Pure Accelerate 2026 | Chadd Kenney, Everpure

Clips
More from Pure Accelerate 2026

Chadd Kenney

VP Product Management

Everpure

play_circle_outline Breaking Silos, Building Shared Context: Shifting from Application‑Centric to Data‑Centric Enterprise Architectures for AI Agents

play_circle_outline Real-time data access enabling AI agents for live decision-making

play_circle_outline Changing buyer personas: infrastructure teams, CDOs, and lines of business

Info
Transcript

Chadd Kenney, Everpure

Chadd Kenney

VP Product Management Everpure

In this interview from Pure Accelerate 2026, Chadd Kenney, vice president of product management at Everpure, joins theCUBE's Christophe Bertrand and Alison Kosik to discuss how enterprise storage is being repositioned as an active data platform for the AI era. Kenney leads product management across Everpure's full portfolio — FlashArray, FlashBlade, Everpure Cloud and Fusion — and explains why decades of application-centric IT have left enterprise data siloed, latent and ill-suited for AI workloads. He outlines how Everpure's data intelligence layer breaks do... Read more

explore Keep Exploring

What challenges did putting data second in enterprise IT create, and how is AI helping to overcome those challenges? add

How does your data management platform differ from traditional architectures that require moving all data into one place, and how does it access, classify, contextualize, and serve data across diverse storage sources? add

Three years from now, will you still be selling to infrastructure leaders, or do you expect buyer personas to shift (for example to chief data officers or line-of-business owners), and if so how would that change your day‑to‑day role and the way you build, position, and sell products? add

bolt Powered by CUBE AI

Chadd Kenney, Everpure

search

Alison Kosik

>> Welcome to Pure Accelerate 2026. We are streaming live right here in Las Vegas. I'm Alison Kosik, joined alongside by Christophe Bertrand and we are kicking off the day at The Hub.

Christophe Bertrand

>> Yes we are. As you can see, it's very busy. Lots and lots of activity, lots of partners, lots of attendees, and more importantly, a great guest.

Alison Kosik

>> Yeah, we do have a great guest to bring in. We just came off the keynote and I know this guest was in the keynote as well. Let's bring him in. Chadd Kenney, he's the vice president of product management at Everpure. Welcome to theCUBE.

>> Yeah, awesome to be here.

Alison Kosik

>> So talk us through what you do at Everpure.

>> So I lead product management across the entire portfolio. So I get the honor of announcing much of the payload of innovation that we built out across all of our core products, our FlashArray, FlashBlade, Everpure Cloud, as well as Fusion. And so I get to play around with a lot of the innovation and work with engineers to get it built.

Alison Kosik

>> Enterprise IT has been built around applications and compute for what, 50 years now?

>> Oh, yeah.

Alison Kosik

>> Always putting data second. Data has been secondary. What challenges did this cause and what is AI doing to finally kind of break it?

>> Yeah, great question. So if you think about it, applications effectively siloed both the data, the context, as well as its internal workflows. And so this made it very problematic when you were trying to understand context across all data. We tried to solve this in the past through analytics by taking copies of data, building dashboards to try to understand it and munch through that data to kind of consolidate it into what would be classified as kind of a system-like record. But it was unfortunately a copy, it was latent and so it wasn't the best use to use for AI. So we have this belief that if you were to able to break down those silos, take the context and share it across each one of these applications and then later build a system of record with all of that data consolidated, AI agents now could actually be running on top of real-time data versus this latent copy. And so if you think about it allows you to have agents now make decisions across applications. Things like if I was to approve a profitable order as an example, and I created an agent to do a workflow to do that. If they only have access to Salesforce data, they would have to infer what the cogs are and maybe just make up what would be profitable or not. But if they understood what our suppliers were, what the costs were and also what the total product cost was, they could actually fully infer what a profitable order is and then make that workflow work.

Christophe Bertrand

>> Right. So there's quite a bit to unpack here. And what's interesting is that you're not talking about storage. We'll talk a little bit about some of the innovations, but you're really talking about data. And look, I've heard this story before, "Give me your data, I'll do it all for you." Okay. How is this going to be different? Why is it different? And you mentioned context and content to an extent, so I'd love to understand that a little better.

>> I think that's one of the fundamental challenges. I think many of the people that are building architectures around this all believe somewhat of the same thing, right? Data should be central and core to the way that AI uses that information and decisions that we make. But the fundamental challenge that most architectures have is they say, "Give me all your data and then I'll be able to create all of the context and semantics of it." Our belief is different. Well, we built a killer storage platform, our data management platform actually accesses data everywhere. It can go to third party storage solutions that are not Everpure. It can go to SaaS, infrastructure as a service, cloud, all these repositories of data and actually interrogate them, classify the data, and then be able to contextualize it so that you can get that one common shared context.

