36. There’s a Gulf Between Storage and AI – Tech Field Day Podcast

When it comes to storage supporting AI infrastructure, it's not at all obvious that the storage companies know how to support ai. That's the topic for this episode of the Tech Field Day podcast, where I'm joined by Curtis Kempel, Brian Boin and Rohan Pur here at AI Data Infrastructure Field Day in Santa Clara. Is it, uh, time to hope for the best that the storage companies really understand AI or well, maybe they just don't.

Welcome to the Tech Field Day podcast, where each time we meet, we discuss a single concept or premise related to enterprise technology. This podcast is produced by the Tech Field Day Group as part of the Futurum Group, and is today being recorded in association with our AI Data Field Day event here in Santa Clara, California. I'm your host, Stephen Foskett, and today we're asking ourselves, how wide is the gulf between storage and ai?

And we're suggesting that maybe it's further than people think. But before we begin, let's meet who's on the podcast today. Hello Everyone.

This is Rohan Puri. I work for Samsung and I'm a file systems developer. Uh, hey everyone.

I'm Brian Bodin, I'm the managing director of Data Glue It, and I work in data analytics and visualization. Hey everyone, I'm Curtis Kempel. I'm a senior director of developer relations at Slack, and I focus mo mostly on the application layer.

So we've been hearing a lot about data infrastructure this week because, well, you know, that's the topic of this entire event as well as ai because of course, that's also the topic of this event. As we talked about a few weeks ago on the podcast, data infrastructure is not just storage, but today we're gonna really take a look at the connection between storage and ai. How close is it?

Is storage really ready? Is storage actually meaningful, a meaningful differentiator when it comes to ai? What's the deal with storage and ai?

Uh, Brian, I'm gonna start with you because you suggested this premise. So talk to us a little bit about the gap between storage and ai. So I think, I think this is a super interesting topic because fundamentally both things are extremely technical, right?

But we really are at polar opposites here, and I don't think a lot of people recognize this out in the wild. I think that, uh, you know, when when you think about storage, you think about very low level, you think about disks, and for me this is a different concept because I'm a data guy. I come from the analytics side of the fence.

And I think that is a point worth stating is that when you come from that side of the fence and you talk about AI and LLM, which is a bit more in like in the wheelhouse, you maybe don't appreciate that there is, I wouldn't say it's a chasm, but I would say that we have a a ways to go. And I think it's been really interesting to hear vendor, multiple vendor perspectives over the last couple of days about where we are in that process because I'm not sure if it matches where we think we are and the vendors think we are. So, I mean, that's my initial thoughts on it.

I think we still have a lot of ways to go. It feels like we're at the start of this journey, Steven. So Yeah, I think that that's abso absolutely true and it's, uh, I mean, it's a truism that people generally don't know what they don't know.

And certainly a lot of the companies in the enterprise storage industry are rapidly pivoting toward providing AI solutions. And yet, and yet, uh, I think they are aware that they don't really have in-house ai, you know, expertise. And so what we've seen is a lot of them are hiring people who focus on AI applications, who focus on data, um, data management, real data, not just storage euphemism data.

And yet, I wonder if the message is getting through, what's your perspective? Yeah, it's funny that you brought that up because we actually see a lot of roles opening up and you know, I'm very closely related to this, uh, with developer relations, but like you, you're kind of like AI specialists in-house, right? Uh, and I think that that is very much a symptom of the, uh, some parts of tech being a bit slower to jump onto ai.

And so, especially when you look at storage companies, like to me what it feels like, what a lot of them are saying right now is essentially just the differentiator is the cost, right? Uh, the efficiency. But if every storage, uh, uh, you know, company is saying that exact same thing, then it's not really a differentiator anymore.

Like, everybody just has the same thing. And, and now I'm just looking at pricing tables and cost tables, right? Uh, and so I think to circle back to what you're talking about, until a lot of these storage companies can better understand the role that they need to be playing in a, in ai, it's gonna be we are storage slap AI onto us 'cause you're here.

And so I don't know that that's gonna bring new people. Like, I don't think that's going to attract people who are know they're going to hit AI scale and are going to seek storage solutions. I think it's more like a stop gap to empower those who are already invested into their systems.

And so for me, I'm waiting to see these, uh, um, file storage and infrastructure companies start to push further out of the bounds of what they've normally done and start to offer a lot of those like affordances and quick wins around the things we know have to happen around AI training data or the insights you have to have around specifically AI data, uh, and storage. And I really didn't see almost any, any of that. I don't feel like over the last two days just hints of it.

