2985: How SSDs Are Powering the Future of AI

[00:00:01] [SPEAKER_00]: How is the evolution of AI dependent on advancements in storage technology?

[00:00:07] [SPEAKER_00]: This is a topic I want to explore today because I'm going to be joined by the Senior Director

[00:00:12] [SPEAKER_00]: of AI and Leadership Marketing.

[00:00:15] [SPEAKER_00]: Add a company called Solid Arm, an air leading global provider, innovative flash memory solutions

[00:00:20] [SPEAKER_00]: and NAMD flash memory solutions and as AI and enterprise workloads demand faster and more efficient

[00:00:28] [SPEAKER_00]: data processing. Solid Arm State of the RSSDs are right at the forefront of enabling these

[00:00:35] [SPEAKER_00]: advancements. So I want to explore the importance of SSDs in accelerating AI, how Solid Arm

[00:00:42] [SPEAKER_00]: is collaborating with customers to deliver on parallel storage solutions and dive into a world

[00:00:48] [SPEAKER_00]: where cooking edge storage meets groundbreaking artificial intelligence. The cost of hosting

[00:00:54] [SPEAKER_00]: a daily show for 140,000 monthly listeners can be significant and I've like the take-a-moment

[00:01:00] [SPEAKER_00]: to thank those who make it possible for me to keep delivering this content every day table.

[00:01:05] [SPEAKER_00]: And I also want to talk about the fact that legacy DRM failed to securely enable external collaboration

[00:01:11] [SPEAKER_00]: on sensitive files and I think it's important to recognise organisations in this digital age.

[00:01:17] [SPEAKER_00]: Face a somewhat risk, trust contradiction. Yep, they must share content with untrusted third

[00:01:23] [SPEAKER_00]: while also protecting that data. So it's time for a more modern DRM solution, one that

[00:01:29] [SPEAKER_00]: solves this dilemma but without compromising security and productivity. So collaborate as

[00:01:35] [SPEAKER_00]: imagine editing files externally without losing control. Streams, zero latency video renditions

[00:01:41] [SPEAKER_00]: to authorize users, but without any actual file transfers needed. The co-author can view it remotely

[00:01:48] [SPEAKER_00]: while you retain full ownership. Your files stay protected in cart work secure, unclave.

[00:01:54] [SPEAKER_00]: Ultimately they never leave your environment. So you can stay goodbye to data, leakage risks

[00:02:00] [SPEAKER_00]: and experience, seamlessly editing across all file types, not just native applications without

[00:02:05] [SPEAKER_00]: any plugins required. So say goodbye to deployment headaches, file transfer risk collaboration

[00:02:10] [SPEAKER_00]: barriers and all those productivity constraints and experience a more modern way to collaborate

[00:02:15] [SPEAKER_00]: on sensitive content, sacrificing control or security and you could do all that by visiting

[00:02:20] [SPEAKER_00]: kiteworks.com to get started. And with my thank yous out of the way, I'm now officially

[00:02:26] [SPEAKER_00]: excited to introduce you to today's guest. So buckle up and hold on tight as I beam your ears.

[00:02:34] [SPEAKER_00]: All the way stateside work to day's guest is way in a join me.

[00:02:39] [SPEAKER_00]: So a massive welcome to the show. Can you tell everyone listening a little about who you are?

[00:02:45] [SPEAKER_01]: And what you do? Sure, Neil. Thanks. Thanks for the opportunity to be on your show. I am Roger

[00:02:51] [SPEAKER_01]: Karel. I work for Solid-I'm is a US based but with a global presence storage company for data centers.

[00:03:02] [SPEAKER_01]: And specifically in terms of what I do, I am the senior director of AI and leadership market in

[00:03:08] [SPEAKER_01]: Solid-I'm. So what does that mean? It means the team kind of develop solution level value

[00:03:13] [SPEAKER_01]: propositions for our storage products, not just across the AI data pipeline but really for

[00:03:20] [SPEAKER_01]: all data center workloads, whether it's in the cloud or whether it's enterprise that our

[00:03:25] [SPEAKER_01]: drive support and then diving in a little bit more in terms of actually what that means that we

[00:03:30] [SPEAKER_01]: deliver again. We come up with a solution value propositions. We drive influencer marketing programs.

