NetApp Insight 2025: Tony Chidiac on AFX, AI Data Engine & Real-World AI ROI

I am Steven FoST, and we're here at NetApp Insight this morning. NetApp launched the new a FX storage Array series, as well as the AI data engine product. And to get a little bit more perspective on what these products are for and what customers are excited about, I'm here with Tony Chidiac, who knows?

Uh, well, you, you got a lot of contact with NetApp customers. Yes, correct. Absolutely.

No, very excited. With the launch of A FX and then A IDE, I think it's gonna be a game changer for our go to market with ai, which I run, and, uh, very much looking forward to it. So I, I think that there's maybe some, uh, skepticism out there that AI is really something that's being adopted in the enterprise, but, uh, considering the fact that, uh, you know, they've got you on staff and they're paying your salary, you must be selling some of this stuff.

So overall, uh, what are people, you know, what are the stage of, uh, NetApp customers when it comes to rolling out AI applications? It can be a mixed bag. There's absolutely pieces and components that, obviously we're seeing more and more software that comes in that makes it more accessible and easy, but it's all over the place.

When we look at it, we break it down really to that data staging piece of it, and there's a lot happening there in the enterprise. And then the training, the training has really evolved and changed a lot because of the, the introduction of being able to bring LLMs onto that training environment. So they're not necessarily building their own models.

They're doing a lot of the fine tuning. And actually what we're seeing more and more of is the inferencing side of it. Now, they're not always these big time, you know, sexy use cases, so to speak, uh, but they're ones that are being real ROI into the businesses.

So we're definitely seeing a lot more of that. And then I look at what we just announced today with A FX, uh, and A IDE, I think it's going to be a game changer, especially on that business outcome perspective. And I'll touch on that in a moment on what the AI data engine can do.

Well, I think that that's, um, so let's kind of take this in reverse order then. Yes. So, uh, today, uh, you know, we're here Insight 2025, um, you know, NetApp announced the afx mm-hmm.

Which is essentially a really big, really fast storage array. Yes. Yep.

And the AI data engine, which gives that array a lot of new capabilities. Yes. So how would you describe this to one of your enterprise customers?

Yes, so, absolutely. So the A FX, um, look for the past four or five, six years, we have served AI CU customers with their AI workloads on ONTAP today, uh, full, full stop all the time. What changed today is the desegregated architecture of A FX, which really gives us unlimited scale, uh, much better density and overall performance.

So it allows us really no limitations for our customers. So it's a foundational piece. We're very excited to bring that look.

AI's not slowing down. The, the workloads are getting larger. The clusters are getting larger.

So we went and deployed the A fx Now with the AI data engine that's coming, which, which is game changing. This just makes the enterprises and their data estate more intelligent. They're able to now look at their data in a different way.

They're able to tag that data, they're able to decide who gives access to that data, if that data needs to be anonymized, needs to be OB obligated, um, and then they're ultimately able to then connect it into the robust ecosystem of the AI environment so they can leverage different chatbots, they can leverage different, um, semantic searches to make this really come full circle. So, very excited about what that's going to do. And, um, you know, it's now basically, again, taking your entire data state, looking at it, and now bringing it to life in ways that you want AI to work for you.

Yeah. That really came from, uh, the keynote. Uh, we heard it loud and clear that essentially the goal here is to enable a company to really just have their entire data set and have that theoretically be accessible to ai.

Yep. But the AI data engine isn't the engine that's running those customer applications. The AI data engine has an important role, uh, to serve before the data gets to that engine.

Correct? Yep. No, absolutely.

And just to give you a few actual use cases, ones that we're working on with like a banking customer, their mortgage, um, applications are way down. And they wanted to look at, wait, in the last 2020, since 2022, who has gone and actually, um, been, what's the, what's the customer profile of the customer type that's been accepting and getting their mortgage approvals accepted with the AI data engine, you can instantly search across that and then bring that to the fold and actually use, uh, the different Gen AI applications to not only give you that list of who it is, but then also be able to create content that can then serve up and go out to the marketing, the sales and marketing to actually go and position this. Another example is when you think of media companies or like the NFL, uh, if you wanted to go and say, Hey, Barry Sanders is getting, for example, uh, inducted in the hall of fame, that's already happened, but if you wanna do, you can go and search through all your data with AI data engine, search through all the videos, and actually see, show me Barry Sanders with all the touchdowns that he scored, and you can pull that out.

That's never been done before. Never even possible to be able to take, you know, billions and billions of parameters of data and actually search through that, pull it and bring it to life. Mm-hmm.

There's so many use cases across that, and that's just some of the things that the AI data engine brings. And, and what it's doing is that it is, so it has essentially unrestricted access to the data. Mm-hmm.

And so it's able To go, well, unrestricted with the governance. That's the other piece. There's a robust governance on that where you can set, and if you're based in this location, if you don't have these credentials, it's all policy based.

So you can choose how the data not only is served, you're a Patriots fan, for example. Exactly. Yeah.

So you can choose who actually has access to it, who, what's gonna be served up. You might be able to serve up the data, but you need to anonymize it in certain ways, social security, anything like that, like in the mortgage example. So there's a lot of different mechanisms and tools.

Sorry to cut you off, but I do want to pull that out there. Yeah. Well, and, and what I'm trying to, yeah.