So Yeah. So my take on this is that infrastructure by itself is just one piece of the puzzle. And it's not just enough.

So storage companies and storage people just saying that, okay, we have fast infrastructure, performance infrastructure, cost effective infrastructure, that's all fine, but how it is getting used as part of solutioning is where the gap is. I feel. So this, this is, and because storage people also do not understand what building a real AI application takes.

So that's why solutioning is very important around this. Understanding how it is being used is very important. There are only two directions that I see that is happening is a lot of storage companies are tying up with, uh, GPU direct because they want to Yeah.

Serve AI applications. And that is where their AI knowledge looks to be limited in only that sense. Yes.

But it has to go further right? Till the application and not just because GPU is still part of infrastructure, it's not outside. So Yeah.

Yeah. I I think you used the word solutioning there and I think that's an excellent way to summarize it, right? Because the solution talks about the end-to-end paradigm of storage at one end and AI at the other, right?

And the reality is that there's, there's probably several layers of meta and extraction that's, no, not no Facebook pun intended, but that sit in the middle of that abstraction. Yes. And It's, It's a hard problem to solve because if you've been a focused storage company to just jump into that space.

And it's one thing to understand technically what's required for the technical limits of, you know, RAG and LLM, but it's another thing entirely to be able to push that as an make it applicable in the real world. And I think, I dunno what you think, Kurt, but I think that's the, the point that we're at right now. Yeah.

Yeah. And I, from what we've seen over the last few days and what I've been thinking about, like, it really feels like for me, the biggest opportunity for folks in storage space to have quick wins are in a couple places. One, you already have all of the data and there's AI tooling available to you outside of what you have to build in-house.

You put those two things together and you could do really powerful insights to help people better understand how to use their data in AI workflows, right? So that's number one. And you don't have to build much anything else outside of what you want.

Second is, we, we've called up to this a couple times, like talking about feeding directly, uh, into graphics cards is egress and ingress, right? Like breaking down data silos, making sure data is exactly where it needs to be, and that process is as easy as possible. Uh, and so we could start to build application on top of that, right?

Like making sure it's so simple for me to connect data sources from all my data lakes, from just, I mean, the millions of sources that we're going to be pulling data from. Uh, and then lastly, again, I still think there is a lot of space, like you said, in when we're training metadata, doing all of the snapshots, like doing all of that work, it all feeds back in there. Like I know there are tools that focus on ML pipeline and generation, but anything that is around like the storage of data, I would be looking to build solutions and it's, or products around that.

And I don't think we're seeing that coming to it. I feel like all of these companies are still just seeing the hardware as the product in this space, and they need to start thinking about building more specific products for AI to match the hardware. I think that's like what I would, I would surmise it as So, so you guys are coming to this from outside the storage industry, well apart.

Sorry, Rohan, the other two are coming from outside the storage industry. Uh, you don't have a lot of background on the differentiators between various storage platforms and so on. Um, did you hear this week or just generally, have you heard, uh, from storage companies any compelling reason to pick one storage solution versus another?

Or do you feel like, um, as it seems on some of the charts that you can basically just swap in a different storage, uh, on that box on the layout and it'll still work? I, I think there's a discussion about maturity here and there's also a discussion about do you know what your place is or where it is right now? And I think, I think the thing that some, um, storage vendors are struggling with is how far do we take this solution and do we build the solution?

And how do we know when to partner and how to partner? And I think a lot of that just falls into the knowledge gap of what is an LLM, what is rag? And although we understand the technicalities of it, how, how far do we try to provide that as a solution ourselves and where, where do we draw the line?

And the feeling I get is that some vendors are still working that out, right? I think there are a few that, that have done pretty well, but there are also a few that are like, gimme some, gimme some, gimme a hand here. I mean, Yeah.

Yeah. I did say one thing I liked, uh, which was, uh, oh, the name who presented as escaping me when they broke down tco. Mm-hmm.

Right? That was, That was Soine Thunder. Soine.

Thank you. I really appreciated that because from coming outside of the storage space, I don't really understand the cost and the breakdown what's there. And so if we're at this stage where that's a differentiator mm-hmm.

Making that readily available to me and very clear and easy to understand is largely gonna push me to want to pick that product because I know what I'll be able to do. I can allocate properly, I can trust that these metrics are there. So technology can be great.

