Delegate Roundtable – AI Workloads Meet Data Operations at NetApp Insight 2025

The Tech Field Day delegates have been attending, uh, NetApp Insight this week in, uh, Las Vegas. And we have heard a number of, uh, product, uh, announcements. We heard a lot about ai.

Uh, we learned about a brand new, uh, AI data capability, as well as a brand new, uh, storage platform, the A FX from NetApp. And I thought that it would be great to hear from the delegates, just get a, a bit of a feedback, uh, session here, round table discussion on what they thought of the announcements this week. So, uh, who wants to kick it off?

Uh, start off with the, with the discussion, Glen. Yeah. I, I was very impressed with, um, how NetApp's changed the discussion on, um, on where embedding and tokenization happen, uh, that it's now a storage function.

I think they, they've, they've made that change and, um, how they've updated their, their storage, the physical storage architecture, but the disaggregated, um, shelves and, and the aggregate virtualization, all that great stuff, um, to, uh, to finally be able to, to deal with a lot of the criticisms they've had in the past, but, um, scale out performance and, you know, independently scale compute and, uh, and storage. So, uh, I think they've taken some pretty big strides, uh, while still preserving the things that were good in, in ONTAP and bringing those into future parities. That's a big deal, I think.

Yeah, I, I liked the, the announcements and there's a lot here. What I'm really interested to see is whether they have just kind of really realized what they've stumbled across. 'cause they could, there's a huge amount of potential here, and it'd be really interesting to see where they take it.

'cause they could be leading the way in the way that AI architecture's built. Um, they've got all the components, it's about whether they realize what they've got there and how they execute it. Yeah.

I, I thought it was really interesting where they're going with the AI stuff. It's, I think they're kind of the first in the market to, to land on some of this stuff. So it's, it'll be interesting to see where they go, where they go with it, and the fact that they were able to, to deal with the cloud and the on-prem world, and from a data person who used to play with storage and having in a, in a little bit, um, it was really interesting for me to see, to have, be at a storage event where they, the cloud wasn't a bad word, um, and where they, they're realizing, okay, we need to deal with all this random stuff that's kind of floating all over the place and, and have it as one big piece that we can deal with.

Yeah, I agree. I, I think that, um, I, I hope that they haven't just stumbled across it. I think there's probably a lot of pushing from the inside to make this really well known, that you cannot do AI without data.

However, a storage company is not an AI company and never will be. We are storage people have always been the guardians of the data, and that's what this, this is all about. I will say that at some points, um, I thought that what I was hearing from the stage of the keynotes was almost like people cosplay being up on a stage and making some of these very grandiose, bombastic comments that they don't need to make.

And they don't need to say they are the only ones that do certain things because they offered no proof that they are the only ones that do certain things. At least not in the keynotes. And I don't think they are the only ones that do certain things as far as storage.

Storage is storage. However, I do think that that taking that, that step to say that data is what's important and working with their partners is what's important. And I hope that they can dial in on that message and really deliver it.

Yeah. Saying that data is, uh, key and important, um, is not, amazingly not set enough in the world of ai. And it's great to see a major brand saying that while simultaneously saying, we are a storage brand.

Storage is what we know and do. They're pretty explicit about that. They're playing their position, um, playing well and in partnership with others.

They've always been really strong there. So I think all of that is, is really, it's really nice to see. And it's nice to see it play out in the products where they're, where they are taking this sort of, you know, horizontal approach.

Uh, what really impresses me in a broader sense is really truly, um, I, you know, I hate jumping on trends or being part of trends, but AI is clearly far more than that. And it is so pervasive. It is so everywhere.

Everyone, every, including the storage, admin needs to stand up and take part. And some of the discussion, private discussion that we had was around how that engenders a sort of a cultural change. I know, Jason, you, you talked about that, uh, quite a bit.

I don't know how that happens. That is far beyond net app alone and, um, that that, that everybody sort of transforming or changing or taking new approaches to what they're doing is really necessary. We see here at NetApp with these announcement tools and capabilities that will help fuel that.

But you said it yourself just now, Gina, storage is storage. That is true. But storage today is not necessarily storage of, of, of yesterday.

No, not at all. Um, like I was really happy to hear all those uses of the word data. Of course.

Um, but like, what I kept thinking about over these two days is they're finally putting data back into data management. So we've had these discussions before about how my world of data management meant nothing the same as storage data management, but, and that was because from my point of view, storage just looked, stopped looking at data at the file level pretty much. And, but now they're going into the files, like classifying the data, tagging the data sensitivity levels, discovering the data.