[00:03:37] [SPEAKER_01]: We manage product launches, product campaigns and kind of lead the development of our

[00:03:45] [SPEAKER_01]: kind of technology and thought leadership narratives. There's a pleasure to have you on the podcast today.

[00:03:50] [SPEAKER_00]: There is so much hype around AI right now. We're already seeing in the last few weeks and months

[00:03:57] [SPEAKER_00]: that everyone from Samsung to Apple are ensuring that the next generation of smartphones

[00:04:02] [SPEAKER_00]: are ready for AI ready to handle that. So I'd love to set the scene for our conversation

[00:04:07] [SPEAKER_00]: today as we move into this new era. How does people listening here in about Solid-I'm for

[00:04:12] [SPEAKER_00]: the first time? How do your solutions specifically enhance the performance of AI applications?

[00:04:18] [SPEAKER_00]: And also what role do they play from the call to the edge of enterprise operations? Because

[00:04:24] [SPEAKER_00]: this is the stuff we don't talk about enough sometimes. I think absolutely right, we're not talking

[00:04:29] [SPEAKER_01]: about it enough. I think right fully so the conversation up to this point has been largely focused

[00:04:36] [SPEAKER_01]: on an Nvidia for obvious reasons and specifically their GPUs. But it feels like now the market is

[00:04:45] [SPEAKER_01]: just kind of beginning to get their arms around, understanding that storage plays an incredibly

[00:04:51] [SPEAKER_01]: important role. And as you said, an important role all the way out from what we call the core data center

[00:04:56] [SPEAKER_01]: to the mid tier infrastructure and all the way out to the edge. And why does storage matter?

[00:05:04] [SPEAKER_01]: Well, you know, AI gets, you know, stating the obvious AI gets better with more and more data.

[00:05:10] [SPEAKER_01]: And I don't want to bore you with how much training data sets are growing in size year-on-year

[00:05:17] [SPEAKER_01]: in the model sizes are growing year-on-year but the amount of data feeding these model developments

[00:05:24] [SPEAKER_01]: and then doing the inferences is just really growing in order to magnitude every year.

[00:05:30] [SPEAKER_01]: And that's going to get even, I guess, more data intense when you talk about multi-modal AI.

[00:05:38] [SPEAKER_01]: So multi-modal AI is when you're combining things like text, image, video, audio all into the same

[00:05:48] [SPEAKER_01]: AI model. And as we've kind of alluded to these compute resources to run these,

[00:05:56] [SPEAKER_01]: to do the model development and then infer the results at the end of the pipeline,

[00:06:03] [SPEAKER_01]: they're incredibly expensive on the front end. When you talk about training, I don't think

[00:06:07] [SPEAKER_01]: you can acquire a DGX H100 server from, from Nvidia for under $300,000 U.S. dollars.

[00:06:14] [SPEAKER_01]: So you don't want those incredibly expensive resources sitting idle.

[00:06:22] [SPEAKER_01]: And as you alluded to, in these workloads are moving to the edge to improve experiences,

[00:06:28] [SPEAKER_01]: reduce costs, reduce network traffic, et cetera. And so what does this environment mean?

[00:06:34] [SPEAKER_01]: What does this landscape mean to storage it means that you've really got to move a massive amount of data

[00:06:40] [SPEAKER_01]: and efficiently store access at speed seemingly everywhere? And this is where high performance,

[00:06:49] [SPEAKER_01]: dense, highly efficient SSDs come in just a quick comparison to kind of the legacy storage

[00:06:57] [SPEAKER_01]: that is running a lot of these supporting a lot of these AI workloads, which is HDDs,

[00:07:03] [SPEAKER_01]: depending upon where you are and what we call the AI data pipeline and SSD,

[00:07:09] [SPEAKER_01]: can be 10x to 4,700 x faster than an HDD. A high-cap SSD can store four times as much data

[00:07:18] [SPEAKER_01]: in the same rack space as an HDD while consuming almost three x less power per terabyte of

[00:07:28] [SPEAKER_01]: data stored. So just incredible gains when you talk about the workload intensity of AI and

[00:07:39] [SPEAKER_01]: the need to keep these very expensive resources fully utilized. Wow, so many powerful stats in that.

