3058: The ROI of AI for the Enterprise: Insights from Lenovo Tech World

[00:00:04] What does it really mean to bring hybrid AI to life? Especially in a world where organizations are increasingly looking to harness the power of artificial intelligence.

[00:00:16] Well, this is just one of the many questions I want to try and find answers to here at Lenovo Tech World. We are reporting live in Seattle this week.

[00:00:26] And today I've managed to catch up with Vlad Rozanovich, Senior Vice President of Lenovo's Infrastructure Solutions Group.

[00:00:34] And Vlad is not only leading the charge in data center innovation, but he's also helping businesses navigate their AI journeys.

[00:00:44] So today's episode is recorded live on the show floor at Lenovo Tech World.

[00:00:48] And Vlad is going to be sharing his insights on Lenovo's comprehensive AI-ready portfolio.

[00:00:55] And discuss how it's designed to accelerate AI adoption.

[00:01:00] Whether that be from groundbreaking liquid-calling technologies to turnkey solutions that meet the demands of today and tomorrow.

[00:01:09] This episode is going to dive straight into the heart of AI's potential impact on enterprises.

[00:01:15] And as AI continues to dominate headlines, Vlad offers valuable perspectives on how businesses can realize immediate returns and strategically approach their AI investments.

[00:01:28] What value does AI bring to the enterprise?

[00:01:31] What is the ROI on AI investments?

[00:01:34] These are some of the questions I want to try and get to the bottom of today.

[00:01:37] So I invite you to join me here in Seattle as we explore what Lenovo is doing to redefine that AI landscape.

[00:01:45] And how organizations can leverage these advancements for lasting success.

[00:01:50] But enough from me.

[00:01:51] Let's get Vlad onto the podcast now.

[00:01:54] So a massive warm welcome to the show.

[00:01:57] Can you tell everyone listening a little about what you do at Lenovo and also what the Lenovo World event means to you, Roo?

[00:02:04] Sure.

[00:02:05] Well, Neil, first of all, thanks for having me on the show today.

[00:02:07] So Vlad Rezanovic, what I do at Lenovo is I run our infrastructure solutions group, which is really our data center organization.

[00:02:14] So responsible for what products, solutions that we're actually building for the market, putting out to the market, that incorporate everything in the data center.

[00:02:22] Whether it's large data center scale systems like 8-way GPU, LLM machines, all the way down to edge portfolio models.

[00:02:31] And really looking at the heart of the market, which is your kind of standard 2P2U enterprise class servers.

[00:02:37] But my responsibility today is to manage that business group.

[00:02:40] And I also manage the customer-facing aspect of our ISG organization as well.

[00:02:45] So for the last year and a half, I've been responsible for all customer engagement on ISG.

[00:02:51] And prior to that, I actually spent two years as being president of North America for Lenovo, where I covered everything from PC to ISG to Motorola.

[00:02:58] Wow. And what about the event itself? What excites you about being?

[00:03:03] Well, what's exciting is having a thousand people here that actually want to hear Lenovo's message and what we're doing to make their IT lives easier.

[00:03:11] What's also exciting is the amount of special guests we had today at the event.

[00:03:16] So we had Jensen from NVIDIA. We had Lisa Sue from AMD.

[00:03:19] We had Pat Gelsinger in person from Intel.

[00:03:22] We had Cristiano webcast.

[00:03:24] We also had Mark Zuckerberg and Satya from Microsoft.

[00:03:29] So the lineup of people who wanted to talk about their relationship with Lenovo was something that was actually very special for me.

[00:03:39] Yeah.

[00:03:39] Because to see that many people, not only here in person from a customer, a press, an analyst perspective,

[00:03:45] but also to see the leaders in the industry wanting to join YY and myself on stage to talk about the relationship with Lenovo and why it's special.

[00:03:57] To me, that was probably one of the most exciting parts of the event.

[00:04:01] I must admit, before I came, I was looking at the lineup.

