We're still an hour or so out from the opening keynote with CEO George Kurian, but while we're waiting why not catch up on what we expect to hear at the conference this week?
NetApp's data management push is likely to be a big talking point. As with its counterparts elsewhere in the industry, NetApp is no longer a simple hardware company, it's offering a full stack setup for customers spanning hardware, software, and open standards.
Elsewhere, you can't escape the spectre of hardware costs. The memory crunch has hit organizations hard, with vendors hiking prices to compensate for the crisis. NetApp will likely look to calm customer concerns on this front.
Crowds are beginning to descend on the conference center here at the MGM Grand, with a steady flow of attendees making their way to the conference ahead of the keynote.
We're in the keynote theatre and now it's just a waiting game until the opening keynote. A lot of attendees making their way in now.
Here we go! CEO George Kurian kicks things off at NetApp Insight 2026 with a message on the importance of data readiness in the age of AI.
"While much of the AI conversation has evolved, shifting focus from discussions around models to GPUs and AI factories to inference performance and token economics," he says
"What has not changed is the critical importance of your data and your data infrastructure.”
Data is the leading constraint for organizations dabbling in AI, Kurian says. The company 'knows a lot about data' - and it's here to help customers.
93% of organizations report at least one data-related challenge is blocking their AI ambitions. This is exacerbated when organizations scale the technology, particularly with agentic AI.
This is why organizations have to overhaul and redefine their data estates. Modern data demands require modern data architectures to help improve visibility and data flows. This is critical given around 85% of data sits distributed in silos.
The actual type of data also matters here as well, Kurian notes. Engineering documentation research data, operational files, media assets, you name it, these are spread across an array of fragmented environments.
This 'dark data' is an untapped goldmine for organizations, but they're simply not able to capitalize on this.
These operational challenges are paralleled by a range of other considerations, Kurian says.
Enterprises also face rising cybersecurity-related threats such as ransomware, as well as new regulatory compliance considerations. Data sovereignty has rapidly become a key talking point for enterprises on both sides of the Atlantic. These are exacerbating challenges for enterprises.
Kurian says that new, modern data architectures require both memory and context.
"Together, memory and context create feeling, impact. And combined with actions, this provides actions that provide for continuous learning," he says.
"For you enterprise to be truly AI enabled, it needs context and memory for agents."
Of course, this is where the NetApp platform comes in, enabling organizations to consolidate and unify their data estates and to provide a solid foundation for AI innovation.
"You've trusted us with your data, your organization's memory for more than three decades, and you can continue to trust us that we will provide you with the tools to turn your data into knowledge."
The NetApp Platform performs several key roles for organizations, according to Kurian. First and foremost, the platform offers intuitive management capabilities for customers to help make their data 'AI-ready'.
Second, the platform helps unify storage estates, tackling those pesky bottlenecks and silos that hamper AI innovation.
"No silos, total visibility," he adds. Security is also a key focus, with Kurian noting that NetApp boasts "unquestionably the most secure storage on the planet".
And with that, Kurian makes way for Syam Nair, chief product officer at NetApp.
Nair begins his talk with a question: "Where do I put my data?"
The answer, he says, is "scalable, durable data storage". That's where NetApp comes in.
"We became the most trusted place to hold your intellectual property, your most precious asset, data. Everybody talks about it, the real truth of the enterprise," he says.
Enterprise data has never been more important in the age of AI, according to Nair. It's the lifeblood that fuels AI applications and agents. And it needs to be 'AI-ready.
"Agents that act on your behalf are only as good as the data they can reach, trust, and act on safely," Nair adds. Data fragmentation hampers this, preventing agents from accessing vital information to support users and conduct tasks.
This is where the NetApp AI Data Engine (AIDE) helps enterprises clean up and consolidate data, allowing them to derive context and value from their data.
"Your data is everywhere, but what has been missing is the layer that knows what is in the data," he says. "The layer that turns a data estate into the memory and context."
AIDE offers enterprises a range of storage and data management capabilities, allowing them to improve data visibility. The service launched last year, and Nair says engagement with customers over the last 12 months has shown its value.
Nair says cleaning up and consolidating data is easier said than done, however. Take medical researchers, for example. They sit on mountains of data sources and various assets dating back decades in some cases.
Pulling this all together and deriving context is a monumental task. This is where services such as AIDE helps streamline processes.
We're onto security and resilience now. A critical focus for organizations globally.
"If it's valuable, it's a target, and if it's a target, it needs protecting," Nair says. NetApp offers enterprises "proactive protection" and a range of tools designed to improve visibility for defenders.
Security and data protection is more important than ever with the advent of agents. These bots have deep access to mission-critical data. Nair says this means we "need a new meaning of zero trust".
This isn't just about network posturing or an "identity check at the door", either. The NetApp platform has zero trust capabilities built in to shore up defenses and keep data secure.
Backups are also critical when one considers the threat of ransomware, Nair adds. "Detection with recovery is just an expensive alarm". Nair says the NetApp platform helps "close the loop" when something goes wrong, supporting recovery and backup capabilities.
