the bits on disk to suddenly make it a new virtual machine without having to copy it, he said.
The new version of the NetApp Shift Toolkit now works with all KVM-based VMs, including Red Hat Open Shift, Oracle Linux, Proxmox or generally any other KVM derivatives, he said.
For Channel Partners
Channel partners are going to be a big part of taking NetApp’s AI-focused data technology to market, Singh said.
“First of all, on the AFX side, the beautiful thing is, it’s just Ontap,” he said. “All of our channel partners natively understand Ontap technology and because it’s plug-and-play, it’s going to be super simple for customers.”
For the channel, NetApp also invested in its validated AIPod reference architectures, which bring compute, networking and storage alongside an AI software stack specifically for the channel, Singh said. The company also offers an Intel-based AIPod Mini and still offers the FlexPod converged infrastructure platform.
“FlexPod just continues to be a fantastic overall converged infrastructure solution that’s jointly developed between NetApp and Cisco,” he said. “We are adding AFX over to FlexPod as well. [That’s] going to tremendously help our channel partners because they are the key in bringing FlexPod to customers.”
Evotek On NetApp’s Message
Mark Pankow, director of platform solutions at Evotek, a San Diego-based solution provider and NetApp channel partner, told CRN that NetApp continues to talk about what customers need to take advantage of the value of their data.
“I really like NetApp’s message on unified data management across all three major cloud providers and on-prem,” Pankow said. “That’s a really compelling capability that is differentiating. The other thing that really piqued my interest was the metadata engine because I think it’s one of the biggest problems in the storage world, and more broadly in the data world, that businesses, regardless of whether they’re younger or established, don’t know what data they have or where it is. The ability of NetApp’s
metadata engine to pretty seamlessly help discover and classify and tag and understand your data landscape, that’s pretty compelling.”
This is a key requirement as customers adopt AI, Pankow said.
“In order to get AI-ready data, you have to know the state of your data, where it is, what it is, in order to leverage it,” he said.
Pankow said Evotek is starting to see customers take a much more balanced approach to where workloads should reside, and the cloud isn’t necessarily the place for all workloads.
“We’re starting to see, at least in the more mature companies, an evaluation of where these workloads should run, whether that’s for performance, or maybe it’s more to do with resilience and security,” he said. “But data is not going to all live in one cloud. It’s going to be multi-cloud, or it’s going to be a hybrid scenario. So the ability to
reduce data movement between these different data silos, whether it’s cloud or on-prem, and to have visibility and a unified experience both operationally and analytically, is going to become more and more important.”
DataEndure On NetApp’s Security Capabilities
Steve Bos, a technical account manager at DataEndure, a Santa Clara, Calif.-based solution provider and NetApp channel partner, told CRN that for his company, which two years ago spun out its WhiteDog security group as an independent business, security of customers’ data is paramount.
The new security capabilities unveiled by NetApp apply to more than AI use cases, Bos said.
“In my mind, the AI thing, I get that NetApp has to have that messaging because the biggest companies are doing it,” he said. “Customers need a storage infrastructure that supports their AI initiatives. That’s cool, but when you get work with smaller customers, they’re turning on Copilot, and they’re just trying to figure out how do I find files when I can’t find files. They’re not building the infrastructure yet. I think that will come, certainly in time, a little bit like the cloud messaging from NetApp five or six years ago.”
Like cloud, when businesses discuss AI three or four years from now, they will all have some sort of infrastructure and storage to support their AI initiatives, Bos said.
“But my point is, the security play that NetApp has applies to everyone across the board,” he said. “Every single person in DataEndure like in every other partner and business development company wants to talk about cybersecurity and data protection. “We have our security platform that we could potentially tie with the capabilities of NetApp as another layer of security. And when you think about resellers across the board, I think very few have their own security platform like we can access. We can tie into those really sophisticated capabilities from NetApp to detect a ransomware attack and report that into our security platform.”
And by adding data breach detection, NetApp is taking cybersecurity to a new level, Bos said.
“Don’t get me started, but this data breach part, I mean, if there’s anybody else out there that has the kind of data breach detection capability that NetApp is talking about, I don’t know about it.”
Bos said he has seen studies that show that bad actors are in a business’ environment for six months before they actually attack, giving them plenty of time to find the most important data and get elevated privileges.
“With our security platform, we’re trying to get that down to six minutes, so as soon as they’re in the environment, boom, we can get them out,” he said. “It’s the same kind of idea that NetApp is doing. They detect an attack within seconds. They can detect anomalous behavior and then take action on that right away. The key to security is to do it early. Of course, we would never propose a NetApp-only solution. You obviously need endpoint detection and a whole bunch of things. But this is another layer, and it’s at the layer that the hackers are after. The data is what everyone’s going for in your environment.”