[00:04:03] It was like a big tech super group, isn't it? Anyone was there.

[00:04:07] It really was. And actually to see them all interact.

[00:04:09] Yeah.

[00:04:09] And even that Intel and AMD announcement today was actually something that I never thought I would see.

[00:04:15] Yeah.

[00:04:16] And they decided to do that here.

[00:04:17] And there is so much hype around AI at the moment.

[00:04:21] And I think for a lot of business leaders, the big question is, what can AI bring to my enterprise?

[00:04:26] There's been a lot of caution to begin with that people sat on the fence.

[00:04:30] That is changing now.

[00:04:31] So how would you answer that question?

[00:04:33] Because I suspect it's one that you get a lot with your conversations with clients.

[00:04:37] It is, Neil. And in fact, over the last year, I have visited with over 110 CIOs or CTOs around the world.

[00:04:44] And number one is, we also did a special investigation from a Lenovo perspective that reached out to not only CIOs and CTOs,

[00:04:52] but also to CEOs of large corporations and midsize corporations.

[00:04:55] And one of the things we saw as part of this, which we outline in what we call our CIO playbook,

[00:05:02] is CEOs are looking for AI to deliver instantaneous returns with productivity, efficiency, and profitability to be implemented immediately.

[00:05:12] And when you talk to CIOs and CTOs and VPs of IT infrastructure, it kind of drives them nuts because they need to answer a few questions before they can get that productivity, profitability, or efficiency.

[00:05:25] And the questions are, what is the business outcome?

[00:05:28] What is the business trying to solve?

[00:05:30] What do they need to use AI for?

[00:05:32] Number two is, do they have to staff up?

[00:05:35] Or what is the expense budget within these organizations to actually migrate to AI?

[00:05:39] Do they have to hire chief data scientists now to understand what the data that's coming out of AI?

[00:05:46] And most importantly, they need to understand, is this something that is going to actually not only look at a business outcome, but also how is this going to be budgeted?

[00:05:56] And is it going to be something that is going to be a long-term strategy for many corporations?

[00:06:01] And what we have seen as CIOs and CTOs, they see AI as having advantages immediately.

[00:06:08] And some of those use cases could be very prominent things like customer service chatbots.

[00:06:15] That has been one of the early use cases around AI that we see many, many corporations adopting.

[00:06:20] The other one we see is code generation for IT, where if you could actually train AI using things like GitHub to actually help you with code generation, many IT users today are finding that they can get a speed up of code implementation by 60, 70, 80%.

[00:06:39] And now those are the early use cases of AI that we're seeing that makes a difference immediately.

[00:06:45] And they could provide instantaneous returns in things like customer service or in productivity of software coders within an enterprise.

[00:06:56] Now, the next step of this is how do these corporations actually use AI to act on the data that's private or most dear to them?

[00:07:04] So if you're a banking or financial services institution, how do you actually act on your customers' data to provide the best advice for financial services?

[00:07:12] If you're in the energy sector, how do you actually use AI to help with maybe it's petrol discovery or maybe it's pipeline routing?

[00:07:22] How do you use AI to actually help you with those very complex use cases?

[00:07:28] Maybe you're in media entertainment and you're looking at things like how do we use AI to translate and film images for multiple languages?

[00:07:36] And so what we're seeing is it's that next level of AI now that CIOs and CTOs are starting to get into that if they can provide business value, then the CEOs will see productivity, profitability or efficiency gains.

[00:07:50] And so that's something that I'm seeing across the board worldwide, whether it's in Asia, in Europe, in Latin America, in North America.

[00:07:59] We see a very consistent viewpoint of what AI could potentially do for some of these organizations.

[00:08:05] And although AI gets all the headlines, of course, it is nothing without data.

[00:08:10] Yes.

[00:08:11] I've seen data many times here.

[00:08:13] And whether it is data sovereignty, data silos, data quality, not to mention how do we keep it safe and secure?

[00:08:20] How do you help businesses with these data challenges with AI?