[00:07:47] [SPEAKER_00]: Here we are talking today in August. You've got a new product launching and for anyone

[00:07:53] [SPEAKER_00]: that missed the announcement today. Can you tell me a little bit more about the launch?

[00:07:56] [SPEAKER_00]: I also have a different choice itself from existing solutions out there in the market.

[00:08:03] [SPEAKER_01]: Yeah, sure. So we're really excited that we're launching a new family of products. It's

[00:08:13] [SPEAKER_01]: PCIe Gen 5. So these products kind of fit into what we call our D7 family and these products

[00:08:23] [SPEAKER_01]: will fundamentally deliver the highest, not to get too geeky here. But in the storage industry

[00:08:32] [SPEAKER_01]: Neil, people talk about four corners performance. So that is the IO pattern between compute and

[00:08:40] [SPEAKER_01]: storage. Is it 100% right? 100% reads, 100% random reads, and 100% random writes. Those are

[00:08:51] [SPEAKER_01]: the four corners. So not only is this drive going to deliver the industries fast just four

[00:08:58] [SPEAKER_01]: corner performance and a PCIe Gen 5 product, but and I think we might be getting into this a

[00:09:05] [SPEAKER_01]: little bit later. What sets all I'm apart is as we engineer for the real world. In reality,

[00:09:12] [SPEAKER_01]: no workload is going to 100% of the time be at those four corners. So understanding the real

[00:09:21] [SPEAKER_01]: IO patterns of our customers we do special things in firmware to optimize them for kind of these

[00:09:29] [SPEAKER_01]: behaviors in a real production environment. And so this really is going to set these drives apart

[00:09:37] [SPEAKER_01]: in terms of just massively accelerating a range of enterprise workloads. And getting back to AI

[00:09:44] [SPEAKER_01]: really quick, we believe we talked about GPU utilization. We believe that this product is going

[00:09:50] [SPEAKER_01]: to be really, really well positioned as direct a touch storage inside of that, you know, we're using the

[00:10:00] [SPEAKER_01]: GXH100 example, you know, inside that same rack to really maximize GPU utilization during critical

[00:10:10] [SPEAKER_01]: AI steps like checkpointing and and restore. And I suspect you've probably got the of a lot of

[00:10:18] [SPEAKER_00]: customers who are communicating with you that they're going all in an AI. They have evolving needs

[00:10:24] [SPEAKER_00]: requirements. I'm curious, how do you collaborate with your customers to tailor storage solutions,

[00:10:30] [SPEAKER_00]: meet those specifically, and maybe you need an and why is this approach important in the context

[00:10:36] [SPEAKER_00]: of enterprise storage? Because I imagine it's a conversation you're hearing more and more.

[00:10:41] [SPEAKER_01]: Again, it's a really good question. I think what sets us apart? Well, let me let me back up a couple

[00:10:47] [SPEAKER_01]: steps. So when you talk about kind of our legacy, our DNA or heritage, whatever we want to call

[00:10:55] [SPEAKER_01]: that we have, solid, has a bit over three decades of storage industry experience. And that really

[00:11:07] [SPEAKER_01]: kind of gives us in the storage experiences is multi level, at multiple technical levels. We are

[00:11:15] [SPEAKER_01]: talking engineered engineer in terms of strategic product planning, engineered engineer in terms of

[00:11:22] [SPEAKER_01]: design, test validation, etc., in these kind of long standing, deep technical relationships,