And that's the interesting thing about it to me is that, um, NetApp is not saying give all your data to the chat, Botts. No, absolutely not. And app is saying, let the AI data engine evaluate that data and in an intelligent way, categorize it, classify it, whether it's video, whether it's files, whether it's text, and then you can make decisions on which parts of that data you're gonna present to and applications.

Yep, absolutely. And then they can take it and even train on some of those data sets when they pull certain data that they wanna train on, they can use that to fine tune and so on. So there's a lot of different applications from that standpoint.

So when you go to these customers and, uh, you know, you mentioned, uh, you know, sports, entertainment, uh, finance, uh, you know, you go to these customers, which of the new capabilities are you going to be most excited to be able to share with them? What is the, what is the hook you think that's gonna, that's really gonna get them excited? The, The, the, I mean, A FX to me, um, it, it puts us in a different category from the infrastructure side.

Mm-hmm. Where you're getting basically the best of both worlds. You're getting unmatched, 30 years of ONTAP now, uh, coupled in with a modern architecture that spans and as I touched on unlimited.

So that's exciting. Um, they're getting that it's a necessary need for all AI workloads, but what's gonna transform their business and transform how they can go to market and transform, how they build on their end is the AI data engine. So I'm very excited for that because now, you know, they have, uh, proprietary, um, data estates is really the untapped market.

We've seen chat GPT, we've seen the LMS that's all trained on internet data. This is now the enterprise's time to leverage their proprietary data that they might have 20, 30, 40, 50 years. And it's gold mines.

And some of it might be sitting in archives, some of it might be in the cloud, some of it might be on-prem. And now be able to take the, uh, the AI data engine and go search and pull that and make use of that data in a way that you've never won, even maybe know what existed, but now be able to pull it into an AI data pipeline that's gonna be transformative for the company. So I'm very excited to be bringing that into the fold into the market.

So these products are segmented though, so the AI data engine is separate from the A FX? Correct. And I'm to understand that you could actually use them separately.

Yep. So you could use the AI data engine with like an A FF array, or you could use the af, the A FX separately if you Wanted to and, and in cloud and across. Yep.

Yep. And soon coming soon will be across storage grid as well. Yeah.

So on the, the object side as well. So all that's coming to the fold. So it's, uh, pretty exciting.

And can the AI data engine actually reach, uh, data stored in, you know, uh, AWS or Google or Microsoft? Yep, Absolutely. And I will say, you know, one thing that we wanna do view this as foundational, right?

We want to provide the tools at the storage layer that touches the compute to be able to have these functionalities. But we also recognize there's a robust AI ecosystem of different, uh, applications and softwares. We will be integrating with those as well.

So if you have something installed or working, there're gonna be API calls that can go directly into that. So we're kind of, again, being very good at that foundational letter level, what we provide, but also making sure that we're going with the ecosystem mm-hmm. To be able to maximize opportunity.

And that's really what it's always been about for NetApp. I mean, I think it's, it's really important to, for example, in the, in the cloud era that NetApp was out there and said, we are going to work with Google, with Microsoft, with, uh, you know, AWS, these are all great partners. Uh, we're gonna develop a product that works with them seamlessly.

And, um, you know, I think that as somebody like yourself who's working with customers, I think that's important because you're not trying to lock them into a specific Solution. Absolutely not. Yeah.

And it's the same with, um, from what I'm hearing about the AI data engine, is that this thing is designed to work with basically any, um, AI platform that a customer has or would want to have. Yep. Absolutely.

And I think one actually neat piece is, you know, a lot of enterprises, they have this untapped data that they, they may or may not, but we are with the AI data engineer, you're now bringing in multiple personas. You know, our every day has been the it. Now we're also gonna be speaking with data engineers and up to data scientists, and to help actually bring those three together with the art of the possible, what we're bringing them.

Mm-hmm. It's gonna spark a lot of ideas and a lot of different use cases that, you know, we haven't even thought of yet today. So, you know, there's a lot of untapped opportunity with the data.

Yeah. And, and a lot of opportunity, I think for people who have been traditionally, oh, I'm the storage person. Yep.

To start thinking about, you know what, maybe I'm the data person a little bit, maybe. Exactly. I can actually be at the, at the table with them.

Yep. Exactly. I think that's gonna be something that we see full circle.

Everyone needs to be, you know, and all hands on deck to, to, again, um, it's one thing to have these cool AI projects, these siloed use cases, but transforming your business. You need to be able to talk to all the data. You need to be able to see all your data, and you need to be able to make it active or take action with all your data.

And those are the pieces, uh, that we're bringing to the fold. I think that's a great way to kind of summarize what has been announced today at, uh, insight 2025. Um, essentially, uh, NetApp is trying to build a system that can literally, you know, handle all the data that can, uh, make it more useful, that it can make it available to AI applications.

And, uh, you know, I think that y'all are, um, y'all are smiling a lot. Yeah. And y'all are pretty proud, I think, of, of the announcements that have been made.

So if you'd like to learn a little bit more about that, uh, we are actually gonna be doing some tech field day deep dives with, uh, the product managers, with the technical folks. We're gonna learn a lot about how these things work and what's under the hood about them. Those will be posted to the Tech Field Day website, tech Field day socials, as well as the Tech Field Day YouTube channel.

You'll also see a lot of the content from NetApp Insight over on our partner site Techstrong tv. And of course, you can learn more at the Futurum Group, which is the parent, parent company of all of this. So Tony, thank you very much for giving us a little bit of an insight into the customer perspective mm-hmm.

On these announcements. And we look forward to sharing more from NetApp Insight.