[00:08:23] Well, it's great because everything you mentioned about the privacy, the sovereignty, it's exactly what I set up on stage today.

[00:08:28] Data is at the heart of what are you going to do with AI to get to a business outcome?

[00:08:34] And understanding that data management is probably one of the most important things that enterprises, corporations need to figure out.

[00:08:44] Because what we're seeing around the world right now is many, many countries are looking at data sovereignty laws or data privacy laws.

[00:08:54] And you have to act on those.

[00:08:56] You have to act on those to make sure that, okay, this is what I need to make sure that if I put this in an AI infrastructure, how do I have the data management tools in place?

[00:09:04] Where the data management architecture to ensure that the data that I am looking at using is kept private, secure, sovereign.

[00:09:14] And it's the whole concept of data gravity.

[00:09:16] Now, what the reality is, is data is being created at the edge.

[00:09:20] It's being created by all the massive devices that we're seeing out there in the industry.

[00:09:26] And acting on that data at the edge with the lowest latency possible is something that many, many corporations are starting to figure out.

[00:09:34] Now, there is an advantage for things like large language models to be centralized.

[00:09:38] Because the speed of which large language models can be trained in the big public clouds is something that enterprises can't, they can't build for that in their four walls.

[00:09:48] So they have to leverage the public cloud for those types of models.

[00:09:52] But if you're looking at small language models or retrain of those models or inference of those models, that's where that AI is going to happen, where that data sits closer to those end devices, endpoint devices.

[00:10:07] And I think something else I thought was really interesting is because I've heard so many stories in the last year or 18 months about companies pushing back on their ESG goals to invest in AI, which has a well-documented sustainability.

[00:10:20] Yeah.

[00:10:20] But one of the things that stood out to me today was Lenovo's next era of Lenovo-Neptune water cooling.

[00:10:26] Yes.

[00:10:26] Which addresses these very concerns.

[00:10:28] Can you just expand on that for anyone that's not here?

[00:10:29] Sure, sure.

[00:10:30] So what we see happening in the industry, whether it's high-performance computing or many of these large language model workloads, it is a power-hungry use case.

[00:10:40] And it's something that we've dealt with now for six generations at Lenovo.

[00:10:45] In fact, if you look at our early water cooling solutions that we introduced for our HPC offerings years and years ago, 10 years ago, everything we have done has been an advancement on that over the last 10 years.

[00:10:57] And today, there is this concept of being at odds of high-performance computing, AI use cases, and ESG requirements.

[00:11:09] And what's interesting is that you have a massive, massive growth rate on how high-performance computing is sucking up massive amounts of power.

[00:11:19] Yeah.

[00:11:19] And it takes massive amounts of cooling.

[00:11:21] But then you have these requirements that so many corporations are looking for, for how do I make sure I do things in a sustainable way?

[00:11:28] So, Neil, what we announced today, which was very exciting, is we introduced our sixth generation of Neptune water cooling, which is really the way we use our copper tubes, the way we use our aluminum cold plates, the way we route the warm water infrastructure.

[00:11:42] What's interesting about that is it's all about how do you actually remove heat from a system so you don't need some of that cooling in a data center.

[00:11:51] In many data centers, what we see today is that cooling accounts for about 30% of the energy usage for many corporations within the data center.

[00:12:00] And if you could actually go to a position where you don't have to have those chillers and air handlers, and you could actually use that power for compute versus cooling and moving air, effectively, it does create a more efficient architecture within your data center environment.

[00:12:17] And so Lenovo is going to continue to make strides and gains on how do we look at the ESG requirements that some corporations are needing to hit with that desire for more power for compute.

[00:12:31] And so we see some companies saying, you know what, I'm going to use Neptune water cooling to reduce 30% of my energy bills, but they're going to keep the compute levels the same.

[00:12:40] Then we see the other aspect of it, which is, hey, if you're going to give me 30% more energy by power savings from cooling, I'm going to go turn that around and do more compute.