Christophe Bertrand

>> Right, and this is going far beyond what was known as storage virtualization by certain vendors a long time ago. We're talking about the fact that you have access and visibility over the whole data set across the enterprise. You're adding content or a dimension of understanding of context to it so you can also analyze the content to make sure personal data, for example, is not in the mix. And what you're saying is, and now you can also automatically serve it to the workloads that require it. Let's talk about that a little bit because there's a lot that has to happen under the covers for that to happen. We saw that through the demo, but demos that show magical things are always taken with a grain of salt.

>> Let me give you a bit of a deep dive on data intelligence, maybe that'll help you understand it. So if you think about it, what they're trying to do first with being able to discover things is you have to know all of your data endpoints. We typically integrate with a CMDB like in ServiceNow and it shows you each of your data endpoints. You get read level authentication as a client. It spins up containers, interrogates the data and then brings it back to contextualize it and classify it. This allows you to build out not only the context of where all your data resides, even though we don't control it and it's not on our core platform, we are accessing it and contextualizing it. This knowledge map is what AI agents actually need to infer data across a wide swath of data. And so while there'll be benefits to running this on top of an Everpure storage solution, it also has great benefits of just understanding all of your data together.

Alison Kosik

>> Go through some of the announcements. You had a lot of announcements during the keynote, updates the unified data plan, the intelligent control plane, AI data system, EDC success blueprint. What does this payload deliver together and what does it mean for an enterprise trying to move from fragmented infrastructure to a production ready AI infrastructure?

>> Great question. Okay, so if you think about what we did last year, we effectively allowed people to build their own enterprise data cloud. It was a concept that we had created or a vision of enabling customers to kind of abstract away the core infrastructure stack and really start to think about data as more of a core thing, less just the infrastructure management. I think it allowed our customers to start not thinking about storage as much and really try to think about autonomy and how you build policy and intent. This year we really made it real. We went and built in autonomous ways of managing infrastructure in policy, so first around compliance as an example. Today, everything is very ad hoc, very siloed in nature. It takes months to really be able to apply consistency across the environment and through some of the compliance remediation and reporting, we enable customers to do this with minimal work at all. Apply it and it just executes. The second part was really around performance management. It's kind of the bane of every storage guy's existence, right? They get a call from an application owner, they're complaining about performance and they fire, fight and try to figure it out. What we do now is we take the same methodology we had in support and predict these issues before they occur, now we're doing it in performance. So think about the benefit of it. It's almost like accident avoidance. You're effectively predicting an accident before it occurs and you never get into them. The same is true, you never have an application owner call you about performance issues because we predicted ahead of time and rebalanced the fleet for it. So these kind of autonomous capabilities I think are going to be complete game changers so that people don't have to think about infrastructure anymore and can now start thinking about more of what we're talking about with data intelligence.

Alison Kosik

>> Right, yeah.

Christophe Bertrand

>> Yeah, and to clarify, this is where AI is used in the architecture. So it's really two sides of the coin, right?

>> Internal and external.

Christophe Bertrand

>> There's AI for the purpose of automation essentially to serve the AI infrastructure itself in the various applications. So you talk about data privacy again, I think that's a very interesting concept. What do you really mean by that?

>> Yeah, good question. I think over time, and this will be a bit of an evolution, there's going to be certain applications that today the data resides in the app and the transactions occur in the app, but over time maybe it will build itself into a common system of record of which both transactions for the application can hit that data in real time as well as agents that are inferring and making decision, and of course, analytics. So if you think about this, over time we believe there'll be a new substrate built in and that substrate is this data intelligence layer of which you build applications through AI, you build agents through AI, of course, and you have analytics that now are also built through AI that are now accessing that context and knowledge graph as well as the local transactional data. I think that's the longer term dream everyone has. They have full control of their data. It's real time. So if you think about it, the last thing you want to... In that same example I gave you of like an agent approving a profitable order, if it's a copy from yesterday, you may have had multiple versions of that quote at this point in time, right? But if it's real time transactional data, it can make decisions right along the human who's actually doing the work. That's what we're trying to get to, real time data access with AI agents running together with humans.

Alison Kosik

>> What do you think is the biggest mindset shift that's required from IT leaders who've spent decades around building around applications instead of data?

>> Yeah, so I think one of the big areas I hear from customers is they're sick of the firefighting. They're sick of like ... I have people saying that they don't get their weekends anymore. They want their weekends back because they're dealing with sev1s. It would be hard for me to sleep at night if I had hundreds of systems and figuring out did I have everything set up correctly. I think by building in this ability to deliver compliance, deliver dynamic infrastructure, we are effectively allowing them finally to get more strategic and that strategic part is actually starting to rationalize the data and there's many steps. What we're doing kind of in my session at least was talking about a bottoms up way of abstracting away infrastructure so that you have the ability to do what Charlie talked about, which is the tops down management of data. Over time, these two worlds connect. Maybe next year you'll hear more about it, but I hinted to it at the end of mine, which is governance will follow the data. Right now, governance is following the storage resources. Over time, that same governance policy will actually follow the data entities within that storage resource. And that's, I think, where we'll see our future innovation head towards.