Like I think that is now that we're getting, you know, sort of now data management kind of means the same thing. And of course, all the wor words that I love, data governance, tagging metadata like that says to me that, you know, they really, when they mean data, they actually mean the data that AI needs to know and all the things they're doing are gonna be valuable even for non-AI workloads. Yeah, and I just wanna go back to what Gary Guy was saying there.

I think it's really interesting to see, um, NetApp kind of refocusing been an interesting kind of, I dunno, not quite a decade, but just under, they've, they've, they went out into lots of different things. Um, they went through a massive set of acquisitions, various areas throughout various parts of cloud and other operations. And there was a time when it looked a little bit like they didn't really know where they were or what their identity was.

Um, but now it seems really clear they've leaned back in and like, we, we are good at storage and we're good at helping people with storage. Um, and we've seen some really clear messages around, okay, well, you know, we've got these tools and they did, you know, they've had some of these acquisitions around data governance and other things, and they're now starting to use them in the right way. So they, they kind of, they shut off some of the acquisitions.

They, they've integrated others, and they're kind of now building that back into rather it being a rather large portfolio of lots of disparate products. It's now feels like a lot more integrated. Um, like they've finally found what works.

And I think that's gonna be really useful. I definitely agree to that. It's, it feels like one NetApp and it hadn't in previous events.

Yeah. It, it going to, to the two previous points that were made. Right.

So I think we can now start to look at, just in the data context, start to bring a little ITIL into this. Now you've got architecture, you've got engineering, you've got operations. What NetApp has always done, and I don't think it's been recognized, is that yeah, they got storage, but what they're bringing is a homogenized data operations platform, not engineering.

'cause that's what that, that's what, that's what we were talking about's, what Karen's talking about. Um, and, and, you know, and in architecture as well, I think, you know, those are the upper layers and they certain, and, and NetApp has never really talked to data architecture folks, and they've never really talked to data engineering folks, but they're all about data operations. And one of the things I do wanna take exception with, uh, Gina, is that the one, the thing that, that NetApp does that no one else does, and this is absolutely fact-based, is they are the only ones who have a homogenized platform that is also has first party storage in the clouds in all three of the major hyperscalers.

Nobody else has that. And, and at, at the current time, and there are a few vendors that do have non man customer managed cloud-based versions of their operating systems, but none of them have integrated the way NetApp has with these three clouds and provide this homogenized data operations platform, is what I'm gonna call it. Um, more than it is.

And now storage is a means to an end. You're absolutely correct. Storage is storage, it's born, and you know that that's, but data operations is where NetApp is sitting.

And what they've now done is taken some of these AI operations that are data focused and said, this is not engineering anymore. Embed tokenization, embedding that used to be engineering. We think it's more operations.

We saw that in regular infrastructure too. Things that came out of engineering and started getting into the operational framework. Right.

So I think it's an interesting way to look at it. Um, and I think it helps put NetApp in, in perspective as to where there should live and, and where dominate should dominate. Wow.

You agreed with something I said that was amazing, that's living. Um, I, I wanna get back to something we talked about in one of the, uh, presentations was, uh, NetApp's use of the word workload. And I think this is what's happening in the industry, right?

People, I think we've had so much hype around AI and what he should do and what it can do and what it's gonna do. And, um, now all of a sudden we're getting to the point where, uh, a, a workload, a method to do this, uh, a pattern of doing AI is starting to emerge. And it's emerging enough that storage people can say, oh, it's a workload.

If it's a workload, as a storage person, I know how to handle that. I know how to operationalize that. I know how to make sure you've got the storage where you need it, when you need it.

I know how to give you the access. Now that I understand other people, I, I, we've figured out ways of, of letting non storage people, like, uh, data scientists make their own little data volumes and playgrounds where they can play in. So they've figured all of that out.

So now that we have this AI workload or workflow might be a better way to put it right now, um, NetApp can, um, can put together, has put together platforms to support that, which is storage. Storage has always supported a workload. If we know what the workload is and what the outcome's supposed to be, we can architect to it.

So they're doing what their heritage, their 30 year heritage says they should do. Yeah, I I will, uh, let's try to zoom in on that a little bit, Gina, because I think that's, that's, that's a message that they were trying to send and to the point that we, that we heard from from Jason. And, and what you just said is, as well, this is a, i I, I think a charge that NetApp is, is, I dunno if they're consciously making it or just happen to be making it, but essentially, you know, storage, storage people.