[00:12:51] And so that's going against some of the ESG guidelines, but it is creating a more efficient system.

[00:12:55] And for anyone listening interested in that side of things, with 100 kilowatt plus server racks now possible without specialized air conditioning, how do you envision data center design changing in the next five years?

[00:13:07] Are there any other challenges in retrofitting existing data centers to accommodate these technologies?

[00:13:13] What are you seeing here?

[00:13:13] Well, it's interesting.

[00:13:14] I even had a meeting today with Schneider, which is one of our data center partners on how do they do things like power distribution units, CDU units for data centers.

[00:13:26] And it could be greenfield data centers or brownfield data centers.

[00:13:29] And so the difference there is if you're building from the ground up, you can use techniques that are using water cooling right away and you can build for that.

[00:13:36] And as we talk to people like Colos, digital realties, Equinix's, we're having those conversations to say, hey, do we retrofit our data centers to do liquid cooling?

[00:13:48] And so those are conversations that are taking place everywhere.

[00:13:51] Now, on an extreme side, we're also seeing that the uniqueness of where people are placing data centers.

[00:13:58] It's where power is cheap or it's where they have access to cooling solutions that could be very unique.

[00:14:07] Like, for instance, we've heard that there's old mining facilities that are underground that are saying, you know what, at a 55 degree ambient temperature, it would be really interesting to put a data center here, especially since you have power lines coming into maybe that was an old mining facility.

[00:14:25] And so that's something we're seeing.

[00:14:26] We're also seeing countries like the Nordics actually really expanding their data center space because of cheap hydropower.

[00:14:33] And so there's a gravitation to where that cheap power is.

[00:14:37] We also see in Abilene, Texas, Western Texas, lots of solar farms that then create lower cost energy.

[00:14:44] And so the cost of energy is so important as part of this, where do you build new data centers?

[00:14:49] And now the question on where do you do retrofit or brownfield?

[00:14:53] Those are the questions of do you retrofit an existing data center to go liquid cooling to potentially add an advantage of 30% more electricity for compute versus cooling?

[00:15:04] Yeah.

[00:15:04] I also think we probably alluded to this at the very beginning of our conversation, but perhaps the biggest question of all when it comes to bringing AI to the enterprise is that ROI, the return on investment of that AI project.

[00:15:18] I think more so than ever, every tech project is under close scrutiny for this and what business value it generates.

[00:15:24] So how are you helping businesses achieve this elusive ROI on tech projects?

[00:15:29] Well, I'll tell you, I think it's even been made public that if you look at the investments made by Meta and OpenAI, the investments have out surpassed the amount of revenue that they've made at this point.

[00:15:40] Yeah.

[00:15:41] But it is getting ready and getting those frameworks ready and getting those large language models for training running.

[00:15:51] Yeah.

[00:15:51] And so AI is a big bet.

[00:15:54] AI is a massive bet for the industry.

[00:15:56] And the amount of money being spent on it right now is something that I've never seen in my 25 years in this industry.

[00:16:03] However, the early use cases and some of those, like I mentioned before, code generation for IT software or customer service chat bots, those are almost instantaneous language, natural language translation.

[00:16:15] Those are almost instantaneous results that people are seeing that have created efficiencies across corporations and enterprises.

[00:16:24] And so the proof points on will this be successful, there's some very, very early promising proof points for is AI going to continue at the pace it is today.

[00:16:34] So if we bring everything we've talked about today together, if we come full circle, if we tackle data sustainability and ROI challenges, what's the future of AI in the enterprise look like?

[00:16:46] Well, I think it becomes more specialized.

[00:16:48] And so it's one of the reasons that Lenovo actually looks at things like our AI innovators program, where we work with dozens of ISVs that are providing very specialized use cases.

[00:16:58] In fact, one of those use cases was adopted by the largest food store in the US, Kroger.

[00:17:04] What's interesting is we were doing business with Kroger in their infrastructure space.