[00:11:32] [SPEAKER_01]: spanning enterprise and spanning cloud service provider customers are. This really kind of gives us

[00:11:39] [SPEAKER_01]: a rethink, kind of a leg up on our competition to do things that are unique in the industry

[00:11:47] [SPEAKER_01]: that really set us apart in terms of differentiating value. It's things like firmware optimizations

[00:11:53] [SPEAKER_01]: that I alluded to a minute ago that improve performance in real-world conditions. It's things like

[00:12:00] [SPEAKER_01]: creating more and more of our customer test into our test suite to kind of, you know,

[00:12:10] [SPEAKER_01]: multi-level benefits there, provide higher quality product to our customers. And because we're doing

[00:12:16] [SPEAKER_01]: kind of if you will more work for them, we are accelerating their time to market opportunity as well.

[00:12:25] [SPEAKER_00]: And before you came on the podcast, I was having a look on, I think it's one of your social

[00:12:28] [SPEAKER_00]: channels. It probably be LinkedIn and there's a great tagline there, the solid RMU expand the

[00:12:34] [SPEAKER_00]: possibilities of data that fuel human advancement. So we just talked about real-world conditions there.

[00:12:40] [SPEAKER_00]: Are there any real-world applications where solid RMSS-Ds have significantly accelerated workloads,

[00:12:47] [SPEAKER_00]: while lowering costs? Is any of those kind of stores you can share with us today? Because I suspect

[00:12:51] [SPEAKER_00]: that many tech projects now are especially in an era of economic uncertainty that challenge

[00:12:56] [SPEAKER_00]: doing more with the same, more with less of equally. They've got a big expensive tech projects they

[00:13:02] [SPEAKER_00]: need to get over the line. So I've looked to bring that to life a little if you've got any use cases or

[00:13:06] [SPEAKER_01]: all of those things. Yes, I can share a couple with you. Yeah, again, sticking with kind of the theme

[00:13:12] [SPEAKER_01]: that we're on here today, Neil, you know, in terms of AI, there is a hyper-scaler customer of ours

[00:13:21] [SPEAKER_01]: who I can't name. But this hyper-scaler moved their AI ingest in data preparation from HDDs

[00:13:32] [SPEAKER_01]: to solid RMSS-Ds and reduced the time that it takes to kind of complete those two key AI data

[00:13:43] [SPEAKER_01]: pipeline phases reduce the time by nearly 50X. We have a customer in the US named OCS. OCI

[00:13:55] [SPEAKER_01]: EMT is a really interesting customer and I would encourage your listeners to look them up.

[00:14:03] [SPEAKER_01]: They are planning that a combination of their really powerful software along with the hardware

[00:14:16] [SPEAKER_01]: stack they use, which includes RQLCS-S-Ds that they can reduce energy consumption of intense

[00:14:23] [SPEAKER_01]: workloads. So things like HPC, data analytics, AI, they can reduce the energy consumption of these

[00:14:30] [SPEAKER_01]: intense workloads by 50 to 90%. And when you talk about the, I guess, kind of the power constraint

[00:14:40] [SPEAKER_01]: or power challenges AI data centers are up against, you can see how this is tremendous value.

[00:14:47] [SPEAKER_01]: And then just across a range of OEM storage innovators, CSPs, companies like Super MicroVast,

[00:14:56] [SPEAKER_01]: Ebola, NetDell, Kingsoft, CheaterRAID, HPE, BiteDance, etc., all of these customers kind of value the

[00:15:08] [SPEAKER_01]: unique combination enabled by our high-cap leadership in terms of density, efficiency and performance.

[00:15:18] [SPEAKER_00]: And again to drill down on that, is anything you can share? Or maybe help businesses understand

[00:15:24] [SPEAKER_00]: the kind of business value and ROI that we're talking about in our conversation with you.

[00:15:29] [SPEAKER_00]: Is there anything else you can share that will help businesses understand solid arms technology,

[00:15:36] [SPEAKER_00]: help them unlock the potential of data for your customers? And also how that contributes to that

[00:15:42] [SPEAKER_00]: that tagline we mentioned a few moments ago about fueling human advancement and he can share around them.