Many organizations are now running hybrid cloud setups, Nair says, which adds an additional layer of complexity when it comes to data visibility.
"Infrastructure in four different places, four different teams, four different consoles," he says. "Admins can't tell with confidence whether the same policy is enforced everywhere all at once."
The solution here, according to Nair, lies in the NetApp Console. This offers users intuitive data management capabilities and "complete unified control". The console provides a single control plane for users to keep tabs and monitor disparate storage infrastructure.
We've moved onto unified storage now. Scalable, flexible infrastructure is vital for organizations to contend with evolving business needs. This is where NetApp ONTAP plays a vital role, acting as the foundation that handles enterprise workloads.
For workloads that need "something more specialized", that's where StorageGRID also provides users with high-performance object storage at a "massive scale".
Put simply, the platform is highly flexible and capable of handling a vast array of different workloads and formats.
Arindam Banerjee, Chief Platform and Technology Officer at NetApp has taken the stage now to discuss how storage architectures need to evolve to contend with surging performance demands.
The rise of AI factories means many storage architectures can't keep pace in terms of performance, he says.
Now for the big product announcements….
The NetApp AI Data Engine now supports heterogeneous data discovery capabilities, Syam Nair says this will allow users to improve data visibility across any environment or storage architecture.
The company is also offering customers the ability to map their data “at no cost” for six months.
“Your data, activated, wherever it lives.”
Keystone Sovereign is a big talking point at NetApp Insight. This is a new sovereignty-focused add-on for the keystone storage as a service (STaaS) offering.
This will allow users to select dedicated sovereign options, with the roll-out initially beginning in France and Germany.
You can read more about Keystone Sovereign here.
And now we have the biggest announcement of the event this week, the launch of NetApp Novus. CEO George Kurian says this marks one of the biggest announcements ever from NetApp.
NetApp Novus is a new storage architecture designed to contend with the huge performance requirements of AI factories. Storage, the company says, has become a recurring bottleneck for these GPU-hungry environments.
The system boasts throughput of up to 100Tbps and is a zetta-scale file system - utterly bonkers numbers. Kurian says this is equivalent to unifying all of the world’s data center capacity that shipped last year.
And with that, the opening day keynote is over. We'll be back tomorrow for the day-two Innovation keynote.
With just over an hour to go until the keynote session, why not catch up on all of our coverage so far?
Keystone Sovereign
NetApp unveiled Keystone Sovereign during the opening keynote at Insight. This is an add-on for the company's Keystone storage as a service (STaaS) offering designed to provide users with greater sovereignty controls.
As part of the move, customers will be able to select where hardware is hosted, and NetApp has pledged to use local staff for operational support. The service is rolling out in Germany and France first, then across Europe and further afield.
You can read more on the announcement here.
NetApp Novus targets AI factory bottlenecks
NetApp has unveiled a file system it claims is the “fastest storage on the planet” in a bid to tackle AI factory bottlenecks.
Dubbed ‘NetApp Novus’, the new system is designed primarily for organizations running AI factories, such as Neoclouds and GPU as a service providers.
You can read more on NetApp Novus here.
AI transformation needs bold leadership
NetApp CEO George Kurian believes that despite the hype surrounding AI, achieving success with the technology still rests firmly on time-tested transformation tactics.
Speaking to ITPro, the chief executive said that long-running adoption challenges with the technology show that it can’t become a case of the tail wagging the dog.
You can read more here.
NetApp wants businesses to ‘bring your own model’ in storage automation push
NetApp has unveiled a raft of updates aimed at driving automated storage management capabilities for customers.
With an expansion to the NetApp Console, users will now be able to directly integrate large language models (LLMs) within the console and make use of new natural language chatbot features.
You can read more on the announcement here.
The keynote theatre at the MGM Grand is starting to fill up now with a steady flow of attendees making their way in ahead of the day-two keynote. Stay tuned for all the talking points, we're just ten minutes out.
And here we go! up first in the day-two keynote this morning we have chief product officer, Syam Nair.
Yesterday, NetApp unveiled Novus. Today "we get to take the hood of that".
Nair says NetApp has always been at the forefront of innovation in the storage industry.
"Every time the industry has gone through a major shift, NetApp has always led," he says. When compute moved from mainframes to open systems, NetApp pioneered NFS, he says.
"We taught the industry how to share data across the network, and when virtualization became real, we led with unified storage. One system with native data management built in."
As we learned yesterday, NetApp Novus aims to tackle the bottlenecks created by traditional storage architectures. AI factories require high performance, lightning fast storage capabilities to keep GPUs fed, and up until now that's been tricky.
Novus solves this.
Taking the stage now we have Arindam Banerjee, chief platform and technology officer at NetApp to discuss how Novus is being used in the real world right now.
Joining him is David Ward, president for cloud services at Lambda.
Ward says Lambda is working closely with NetApp to hand over a 10,000 plus GPU superintelligence to an unnamed customer.
"Lambda and NetApp have been working together to provide storage for that environment, and it's one of the first of its kind," Ward says.