You can tell me you're saving me a bunch of efficiency for X-ing this or doing that, but like, if you could break that down for me physically in a dashboard and I can do predictive things and set alerts and limits on, you know, expenditure and stuff like that, right. That I think would also probably be a pretty big differentiator. It's interesting that you picked the lowest level product that we heard of.

I mean, they literally make SSDs and yet their presentation resonated with you because they were talking your language. Absolutely. Right?

Absolutely. Yeah. I, I, I think that cohesion, and again, just coming from the different angles, that that storytelling part is natively part of the, the data and analytics and visualization industry.

You cannot get anywhere without the story that takes you along the journey. And I think I I, I, I'd love to hear what, what Rohan thinks about that from a, from a data storage angle, but from a visualization side of things and an analytics side of things, yeah. You have to be able to net that together.

Right. And I think Soine did a, did a great job of that as well. But I'd love to hear what you, what your thoughts were as Well.

I I I feel there is a bigger problem here in the sense that there was also a topic that was charged about data silos. Mm-hmm. Yes.

So if we, if we just see how many sources of data we have, we have databases, we have data warehouses, we have file systems, block object, traditional storage companies, appliances mm-hmm. With so many data sources available, how to make sense, what is ai, right? It is about making sense of this data Yeah.

And predicting future with this data. Yeah. With patronizing it.

Yeah. How, how to, how to group it together. Because what I see is storage, traditional storage companies and public cloud vendors and warehouses, these are all so disjoint with each other.

One good thing in warehouse and data, uh, like, uh, databases world is happening is that they're decoupling individual layer layers of databases with open formats. Yeah. Like open table format, open storage format.

This gives ability of, to all the vendors to connect together at least on that open platform. Yeah. And then applications can only work with those open platforms.

Yes. Something of this sort is not happening very much in storage. Well, we do have protocols.

S-M-B-N-F-S people talked about it. Mm-hmm. But those are more of access protocols.

Yeah. How to consolidate data from different vendors with some open format on table and then communicate. I feel like that was a little bit of what we just saw at the end, which is what they wanted.

Like you take all of your data and you put it into, uh, But the problem is every, your storage and every vendor says that only like we are the we, uh, give ev give all of it to us. Absolutely. But that's not the reality.

No. In real world. It is, it is going to be distorted.

It is going to be at different places. Now the question is, do we have any software that works with all of it and make life easier for application developers? Yeah.

Or tell them, okay, now you have this vendor, that vendor cloud on-prem something we need Right. To connect all these things. I Agree.

It, it is funny 'cause um, earlier when we were chatting during, one of the things, you know, it was like some, like some ideas and systems, like mental models are starting to take shape. Mm-hmm. And a very big one that came up was centralized versus decentralized.

Yeah. And so, like you look at some of these solutions, just like you said, and they're like pulling it together, but Mint IO is going in the opposite direction. Iona said, no, no, no.

We've built the open source layer, the tooling so that we can be anywhere. Your data is instead of your data having to come to us, and then we can move it around for you and help access it. Now you can put your data there as well, but they offer just the open source tooling that you could tie into data that's pretty much located anywhere.

I, I think also it's worth just drilling down when we talk about data that's a highly all encompassing term, right? And even even coming from like an LLM and a rag and analytics angle, unstructured versus structured. So you're, you're, you're very structured file data that sits in databases and is easy to pull out is very different from pulling data out of a PDF or a Word document or a PowerPoint or your whole SharePoint site.

And I think the variety of what we saw over the last couple of days, and the overarching message is that storage companies have a ways to go to understand the differentiators there at, at LLM just doesn't sit on the top of anything. It has a specific purpose. And we've not even touched on that yet.

Right? Yeah. An LLM is not just to get data for everything and ask all questions, it's to, uh, in the case use cases that we have been seeing, it's much more to do with the fact I want to see a limited set of my structured or unstructured data.

I want to install that locally. And for a storage vendor that's, that's pretty tough because that, that could be an on-prem, it could be a cloud, it could be hybrid. And we've seen all kinds of those factors coming in over the last couple of days.

That's Circles back to what I'm talking about of where I feel like, you know, these, uh, storage companies can really start to push away from just being storage. Meaning that like, okay, we know that when we retrieve certain files, like unstructured data, things need to happen to that, right? Data prep needs to happen.