Let's go, let's elevate what we're doing. Let's become more involved, more part of the, part of the work. And I agree with you, Gina, from a, you know, you, we, we go way back in storage.

Workload was not a word we used, but I heard it used a lot to refer to a data set, you know, a consistent set of data that served a set of applications, um, a set of systems and that it had its own sort of paradigms and needs and performance and capacity and things like that. And, and all of these things. It's a much more, I think, user-friendly way to describe storage than the archaic terms that we've used for, for decades.

And I think that that reflects a lot of the messaging that we're hearing here as well, you know, that it's, it's things that are happening up the stack, but at the same time, I'm worried that some storage companies have gotten, I guess two, uh, uh, out over their skis when it comes to, we're not storage companies anymore. We're data companies. We're AI companies, we're compute, we're cloud.

Um, I think NetApp is the only storage company that still understands that it's a storage company, you know, and, and I think that that's good because frankly, I don't find a lot of their, uh, a lot of the other, uh, messaging compelling. So, um, yeah, let me throw that at you. At you.

What do you think about a storage company? Um, we'll, we'll go to, we'll go to Guy and then, and then Gina. Yeah, I think it's worth exploring, um, AI data engine, which is one of the big announcements that that mm-hmm.

In, in a little bit of detail because I think, you know, uh, where these claims and these big statements, you know, um, another one was, uh, um, we don't try to pro produce like a lot of companies do. There's a claim from NetApp. We don't try to produce what a lot of companies do, and like an entire solution.

We know where we sit and we use partnerships to, to, to fill in the rest. Um, the test for these claims is what the product actually does, and we all got a chance to, to do a little walkthrough of the product, see what its functionality is. And we also got some perspective during these sessions as to how they've designed it.

They've designed it for these two, uh, personas, um, the figure maintainer and so forth, the storage or it admin on the one hand, but then the user on the other hand, the, the, the proverbial data scientist. And when you look at the functionality of AI data engine, um, what you're seeing is these sort of fundamental, basic, uh, components of data curation, data governance and so forth to tune or to adjust or it to a particular application and use where you can see what the admin is doing. You can see what the data scientist is allowed to do.

And you can see how these capabilities can be really beneficial, close to the close, to the, to the, you know, to the flash close to the medium. And that helps you see that they are limiting what is actually a happy hunting ground for them of product development anyway. I think even limiting it, that's, it's such a lot to do.

I Mean, is that, that's what I meant by happy hunting ground. Yeah. Okay.

They limit it, but it's still this wide range of potential innovation and value. There's so much to do. And I think when I was describing kind of that workflow for AI that they've been describing, that's not, hasn't come to a head yet.

There's not a set of standards that says you do this, then you do this and this. That's all a emerging. And I think they've put themself into like a very good place.

But I think from a, from a storage perspective, there's a lot of background in that. So if a new project is spun up somewhere where you're working, and it's gonna be, uh, this is how long ago I was storage diving, it's gonna be a three tier web application, right? That's what we, I used to do.

So don't judge me. I'm gonna use This example because People still build three tier web applications. I will tell you, yeah, they have the same, same color hair as you though.

But it's gonna have a database, it's gonna have a database component, it's gonna have storage component, it's gonna have, um, server components. And all of those things are gonna have to work in harmony. And you have to think about all the other things.

You're gonna think, will it grow? What do I have to work? What volume am I gonna put it on?

How, how much, um, how busy is it gonna be? How much processing speed does it have to have? Um, where's the data gonna go?

Who needs to have access to it? How do I back it up? How do I restore it?

We're gonna have to put this into our whole disaster recovery plan. All of these things. And that was then, so there's now a whole lot of evolution of, of the field as Becky could currently go into details, all sorts of details on.

But the whole deal is we had to understand the application. We had to understand not only the applications being created, we had to understand all the other underlying components to it. So that's kind of how I take this whole idea of workload.

You have to understand for the, the end application, a lot of what I was supporting then was, uh, medical trials and websites for, to support medical trials. Lots of lots and lots of, um, what's it called? Lots of medical compliance stuff around it, everything, all of that stuff that a storage admin has to know so that we understand that that data is always available when it's supposed to be available.

And I think that's kind of where NetApp's going back to with ai, we've still got this happy hunting grounds. We don't really know what that workload's gonna look like, but they've kind of said, we know this is the storage part of it and we know that our ONTAP can do all of these things to make these things we're hearing data scientists tell us are gonna be part of that workflow relative and available, and we can build things to make it easier for them. So that's kind of where I got out of this.