"For us as Lambda, we're evolving the services that we offer to our customer to expand more into storage."
Banerjee says the industry is reaching a "profound" moment that will be defined by how efficiently AI factories and storage interact. With Novus, the company hopes to streamline operations and keep those data hungry GPUs fed.
This also has a big impact on energy efficiency as well. With GPU utilization as low as 30% in some instances, that's a massive waste of resources. Novus tackles this directly by maintaining those vital data flows.
Banerjee says older architectures in storage are outdated and can't contend with Ai factory demands.
"Two architectures got us here," he says. "But neither was designed for this."
"Enterprise NAS-scale shared access and data services at human-based concurrency. Parallel file systems scale bandwidth for a number of sequential jobs."
AI factories do both at once, however. Billions of reads for data loading, while terrabyte-scale checkpoint writes land across entire fleets at the same time.
Banerjee says that at this stage, "failure is a steady state".
"A stalled read in a synchronized training job does not slow down a node, it leaves an entire cluster waiting."
Take a cluster of 100,000 GPUs, and this amounts to tens of millions of dollars spent every day on wasted GPU capacity, he says.
Parallel file systems have a fundamental problem when it comes to Ai factories, which "push the design into a territory it was never built for," according to Banerjee.
A big advantage of Novus, according to Banerjee, is that it "does not share a bottleneck" when it comes to storage. Metadata, for example, scales independently, as does data. Put simply, the traditional storage architectures which held back data flows to GPUs no longer slow down the entire process.
Previous processes for improving data flows including 'sharding' - but this didn't completely remove the bottleneck. The answer to this lies in Lattice, a metadata architecture that disaggregates file systems.
"The protocol service is disaggregated from the actual metadata store," he explains. "The protocol service managed NFS sessions and scales with clients. It's lightweight and, by design, ephemeral."
"On the other hand, the metadata store is a sharded KV store. It's a KV database that scales with the file count. Multiple planes, multiple axes, independently scaling from one instance."
The numbers touted by NetApp in terms of performance with Novus are eye-watering. The system boasts 100Tbps throughput, capable of contending with up to 50,000 GPUs.
Underpinning this, the software is designed to perform "at hardware speed". Banerjee says this prevents a tail wagging the dog type situation where the hardware is beholden to the associated software.
That removes another key bottleneck for organizations running AI factories.
"The future of AI infrastructure is being written right now," Banerjee says. The company's lead partner on the Novus push is Nvidia. Jason Hardy, VP of storage technology at Nvidia, has joined us now to discuss the impact of storage on AI factory efficiency.
"There's no denying that the impact of storage on the AI factory is continuing to be well understood," Hardy says.
Hardy says that when Nvidia first started AI training, throughput became an issue. We've gone far beyond that now, though. Hardy says the chipmaker has had to create a "whole new classification" for storage.
Put simply, it's now critical, and it's the leading bottleneck.
Data is the "raw material" that makes the AI factory ticks and drives intelligence. If you're unable to feed that fast enough, then you face big problems.
"The ability for storage to continue to deliver on the data, the speed the factory operates, while also addressing the new memory requirements like context memory and everything else ... it's pretty significant challenge that we have to undertake."
Banerjee once again emphasizes the fact this is the world's first zettabyte-scale file system. It's dominated the messaging from the company so far this week and clearly a point of pride for NetApp.
Banerjee makes way for Gunna Marripudi, VP of product management at NetApp now. Marripudi is one of the lead architects behind the Novus system.
Marripudi says "AI resource efficiency" is one of the longest-running challenges when it comes to AI factories. Power efficiency, for example, is a big pain point.
The latest GPU architectures and AI factories built around them demand "more and more power and cooling than ever before".
"In many environments, growth is no longer limited by GPUs. It is limited by the availability of power," he says.
Marripudi says that compared to other NFS storage alternatives, Novus can reduce power requirements by up to 31%. Pretty significant savings here that will most certainly add up as AI factories scale from 10,000 GPUs to 50,000 and more.
Scalability is also a big advantage for Novus, according to Marripudi. As demand increases, users can add resources that are constrained, rather than "scaling the entire system".
Many AI factories are also constrained by floor space. Marripudi says NetApp will extend the Novus architecture "beyond traditional appliance-based deployment models".
The actual use of AI factories differs vastly based on the different needs of the organization.
Some may be training frontier models, for example, while another may be fine-tuning. Elsewhere, others might be running inference at huge scale.
Each of these brings "different requirements for performance, capacity, isolation, and operations," Banerjee notes.
With Novus, NetApp aims to provide the same level of performance, capacity, and everything else regardless of the particular task or workload.
"That gives operators flexibility without forcing every workload, every tenant, or every customer under one rigid model," he says.
We've reached the end of the day-two keynote now, and it's clear the company is very confident about the potential of Novus.
NetApp's messaging on this front has been very bullish across the two days so far, and we can expect the company to keep driving home the benefits and advantages of the system.
We hope you've enjoyed our rolling coverage of the event this week. For all of our news, views, and analysis of NetApp Insight 2026, check out our website or sign up to the ITPro newsletter.