What if that happened for you along the pipeline? Right? Like things like that, you know, I don't know exactly what the developer experience and user improvements and efficiency improvements are, but I think that's the only way to start playing.

Mm-hmm. You know, and you have to have something that ties you directly to AI that makes my life my job easier. See, I think that's where standards should come into play.

Like, for example, storage does have standard for drives. Like we have NVME spec and all those things. Uh, we have S3 also, but, but then we need standards at one layer or couple of layers above, like, and then the application should deal with that standard.

And that's done. Like if, if you guys want to work with vector databases, APIs and connectors are different. Traditional databases Oh yes.

Is different. Yeah. We work with files.

It is different. So I, I I think we are, although it's maybe doesn't seem that way because of the explosion of chat GPT over the last year, we are still at a very early phase of this process from an AI and LLM perspective. And I think standards come with time.

Right. I don't think standards start to just happen overnight. So, right.

For me, especially for storage companies that want to reach actively into that space, they need to niche down into something and they need to latch on. It's not enough nowadays just to pull, uh, you know, a diagram that says, here's a rag, here's an LLM and let's give you a chatbot interface. And you know, chatbot's kind of been the moniker word a little bit of the last couple of days.

It sure has. But again, it's a paradigm of are we really seeing what's needed to build that middle middleware arc? I shouldn't use the word middleware, but that, that traversal architecture that sits in the middle.

And I think that's, you know, we're getting there. We're, we're seeing signs of that, but it's only, it's only seeds right now. Right?

Yeah. Yeah. Would You Agree with that?

Oh, yes. I absolutely agree with that. Uh, we're in the, especially from storage to like just these, I mean, somebody had a great slide that showed just the exponential growth, and I'm gonna say it, data growth, the size of data.

Did, you know, data is growing, data is growing, the data is growing, um, but it really is. Right? And, and, uh, I don't think anybody is really prepared to handle data of that magnitude and run workloads of that magnitude.

Um, I, big companies are doing it because they have the engineering power and know-how to throw behind that. So I think if any of these companies could start making that more readily available, like to other people, I think, I think that's what they're trying to do. And we did hear a lot about how these systems scale and how they can scale IO and they can handle massive numbers of file access.

But, but it still seems like there's, to me, it seems like there's kind of two ways in which storage companies are seeing ai. Either they're seeing it as we can have our system plug in specially to have special support for AI applications and training and so on, and, and we can become, you know, more integrated. Or they're saying, Hey, your data is here.

Mm-hmm. Don't you want to be able to access that with your LLM? Yeah.

Yeah. Wouldn't it be great if you didn't have to move your data all around? And, and really those are two fundamentally different, you know, perspectives on how to service ai.

And we actually heard from, um, you know, the infin at, uh, guy who was presenting in, in, uh, kind of after hours, he was saying, you know, look, as a storage company, we have to decide what, what is our strategy gonna be when it comes to ai? Are we gonna be storage for ai? And he, he didn't say this exactly, but this is kind of what I heard.

Are we gonna be storage for ai or are we going to be storage that can connect to ai because this is where your data is. Exactly. I I, I think that's an amazing point.

And I think HPE did a great job of that yesterday of like defining the way that they wanted to deal with ai. They, they, they understand that they want to containerize and they want to be portable, and they want to be able to connect via different ways, and they want to make it available in an engineering way from an analytic standpoint so that you can pick that up, right. Mold a model and push it somewhere else.

And I think that kind of narrative, I think is quite compelling given the infancy of this market at the moment. I mean, that, that's, I think that was one of the aggregated platform as well. Yes.

With connecting to a lot of different types of data, model storage and everything. Yeah. But my, my, my concern with that is that it seemed, even for, from a data analytics perspective, it was pretty technical.

Yeah. You would need to be a data engineer that knew quite, quite a lot of storage and quite a lot about, but it was defined. And I like that.

And I maybe, maybe there's a bit of give and take there. Maybe you can't have that containerized or we're not in a place where you can have that containerization without being a little bit technical. And maybe that is the trade off of where we are in this space.

I feel eventually we'll have more abstractions on top of that also, but this is the first step to towards getting there. I I, I think no one really knows how many layers are gonna be in the middle as well. And I think, I think that's the tough thing.

It's just you don't know what you don't know, what you don't know right now. And the, the companies are trying to work out those, those layers at the moment. And it's not easy.