Yeah, It's like a, a sense of humility almost. Like we know who we are, we know what we know, and we know what we don't know. Yeah.

And we wanna work with those people, you know, we wanna work with you. Yeah. So I understand why they've branded this AI A IDE, right?

Um, but I also keep thinking about all the steps, nearly every step. They talked about, even the things you talked about just now, Gina, they're also needed for analytics. They're also needed for reporting and data visualizations and all of that stuff.

So it makes me wonder if that branding is gonna over limit people's paying attention to it, because it is really the, it's the sum of all the parts that they've focused on ai, but we have a lot of the same things that we need to do for analytics. I think one of the reasons is, is that we learned all those lessons when we went to data warehouses. So the beginning of analytics, and now we're bringing in ai, it's a different group of people and they're starting to realize if you're going from the source data to something else, there's all these extra steps you need to do.

You have to find your data, you have to figure out what it means. You need to store what you found out about it. So I think I'm really happy that they're doing this, but they're also, it's very lean products and they admitted to that.

It's scoped to, you know, supporting their particular services. It's doing the, the minimum of what they need to do for inventory and classification and guardrails and all of that stuff. And certainly they're gonna, you know, add to that, expand the scope.

Um, but I do think there's, there's a lot of competition they have from other providers, like even the hypervisors that have products that do that right now in the, in their cloud services. So I wonder how that's all gonna balance out. Yeah, it's a really interesting point as well.

And I think I, I like that they've started small, Um, yeah. Oh yeah. Because I've been around long enough that going through the, the seven mode to see transition where they went all out was painful for very, a very long time, um, for many customers.

And it's good to see they haven't done that again, and that, that they've started small and they've said, right, this is what we can do now. But as we were talking to them and as we introduced things to them, you could see them going, oh, actually no. Okay.

There's good, good input there. There's there's more ways that we could use this. There's other use cases that we haven't necessarily thought about.

Um, and I, I, I think that sort of analytics bi, those sorts of things are gonna be the ones that change the most, um, when it comes to, to storage admins and, and, and data at the moment. Even, even with the way that AI is changing, I think those things are the, the ripe candidates for the first kind of major changes. Um, and that's some of the stuff I think is really gonna be interesting.

And it actually ties into a couple of the customers I'm working with at the moment, um, where we have been working and looking at how we store all of the different types of data across the systems, because we might want to use some of it at some point. We dunno exactly how mm-hmm. But it's like, get it all there and then just start learning the intelligence of that data and then we can work out what we wanna do it later.

But as long as we've got it and we know we can sort of curate it a little bit, then we can deal with it later. And I think these sorts of systems will help with that. And last year they talked about these, these same topics and they really didn't have a lot to show for it.

Um, and it's really interesting to see like what progress that they've made in the past year. Um, and even going through like the can demo for A IDE, it looks very accessible. Like it doesn't look like you have to be like the top of your industry to be able to run this.

That's very impressive. Yeah. So I, I was just thinking back to the, what we were talking about a couple of sessions ago with, um, how it was gonna gather all the data up and it'll gather all the metadata for all the data in your estate.

And I was thinking about some of the stuff Karen was talking about, um, with, you know, how they're gonna be doing some of this. And I was thinking back to a field day I was at three or four years ago when one of the companies was talking about the fact that they were building a platform to do just that, to suck up all the data in the enterprise and gather all the metadata. And they were talking about it would take days to crawl the entire enterprise and look for changes.

And today we're, they were talking about the fact they, we were gonna be able to do that in minutes or hours as opposed to what I saw just a couple of years ago where, and they were talking about having a fleet of servers to do all this crawling. Um, and now we're doing it just in a couple of quick little appliances that sit there and just get a snap of the data. And the fact that we can do it all generated at the storage layer without having to push it all out to an app server somewhere is, is a, a great thing.

I think I do wonder though, about data gravity. I, I, they, they talked about it, talked about side snapping it. I don't really know how Snap works.

Um, I don't know if it works in a snap. I, I, there there's a, there's a little bit of, um, gravity towards the new hardware. Um, yeah, I think it's definitely pushing you towards the new hardware to a Lot of that.

And yeah, and I, I, I think that that is a critical issue to focus on is, um, you know, you want to, right now you want to get the workload to the data rather than the other way around. Um, and with all this flexibility and everything else, um, you know, architectural decisions that, that create friction time, whatever, um, they, they appear like a risk to me. Yeah, yeah.