Christophe Bertrand

>> Well, let's talk about governance because obviously that's a big topic and there are multiple layers here. There's governance in terms of do I have the right data for the purpose, the business purpose I'm trying to serve? Is it compliant? It's obviously the low hanging fruit. But look, I mean, we started talking about this 20 years ago with Sarbanes-Oxley.

>> Yeah.

Christophe Bertrand

>> Okay, so what has really changed there because it's not new.

>> So I think understanding the data is a little bit more of it. I think there wasn't deep interrogation. I'll give you a couple examples. You can imagine a Word doc. In many cases, we would have an OCR of the text, but what if there's an image with a check or credit card information there? A lot of times it was missed, right? And so being able to actually use large language models to do classification nowadays as you discover starts to open up new avenues to really truly understand the data in its contextual nature. The second big part is we didn't have a knowledge graph, which is with the big innovation we didn't talk too much about, but it really is the thing that allows you for semantic level search but also connecting data sources together. So in the example we used in the demo, it was like correlating medical data across two individuals, while many times you would know there was medical information out there and maybe even understand what was in it, the correlation of that information was always missing. I think that's where you're going to see the true impact of being able to really understand the data more effectively.

Alison Kosik

>> So if we're back at another Accelerate, let's say in three years from now, what do you think will be the defining characteristic of companies that have successfully embraced data privacy compared to those companies that didn't?

>> Yeah, great question. So the first I think from the bottoms up is they built an autonomous infrastructure. The cloud kind of delivered this, by the way, but infrastructure on prem really has never had it. And so this means that it self-manages, self-heals, dynamically uses the infrastructure effectively, so there's not people managing it anymore. Beyond that, there's an understanding of the data and they're getting full use of it. And so I think in a world where we're actually applying things like knowledge graphs and those types of things to interconnect data, these guys are going to be able to use AI well ahead of anybody else in the market. If three years from now people don't take advantage of this innovation, they're going to still be stuck managing infrastructure and not actually understanding their data yet.

Christophe Bertrand

>> So this is interesting because three years from now, are you still selling to the infrastructure guy? You are head of product management, so personas have to be front and center to a lot of discussions you're having, how you build the products, who you build the products for, and how you architect even the sales force to go talk to the right people at the right time, not even mentioning marketing. So I'm curious about that changing persona. We talked about compliance, now we're talking about a chief data officer. Well, that's definitely not an infrastructure person. We're talking about understanding the actual, having cognitive intelligence on what to do with the data for the purpose of business transactions. Well, that's probably a line of business type of person. How does it change your job every day and how you build the products?

>> So what's really, really fun, by the way, is yes, we talk to infrastructure folks a decent amount, but what ends up happening is we introduce this concept of data intelligence as well as data privacy and all that stuff and all of a sudden they kind of light up and they're like, "This is super top of mind." And you'll be surprised how many infrastructure leaders, especially at the higher levels of things, that this is very top of mind, that they're being pressured by their CDO to say, "You need to help me with this. This isn't just a chief data officer and data team. This is a whole squad we need to get together to operationalize this." And so it opens up conversations for us to access those individuals and learn a lot more about how we can both help from bottoms up as well as tops down. And so we're learning a ton. This is early days in the data intelligence game for us with 1touch, but we've been gaining a lot of ground just in understanding the customer motion, understanding how we make AI ready data with things like data stream as well as data intelligence and it's been a blast to learn to tell you the truth.

Christophe Bertrand

>> Well, let's carry on with one more thing. We haven't talked too much about cyber resiliency. Just to clarify, all of this has to happen with full cyber resiliency, you've introduced a bunch of new capabilities about a couple of weeks ago, cool demonstration on automatic failover for file environments, which was one of the innovations, very cool stuff. All of this participates in the AI infrastructure, right? So there is no AI without good classified data. There is no AI without secure data, right?

>> 100% agree. Yeah, if you think about it, we did a big launch on this earlier and we kind of have this concept of storage is actually your last line of defense. If you think about it, people are going to come in, you try to catch them as much as possible, but the recovery of that data through SafeMode in particular is one of the key capabilities to save customers more than anything else. We've added in resiliency capabilities. ActiveCluster was one of the coolest technologies ever to be able to synchronously replicate between the two arrays, but also make it look like one single volume. Bringing this to file was quite interesting because we actually made it a fleet level component. If you think about synchronous replication for the most of storage vendors, it's hardware centric. It's even like port bound in many cases, it's kind of nuts. What we've done here is actually built it as a fleet view. So this thing can move across the fleet and there's no affinity to a box anymore. I mean, I don't know if you caught it, but some of the data controls I talked about as a workload being the new management paradigm is effectively trying to make the storage system just be a resource. Your abstraction layer now is a workload, a thing that moves around and you apply controls to that thing through service levels and policy. Over time though, we're going to understand the data and context and be able to manage it from top down as well and that governance and cyber resiliency will now actually follow the data versus it being stored array bound, if that makes sense.

Alison Kosik

>> Chadd, such a fascinating conversation. Thanks so much for stopping by theCUBE. We loved talking with you today. And you've been watching theCUBE, the leader of live tech coverage and live tech analysis and we will be right back.