It's not easy. It's not. And like, you know, we see a bunch of companies popping up doing, focusing just like on the AI and ML workload portion of this.

And then you've got data companies jumping in. I, what it feels like right now is like we're very much siloing along technologies agree for ai and in the history of tech, it seems like that's generally how it goes when something new comes out. Yep.

But I think in the future we'll start to see a blend. We'll start to see that those we verticals change and like what used to like be locked off into certain areas will start to be a single product part of your, your architecture and pipeline. Right?

Right. Um, and we see that happen all the time. And so for me, I'm just curious, I think that's why I'm so bullish on seeing these, uh, storage companies not go the I'm a connector connect me route.

Mm-hmm. Um, because then you're just yet another data storage. Um, yeah.

I'd rather see them start to push the boundaries of saying, no, no, you don't need to also run what space, hammer, sledgehammer, whatever it was. Like something Hammer space. Hammer space.

Thank you. Uh, you know, to the point where you don't need to run that. Like, I want my storage solution and my AI and model training solution and all of that stuff, ideally in one One.

Yeah, that's right. Like I, yeah, So we only heard from one company that actually does AI specifically and that was Google Cloud. Yeah.

Yep. They have their own AI models. Yep.

And they do, and they're doing cloud. Absolutely. What, what was your take on that angle of it?

You know, I mean, it's the only company of, of this bunch that does both AI and cloud. Like absolutely. They're in those two spaces.

Yeah. I think, uh, their story was a lot closer together. Um, but it's still felt like you have Google Cloud and then you have Google ai and like the, the, the bridge connecting them was definitely thinned.

And like, I've just been through this, like Slack was acquired by Salesforce, and so it's like when we started to bridge those two platforms, it was a very narrow and thin bridge. And those can get much stronger, uh, over time. But like, you know, when I started diving into like, well, oh, oh, this makes sense and you must be able to offer like amazing insights and do all this, I expected to see that in product, not I go to a report board, I click a button, generates me an insight like, um, to have such a closed system and not as tight of a product integration.

But I think that they're doing what I'm saying. Yeah. Which is you can store your data with us, you can run your workloads with us, you can get insights to that with us.

And I think like that is it like, it, it's AI platforms at the end Of the day. I think one, one problem, I agree to your point, I think, but one problem I saw in their, uh, portfolio is that they have too many things. Oh, yeah.

Like even in, that's even in storage, they had like 20 different offerings. Absolutely. Which ones to choose which ones are good enough, how it maps to what applications you want to run.

I think those solutioning, and again, those reference architectures are, is, is probably missing at this point, but eventually will get better at that. This Is like a rock and a hard place situation for Google, I think because they have a lens on both sides, right? They have the storage angle, they have the cloud angle that sits in the middle, middle, and they have a bit of a, a finger in the, in the BI piece as well.

So tying that together while giving the versatility and variety of solutions is a very tough thing to do. Exactly. The the LA the last thing I'll throw in there is that we haven't even touched on BI vendors and how they're approaching the LLM from a, a, a position of knowledge, right.

And a position of power. Yeah. And how they are starting to delve.

Maybe it's happening under the hood, right? But they're starting to deal with these storage type of issues. It's just, you handle it a different way, right?

You point to a file system and it all gets ingested. And we, we all talk about white box and black box. So I think, I think that's another angle to talk about, maybe on another field day topic.

But I, I think, I think the way that bi vendors in general are trying to solve this problem is a very interesting, um, uh, flip side to what we're talking about here, because we're talking about storage and driving up the way, but there's also the, a very down, very strong downward compression that's happening. Yeah. And Google, certainly of the vendors that we saw is the closest to matching that, that use case right now.

But, but Microsoft would be incredible to hear from them about their strategy. Yeah. Because they're one of the big biggest BI vendors.

I, And you, you know, I, you know, I I'm happy to say I have some affiliations as well with, you know, I, I know that there's other tools out there that do that, like click and they have rag models that, that can do that from a push down perspective as well. Yeah. So like we, we still don't know what's right and we still like, comparing these different paths is something that's going to take Well, I I still some time.

Yeah. It will take some time. I still see right now, everyone, if you are listening to this new deal with storage or honestly anything with ai, yeah.

You already have ai, you know how to use ai like build AI into the products that you are trying to use to serve folks. Like if I'm dealing with storage, like I want to know way in advance when I might hit certain usages. I want to know that by the way, I'm training my data and running it through your pipelines, that I'm not actually doing this the right way or using the right database.