Yeah. There's some really, um, key points about that I think where NetApp of leverage really well. Um, it goes into what you were saying about minutes and and hours is their snapshot technology and their snap diff engine, which never really gets talked about much, um, or at least enough, is really incredibly efficient.

And it's been used for backups and other bits and pieces for, for a long time. But the, the way that that works and all the work they've put into that over the sort of 30 years, um, is a really incredible way to then fuel this. Um, yeah.

'cause you get down to the absolute, you know, block level changes and you can scan through, um, you know, petabytes, exabytes of data in, in seconds and say, well, it was only these half dozen blocks that changed. Right? Because they, they keep all of that.

Um, and so while there is that little bit of gravity of, I wanna bring these things in, it's not necessarily that I wanna bring the data there. I wanna bring the data about the data. So it's the, the, the, the snap mirror brings you the metadata, it brings you all the IO tables, it brings you all the differences.

Ah, okay. Yeah. Yeah.

So you're snap mirroring it to get the differencing not necessarily for all of the data all the time. You're getting it to get all the metadata and all the changes. Um, and you, those, those snap mirrors could just be, you know, kind of bend off again after a very short period of time.

'cause you're just, you're pulling all the metadata in via the snap mirror to get it into the engine. Well, yes. When we're using NFS or SIFs, but the S3 engine is a little bit different with Snap Mirror today.

It is based on object by object today it has to be. Yes. It's just, it, it's the way they implemented an ontap.

So, um, I think there's, there could be some challenges there, but no more of a challenge than would be with any other tool. Right. So I'm not too concerned about that.

Um, but to the point of, you know, a lot of the vendors, um, a lot of CEOs have taken credit for, I'm gonna bring AI to the data and not data to the ai. I've heard three of them already. And so, um, challenge with that is that, you know, I might wanna bring AI to the data if I've got one of these a FX clusters in a data center somewhere.

I mean, that's a great thing, right? Um, the problem is, you, you may not be able to have GPUs right next to that thing to do the other stuff, right? And so you may not be able to bring AI to that data or the, um, the, the nature of the tasks that you're doing may require a lower latency than even the two or three milliseconds it would take to get to the cloud region that's adjacent to that data.

Right? It may not be appropriate. So sometimes you're just gonna have to move it out there.

And I think there's gonna be more of that. People think, um, that's, that's my essential point. Yeah.

I think, I think it's not gonna be in risk there that you may not, yeah. You just don't know now. Mean it could be perfectly fine.

I mean, there'll be certain things. Um, and, and also not everything requires GPUs, uh, let's be honest. Yeah.

Um, but I, I just think that there's, there's so many permutations. I think that this is gonna be the core, uh, this is gonna be the core tenet of a, of, of a company's AI data hub, right? And then from that, they've got all the, all the choices in this clouds, especially now that you've got snap mirror with Azure, snap mirror with Google, you've got the ability to move this data all these different ways down to your edge.

So you'll be able to have this core, have your, you know, and do your data pipeline stuff in that sense. And then push it out to where you need to go at, whether it's a neo cloud or it's a cloud, whether it's an edge or, you know, some other place where you do have GPUs. I think there's gonna be more data motion than, than, than not.

But NetApp is the one who they own that space. I mean, they're the ones who have been always better at that space to do it consistently. Um, the S3 thing maybe juries out, we'll see.

But, um, I think they're, they put themselves in a great spot and I think, um, enterprises are gonna need this for their architecture going forward. 'cause it's just, they have no idea where they're going. I came into this, you know, we knew ahead of time is gonna be very AI focused and everything.

Um, but I was hoping to hear more, more references to things like AI specific security needs, ai, threat mitigation needs. Not so much that I expected them to deliver products or services that do it, but I think all people delivering AI infrastructure presentations should be talking about how does this work with threat mitigation? How does this work?

What sort of granularity of security is, and I think part of it was they're coming into it saying, Hey, we've got the whole access to data thing. We've been doing it for decades, so we got that part. But I'd want them to talk about even things like AI ethics and how, and bias and all those things.

Like how are they gonna support those tools or those actions. And I, I wanna hear more of more vendors talking about that. It's, it's a, it's a, a really interesting point that you make.

And I think that that's a corollary to what we were talking about earlier in this discussion. That if you have, uh, this sense of, you know, sort of the sense of self and of boundary and, and and, and partnership and so on, uh, the flip side of that is that sometimes you don't hear the, um, you know, the messaging that people expect you to hear. I mean, I actually felt relieved that I didn't hear a bunch of hand waving about the future of ai.