[00:15:47] [SPEAKER_01]: I think maybe the best example is so let's assume that AI is primarily used for altruistic reasons.

[00:15:57] [SPEAKER_01]: So let's assume that. I think the best example then is if you just look at

[00:16:05] [SPEAKER_01]: the potential of AI, the solve massive problems, it improved the quality of life.

[00:16:13] [SPEAKER_01]: Okay, so how do our products kind of play into that? As that being an example of fueling human

[00:16:20] [SPEAKER_01]: advancement will, again at the risk of getting repetitive AI doesn't happen without massive amounts

[00:16:28] [SPEAKER_01]: of data. Data doesn't happen without storage and scaling of AI to reach its potential doesn't

[00:16:38] [SPEAKER_01]: happen without high capacity, highly efficient and performance storage. So that's kind of AI as an altruistic

[00:16:48] [SPEAKER_01]: example of how we believe that our technology can help fuel human advancement.

[00:16:57] [SPEAKER_00]: Love the uncurious considering solid arms origins in entails innovation and SKA on

[00:17:04] [SPEAKER_00]: leadership. How has it influenced your approach to developing and delivering

[00:17:10] [SPEAKER_00]: cooling edge storage solutions? I feel there's going to be some kind of

[00:17:13] [SPEAKER_00]: synergies there or anything that you can share around them? Well, yeah, I think it's a combination

[00:17:21] [SPEAKER_01]: this heritage within Intel which I'll touch on a bit more and then our ability to

[00:17:30] [SPEAKER_01]: invest to sustain leadership, you know, bad and able by SKA, Inax is our parent corporation.

[00:17:38] [SPEAKER_01]: So kind of getting back to the Intel element of that, what makes us unique again is this

[00:17:48] [SPEAKER_01]: platform solution, I guess, customer centric approach that Intel took in stilltikes and it is

[00:17:57] [SPEAKER_01]: part of our DNA. And this mindset, if I could be a, I love history so I'm going to kind

[00:18:07] [SPEAKER_01]: of click into a bit of history here, you know, this mindset has delivered this long legacy of

[00:18:12] [SPEAKER_01]: storage industry innovation. What many of your listeners might not know is that we ship the first

[00:18:19] [SPEAKER_01]: flash disk in the form of into an IBM PC in 1992. Wow, our heritage goes back to us co-creating

[00:18:30] [SPEAKER_01]: the PCIe spec in O3. We co-founded the Open Compute Project which is an incredibly important

[00:18:40] [SPEAKER_01]: industry consortium that sets hard worst standards for hyperscalers and more.

[00:18:45] [SPEAKER_01]: We invented the ruler, Neil, if you've ever, if your listeners are seeing these long,

[00:18:51] [SPEAKER_01]: skinny SSD form factors that was called the ruler and then we donated that to Snea and that

[00:18:56] [SPEAKER_01]: became the EDSF specification which is increasingly the preferred kind of modern form factor for

[00:19:04] [SPEAKER_01]: multiple reasons. We were the first to ship QLC SSDs into the data center. We lead in high cap storage

[00:19:12] [SPEAKER_01]: now with our 61.44 terabytes drive. So kind of, if you, I guess step back and kind of look at all

[00:19:19] [SPEAKER_01]: these things coming together. This, this investment enabled by our parent corporation, this history

[00:19:30] [SPEAKER_01]: of legacy, this kind of customer first mindset. All of this is now kind of coming together into

[00:19:37] [SPEAKER_01]: we believe that our technology is, you know, kind of led us to, you know, our technology meeting the

[00:19:45] [SPEAKER_01]: AI moment in terms of storage challenges and opportunities. I don't know idea, if we look to

[00:19:53] [SPEAKER_00]: the future, what trends do you see in the future of SSD technology, especially in relation to AI?