Mm-hmm. Um, you know, I want things like that. And honestly, any company can start doing that today.

That's like, I Will point out that a lot of companies are doing that, and we've actually heard quite a lot about companies that are using AI in their products, but they didn't present that here at this particular event, mainly because I told them that this was more about how are you serving AI and how are you supporting ai? So for example, you know, pure Storage, you mentioned them, uh, they absolutely are using AI as part of their management software. I know HPE is okay, I know that, uh, Google, uh, I mean, they, Google Cloud actually showed us some of the ways that they're using ai.

So yes, we're, there's definitely companies doing that. And that's yet another angle, yet another aspect to this. So we we're actually kind of running outta time here.

Before we go, I want to give you guys a chance to sort of sum up your opinion. Let's get back to the premise that there's a wide gulf between storage and ai, and I wanna give you a chance to sort of reflect on that and respond to that. So, so, so go ahead.

What do you think? So I feel integration solutioning is the way forward. And it seems like a lot of the companies, the storage companies especially, are working on doing that.

Partnerships and integrations. Yeah. Um, I, I, I want to see a lot more and definition and niche.

I think those are the two big things for me now. I want to understand where companies are focusing and where their efforts are, rather than trying to solve all problems for all things. I think that's the biggest thing I've taken outta the last couple of days.

I think for me, education, I would love to see these storage companies bring more and more people in-house, talk to more and more potential customers get nitty gritty and just, it honestly, it feels like they don't fully understand the AI space. And, uh, until you do, it's gonna be very difficult to play. And especially for something that plays such a critical role in the AI space, It seems to me like the companies, the storage companies understand that point and, and they're, they're, they're hiring people who can talk the talk and they're trying to get those people out there.

But it is an upheld challenge. And my fear has happened when storage and data or storage and cybersecurity tried to, to come together, that there would remain a golf within the company that they would have, you know, well, she can talk the talk. So she's gonna go out there and talk about cybersecurity with the customers, and then she comes back in and pulls her hair out because the, uh, the developers inside the company who are making the product don't, don't know anything about cybersecurity.

You know, and I worry about the same thing happening with ai. Absolutely. You have to empower these people to actually change the product product planning in order to have a better product.

Not just go out there and talk to talk the talk to customers. No. And this means you need to be hiring like AI specialized people into very critical and key roles within your system.

Like leading up product, like leading up customer. Yeah, that's right. They do something about it and they have to be empowered to do something, which means leadership.

You need to be hiring people in leadership who have this experience, who have the ability to make sweeping changes that will be needed in reeducation, and it's gonna have to happen, happen top down. Well, Before we go, I'm also gonna give you guys a chance, uh, tell us, uh, where we can interact with you online. So, so tell us your name and tell us, you know, what's your social media platform?

Where will we hear more from you? Uh, my name is Rohan Puri, LinkedIn, primarily I use and Twitter, uh, Rohan Uri. Well, the, um, I'm Brian Boin.

You'll mostly find me on LinkedIn. Um, I'm there most days, uh, and you can find me with Click Luminary on LinkedIn or just find my name. There's only one Brian Boin on there, so it's easy.

Oh, Well, there you go. Yeah. Uh, and I am Curtis Keel.

com, uh, or on LinkedIn as well slash Curtis Kempel. Uh, and if you really wanna find me on Twitter, you can, that's under Digital Vandal, but it's really not gonna be this kind of content that's just like everyday life, uh, you know, trash posting. So if that's your thing though, come along.

And as for me, you'll find me at s Foskett on most social media networks. I'm pretty active on LinkedIn, but I'm also posting on Mastodon Threads, blue Sky, Twitter, et cetera, et cetera, et cetera. And of course, hosting the Tech Field Day podcast, the Gestalt It News Rundown and the Textron Gang every Tuesday.

So, thank you very much for joining us for this episode of the Tech Field Day podcast recorded live in Santa Clara as part of AI data Infrastructure Field Day. If this piqued your interest on the whole topic of AI data infrastructure, check out YouTube. By the time this gets published, all of our videos will be posted.

Just go to YouTube slash Tech Field day and, uh, you'll be able to find those posts. com on the Techstrong websites, including Check Strong TV and, uh, other, uh, Futurum group outlets because we are a part of that. Thank you for joining us, and we will see you next Tuesday.