Oh yeah. I heard a lot of nuts and bolts about supporting the future of AI with partners. Yep.

Which I actually really enjoyed myself. But I, I, I see what you're saying too, that, that, that maybe we need a little bit more specificity on how we're gonna do some of these things. Yeah.

I think we may have got the mo nuts and bolts of how the systems work, but I agree with you. But I, I, I also just agree about the basics. Like, I would've loved to have heard about a customer who, here's where the customer was going with what they were trying to build with ai.

Here's the challenges they had of having to get the data in. So then we had them try out this new thing that we mm-hmm. We created and we were gonna launch.

And here's how they connected their old filers. 'cause you've heard all the questions I asked about, but where's the data? What happens when you have, um, somebody that doesn't know storage, create their own volume?

Where does it actually create it? How many copies is that making? How are you keeping track of all that?

Like let's talk about, 'cause from a storage perspective, that stuff all matters. That's part of understanding how big it's gonna be. They didn't talk anything about kind of what's in the pipeline.

Here's a customer that went through all of that with NetApp. I would've loved to have heard that. Yeah.

I think, um, to answer a couple of these questions that just come up is that I think there was a little bit of an assumption around the baseline of knowledge, um, because of them being at Insight. And so, um, they have, uh, DataOps toolkit, right? Which plugs into JU Labs, uh, Jupyter Notebooks, and you know, it handles all those things for you.

Um, and there's the same with what you, you were saying about the, the data gravity, um, and moving things around. There's already parts of the toolkit that would, um, and, and I think this is where the A IDE stuff will come in really handy is that yeah, I may still need to move some of that data around, but they've got the most efficient way of identifying and moving just the data I need. Mm-hmm.

And so that's something that NetApp's got that they can play on really, really well. Um, they had A tool a long time ago that they were, they were kind of teasing a long, long time ago, and then it just got killed. And so I'm glad to see them get to market with this.

Yeah, exactly. And then, so on, on your part, I, I don't think I would necessarily like them to see, see them have spoken too much about the ethics Exactly. Some of the other things.

Mm-hmm. But what they probably could have talked about a little bit more, and I, I made some assumptions 'cause they didn't cover it, but I made some assumptions around how the policy engine works and I'm, I'm assuming it's based on F policy. So they've got this entire framework and API, um, that partners already integrate into.

Mm-hmm. So you could just have new integrations into F policy and A-A-I-D-E that could handle that, um, bias and governance and other things. And they, they probably could have done a little bit better job talking about that, what the ecosystem is.

'cause they were very good at saying, this is where we fit right now. Um, and I think yeah, there could have been a bit more of a, this is where the other players fit and how we build it and how We integrate with them. Yeah.

Like that's what I wanted to hear. Yeah. Yeah, exactly.

Well, just as they were providing an, this new a FX thing provides an aggregate, a, a virtualization layer, abstraction layer above the aggregates. What they're providing here is the abstraction layer above those primitive tools to make another layer of tools that are more consumable, but are still building blocks to be used by others to build other things. Right.

So, layers upon layers, Turtles all the way down. Does anyone wanna have a last, uh, last word or should we wrap? Yeah, I will.

Um, I, I, so the, the demo you're saying they should have done where, you know, this, this is where they were going, and then we plugged our thing in and, and it was, I love that idea. I would love to see a demo like that or just talk about that at a keynote. And it doesn't need to be long, you know, two minutes of, but I think that would be super compelling and really would highlight and show off the bells and whistles of the new platform.

And I, I, I, I now want them to do it. I will just point out that the platform was literally released today. Yeah.

But it was preview. And so, um, I, It was in preview. They had, they had people on it.

So how about this? Why don't we, why don't we say this next year, NetApp site, um, that's what we want to hear, because I agree with you. I think that would be very, very powerful to set.

Because, you know, as we heard again and again and again, and as we've heard at literally every conference related to AI this year, that m MIT study mm-hmm. And the internal study here and so on, that, that companies are trying to do AI and they're failing. Yeah.

And, and, and I Was this and we can make it better. Exactly. And, and, and a vendor who can come in and say, companies we're trying to do AI and are failing and they plugged our thing in and it worked.

That's the story I think that all of us want to hear, because I think all of us are hungry to hear it worked. Yep. And unfortunately, we've heard a lot of it didn't.

So that's the challenge, I guess, for next year. So, um, thank you for, for being part of this. Thank you all for giving your time and, uh, your thoughts to, uh, during the tech Field day presentations as well as, uh, during this round table discussion.