[00:20:01] [SPEAKER_00]: And how would you say your position to lead on this continuously evolving landscape that seems

[00:20:06] [SPEAKER_00]: unfolding in real time before our eyes at the moment? Yeah, that's an interesting observation.

[00:20:13] [SPEAKER_01]: We think that, you know, I don't know, let's go up. What is today Wednesday? We go back

[00:20:18] [SPEAKER_01]: back last Wednesday. Boy, we've really got an understanding of the AI landscape and what it means

[00:20:25] [SPEAKER_01]: to storage and, and here we are, we later, oh my gosh, this new thing is emerging and that seems

[00:20:31] [SPEAKER_01]: to, you know, be having implications on storage. So this is just, this is the wild west. This is such

[00:20:38] [SPEAKER_01]: a rapidly evolving industry but super exciting. And so I think our portfolio is well positioned

[00:20:47] [SPEAKER_01]: now and the investments that will make as well positioned to continue our portfolio leadership.

[00:20:55] [SPEAKER_01]: So drill it into that a little bit. As we've alluded to several times, you know, efficient

[00:21:03] [SPEAKER_01]: performance high cap SSDs is certainly critical. That's critical to overcoming the, you know,

[00:21:13] [SPEAKER_01]: it can be a key part of solving the power requirement challenges for AI data centers.

[00:21:20] [SPEAKER_01]: We will continue to invest in where a hundred percent vested in sustaining high cap leadership.

[00:21:28] [SPEAKER_01]: So I think we kind of start there. We will have the next capacity point beyond 61.44

[00:21:34] [SPEAKER_01]: terabytes in market by the end of the year. We talked about the edge a bit. It if you kind of

[00:21:41] [SPEAKER_01]: ear listeners are probably familiar with the term data gravity and the more and more as much as

[00:21:49] [SPEAKER_01]: possible you want to store, analyze act on data where that data is produced. It's a very expensive

[00:21:59] [SPEAKER_01]: to move data up and down the wire. AI workloads are moving to the edge. And so we're looking

[00:22:05] [SPEAKER_01]: at capabilities on what more can we do with data at the edge? So kind of stay tuned for more there.

[00:22:13] [SPEAKER_01]: And we think performance leadership is key as well. And it's especially key, you know, if we

[00:22:18] [SPEAKER_01]: talk about GPU utilization and that kind of direct attach storage element to keeping these 300,000

[00:22:26] [SPEAKER_01]: US dollar GPU servers fully utilized. So we will invest to sustain our performance leadership as well.

[00:22:38] [SPEAKER_00]: Wow, so much food for thought there and a lot of a few teases as well. So I think we will have to

[00:22:44] [SPEAKER_00]: get you back on towards the end of the year to find out more about some of those things. But

[00:22:49] [SPEAKER_00]: I kind of say, you know, for coming on here and showing your insights around AI and the power

[00:22:53] [SPEAKER_00]: of SSDs too enabling and also mentioning the product launch of course. But before I let you go,

[00:22:59] [SPEAKER_00]: I'm going to ask you to leave everyone listening with one final gift. And I always ask my

[00:23:03] [SPEAKER_00]: guests, is that a book that you could leave to add to our Amazon wish list or a song to add to our Spotify

[00:23:10] [SPEAKER_00]: playlist, guilty pleasures all allowed there as well. But all I've been asking is what you'd like

[00:23:14] [SPEAKER_01]: to leave everyone listening at one? Okay, I'm actually going to answer that in two parts and

[00:23:20] [SPEAKER_01]: you can cut one of the mouth if there isn't enough time. But in terms of book, I often refer to

[00:23:26] [SPEAKER_01]: there's a book by a British author Richard Holmes called The Age of Wonder. I think your readers,

[00:23:33] [SPEAKER_01]: you know, you have in a technology podcast, I'm sure they're interested in in the scientific discovery.