[00:17:08] But their security team started investigating very small ISVs for they wanted to do machine vision technology for theft, deterrency and detection at the self-checkout kiosks.

[00:17:20] But it was a very small ISV.

[00:17:22] And so they were a little bit concerned on, can I trust this small ISV within my large enterprise?

[00:17:28] And they came to Lenovo and said, would you do a certification on this software?

[00:17:33] Would you help support this software?

[00:17:36] And that was something that Lenovo had done.

[00:17:39] And so this is an AI use case for security, theft deterrency, theft detection.

[00:17:44] And by using those AI innovator partners, we've created a very narrow solution or business outcome.

[00:17:52] And what I think is going to happen is you're going to see more of these very specialized use cases that are going to be very focused for a very set application that is going to use AI to make very efficient, quick, profitable decisions for corporations.

[00:18:07] Wow.

[00:18:07] Wow.

[00:18:08] Powerful moment to end on.

[00:18:09] But before I do let you go, when you finish here at the event and you're on that plane ride home, what are you going to be thinking about?

[00:18:14] What's your conference takeaways?

[00:18:16] What are you going to be thinking about on that plane ride home?

[00:18:18] Yeah.

[00:18:18] Well, I'll tell you, there's a couple of things that I'm going to take away with me.

[00:18:21] Number one is that Lenovo truly is leading the industry when it comes to pocket to cloud and smarter AI.

[00:18:29] We are a powerhouse in the industry.

[00:18:31] And even though we're number one in the PC space, we are up and coming in the data center space in a very, very fast way and in the services space.

[00:18:39] And so what I'm going to be proud of is all the hard work that our engineering teams, our sales teams, our marketing teams, our communications teams have put in place, our operations teams, our manufacturing teams.

[00:18:51] I'm going to think about all of the hard work that those employees have put into making this a pivotal moment for Lenovo to really be on that forefront of AI and hybrid AI.

[00:19:03] Well, from the outside looking and I've seen you on stage today, back to back interviews all day.

[00:19:07] But I think it's such an important question about what value AI brings to the enterprise.

[00:19:12] A lot of business leaders thinking about it at the moment.

[00:19:15] And just thank you for taking the time out to sit down with me and share that with me today.

[00:19:18] Oh, it was a pleasure.

[00:19:19] Thank you.

[00:19:19] As we wrap up our conversation with Vlad today, for me, my big takeaway was it's clear that Lenovo is not just a player in the tech industry,

[00:19:27] but somewhat of a pioneer in making AI accessible and actionable for enterprises.

[00:19:33] And Vlad's insights into some of the challenges of implementing AI, along with the importance of data management,

[00:19:40] offer an almost roadmap for business leaders to tap into the benefits of this transformative technology.

[00:19:46] And reflecting on his key takeaway from Lenovo Tech World, I think it's evident that the future of AI in enterprise is not just about immediate gains,

[00:19:55] but it's also about building sustainable solutions, sustainable solutions that align with corporate goals.

[00:20:02] But the big question, of course, is what strategies will you and your organization employ to leverage AI effectively?

[00:20:10] This is a dialogue, not a monologue.

[00:20:13] So I do invite you to share your thoughts and your experiences on this evolving topic.

[00:20:17] As always, techblogwriteroutlook.com.

[00:20:20] You can connect with me on LinkedIn, X, Instagram, just at Neil C. Hughes.

[00:20:24] Let me know your thoughts on this one, because it's something that will impact every individual and, indeed, every business.

[00:20:30] But that's it for now.

[00:20:32] Now, I'm getting back out on the show floor and trying getting some other perspectives here at Lenovo Tech World.

[00:20:38] So stay tuned.

[00:20:39] I'll be back again tomorrow with another guest.

[00:20:41] But thank you for joining me today.

[00:20:43] And hopefully I will speak with you all again tomorrow.

[00:20:46] Bye for now.

[00:20:49] Bye for now.

[00:20:53] Bye for now.

[00:20:53] Bye for now.

[00:20:55] Bye for now.