[00:23:39] [SPEAKER_01]: This is about the wave of scientific discovery that occurred in England,

[00:23:43] [SPEAKER_01]: kind of between the late 1700s through mid 1800s and it really is a testament to kind of the

[00:23:49] [SPEAKER_01]: capacity of the mind and human inquisitiveness to solve seemingly intractable challenges.

[00:23:59] [SPEAKER_01]: And then I love the question because I'm also a huge music fan my favorite artist by far is

[00:24:05] [SPEAKER_01]: Ray LaMontaine, not sure if he's big in the UK but it's encouraged your listeners to check him out.

[00:24:12] [SPEAKER_01]: Chill vibe he's dedicated to his craft. He's hardworking my absolute favorite song of his is such a

[00:24:18] [SPEAKER_00]: simple thing. Wow, a fantastic artist. Yeah, he used to be a huge fan. He was holding

[00:24:25] [SPEAKER_00]: me in my arms. That must be a big. Yeah, he feels like a few years ago. He was probably

[00:24:29] [SPEAKER_01]: like 20 years ago, you know? Yes. He just dropped a new album the other couple weeks ago

[00:24:34] [SPEAKER_01]: and my wife and I are going to see him in concert for probably the fourth time sometime later this year.

[00:24:41] [SPEAKER_00]: Wow, well I will certainly get both of those added to Amazon wishlist and the Spotify play

[00:24:46] [SPEAKER_00]: is great choices. But anyone listening want me to find out more information about solid I'm

[00:24:52] [SPEAKER_00]: and explore some of the topics we talked about today. Anyway, in particular you'd like to point

[00:24:56] [SPEAKER_01]: everyone listening. Sure, I think the best spot to start is probably our website www.solidime.com.

[00:25:05] [SPEAKER_01]: That is how you pronounce it. Solid I'm combination of solid and paradigm. It's spelled S O L I D

[00:25:14] [SPEAKER_01]: I G M. And since we talked also Neil since we talked quite a bit about AI we have a really informative

[00:25:23] [SPEAKER_01]: AI page. It's under the Solutions tab on our website. And test it. Well, I will have links

[00:25:29] [SPEAKER_00]: and for the blog post that will accompany this episode so people can find all that stuff

[00:25:35] [SPEAKER_00]: from the ISOAZ. One of the things I've looked down with you about today is talking about AI

[00:25:39] [SPEAKER_00]: but looking beyond the hype and how your story solutions are helping to unlock the performance

[00:25:44] [SPEAKER_00]: and are tuned for modern real world applications. But I think the end goal here is accelerating

[00:25:50] [SPEAKER_00]: work like while lowering costs and scaling efficiently. I think it's a bit of a utopia for a lot

[00:25:55] [SPEAKER_00]: of businesses that are aiming for that but you're proving you can do just that so thank you for sharing

[00:26:00] [SPEAKER_00]: your story with me today, much appreciate it. Thank you Neil thanks for the opportunity to appreciate

[00:26:04] [SPEAKER_00]: it. I think as we've seen today the future of AI is closely into twirling with the advancements

[00:26:10] [SPEAKER_00]: in storage technology and a big thank you to Roger for giving us a glimpse into how they're latest

[00:26:17] [SPEAKER_00]: SSDs are not only setting new standards for performance but also efficiency and reliability

[00:26:22] [SPEAKER_00]: across the entire industry. And I think by optimizing storage solutions for real world applications

[00:26:28] [SPEAKER_00]: they're not only accelerating AI and enterprise workloads but they're also paving the way for future

[00:26:35] [SPEAKER_00]: innovations but the big question is how do you see the role of storage evolving in the realm of

[00:26:40] [SPEAKER_00]: AI and LeBee your thoughts. So let me know, TechRug right to Outlook.com, Neil Seahouse on all the

[00:26:46] [SPEAKER_00]: usual social channels but thank you for joining me today, stay tuned for more insights into the

[00:26:52] [SPEAKER_00]: technologies shaping our world. Join me again tomorrow where I've got another guest lined up

[00:26:56] [SPEAKER_00]: an evil speak to you all right in early tomorrow. Bye for now.