On this episode of The Six Five – On The Road, hosts Daniel Newman and Patrick Moorhead welcome Robert Daigle, Director of Global AI Business and Blake Kerrigan, GM of Global ThinkEdge Business Group at Lenovo, for a conversation on Lenovo’s AI Innovator’s program, their AI investment, and how Lenovo’s solutions simplify and improve AI-adoption.
Their discussion covers:
- What initiated Lenovo’s AI journey and how are Lenovo plans to stay ahead of the rapidly evolving AI competitor landscape
- How Lenovo’s AI Innovators program has outperformed its initial expectations and how the program expects to grow over time
- Why Ethical and Responsible AI is a primary focus area for Lenovo and what they are doing to ensure what they deliver is just that
- Lenovo’s perspective on the recent challenges or pain points they’ve observed in the adoption of AI and how Lenovo’s new solutions uniquely simplify and improve AI adoption
- The advantage of Lenovo’s edge portfolio in enabling AI-adoption and improving scalability at the edge
Watch the video here:
Or Listen to the full audio here:
Disclaimer: The Six Five webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.
Patrick Moorhead: Hi, this is Patrick Moorhead, and The Six Five is live on the road at VMware Explore 2023 in Las Vegas. And as you can hear, we are on the show floor. It is loud. There’s a lot of activity. And in fact, about 150 people just exited stage left, so we can come in and do a video for you. I’m here with my co-host, Daniel Newman. Dan, how you doing?
Daniel Newman: You’re so selfless when you offer to do these videos for all these nice people out there. I do this for me. I just want you to know that. No, in all serious, it is really great to be here. It’s always great to see the evolution of technology happening so fast.
Patrick Moorhead: I’m still laughing at what you said. That was good.
Daniel Newman: Do it for you. Remember, sales 101, say you a lot.
Patrick Moorhead: Yes.
Daniel Newman: But we’re here, and it’s interesting because we’re in this really big inflection, it’s kind of a seminal moment in tech. The multi-cloud journey has been really interesting. You called it 10 years ago when I was still in high school, and I basically have watched now digital transformation 3, 4, 5, all take place in like three years. You can call it 2.0, 3.0, 4.0. You can use whatever you want to use, but AI has come on so fast. It’s changing every business, every industry, and it’s basically become the theme of every single event. It’s no different here at VMware Explore. I’m going to keep calling it VM World, Pat, no matter how many times I say that.
Patrick Moorhead: No, it’s okay. I think the stalwarts would love you for it.
Daniel Newman: I’m committed.
Patrick Moorhead: Yes, so do that.
Daniel Newman: I’m committed.
Patrick Moorhead: By the way, the other trend that we’ve seen is the Edge. Everybody, what, seven or eight years ago, hey, it’s going to be monolithic. Everything’s going to be inside of this massive data center. We only need three of them. Well, it reminds me of what was said about the mainframe about 40 years ago, but the AI plus the Edge is powerful. And imagine that, we have the leaders of AI, Robert, great to see you, and the Edge, Blake, here from Lenovo. Thanks for coming on The Six Five.
Robert Daigle: Thanks for having us.
Blake Kerrigan: Great to be here.
Patrick Moorhead: Yeah, we have a great history with Lenovo and The Six Five. I mean we’ve had your fearless leaders on, we haven’t had YY yet, but we will see in the future. We’re looking forward to your amazing big tent event coming up in Austin as well. So thanks for coming on the show.
Robert Daigle: Thanks for having us. We’re really excited about Tech World coming up. Lots of really good announcements that we’re going to be making around AI and our Edge portfolio. So it’s an exciting time here at Lenovo.
Daniel Newman: I love it. Right on the heels of the big race. That was the last time we hung out with YY actually.
Patrick Moorhead: We were. We took up a lot of square foot in your paddocks, and we appreciate you. Thank you. And as I say, I only begged for one event and that’s it. So that’s all we do, for one.
Daniel Newman: All right, so we’re having fun here. Robert, I’d like to start with you talking a little bit about the AI journey. So Lenovo’s clearly made some big proclamations around its plans in AI, billion plus dollar investment commitments, building out global AI data centers, AI ready solutions. Talk a little bit about this kind of journey that the company’s taking and how it’s sort of using this to really build a reputation as being a big part of the global AI story.
Robert Daigle: Yeah, I think one of the most important things here is that we’re not new to this journey. This is not a new commitment. We’re not jumping into it. I’ve been with Lenovo since 2017 when we launched our AI business. So we’ve been at it for some time now. We’ve made some changes and adjustments along the way, but even from that initial commitment of 1.2 billion investment into our merging technology space, including Edge computing and AI, that’s really been the foundation that we’ve been building upon. We’re one of the largest AI infrastructure vendors in the world, one of the top three in the world, and that’s based on those investments. What we’re doing right now is we are doubling down on that. We’re putting another billion dollars into our AI business to scale it out very quickly, to take what we’ve built as a foundation and bring it globally for enterprise organizations.
Patrick Moorhead: So I’ll shift to Blake now. Blake, if you look at the last 40 years of it, compute goes as far as it can to the edge as possible. Let’s just say shorthand, we went from mainframes to minis to client server, mobile, but we have a lot of heavy duty processing going on the edge. And whether it’s in a hospital like we talked about in the green room or at retail or any Web 4.0 that you can think of, smart manufacturing, smart this. Can you talk about how you’re infusing AI into the edge?
Blake Kerrigan: Yeah, absolutely. In fact, one of the key drivers for edge computing is this idea of new different AI opportunities at the edge. So when we think about predictive analytics, computer vision, generative AI, believe it or not, in a few corner cases, we see that starting to unlock a lot of untapped potential for companies. And like Robert said, we’ve been at this for just about six years, building purpose built edge computing hardware, and there’s kind of two things or two transitions that we’re seeing happening. One is the business cases you alluded to, which is there’s a massive opportunity for consolidation of resources at the edge.
So similar to the mainframe we see in a retail example, we’ve got point of sale video management, surveillance, smart IOT applications. We can build a case for edge computing just around consolidating those workloads. But at the same time, there’s this transitional period around AI that’s also introducing this new opportunity, but also some complexity, and that we’ve got to figure out how to put data center like GPUs inside systems that don’t overheat in the back of a store closet or something like that.
Patrick Moorhead: Yeah, for sure. It’s a big challenge.
Daniel Newman: So edge to cloud, Robert, I mean the edge is clearly going to be a massive part of the future, and I think this early wave of data center AI has kind of been all about the GPU, just kind of, but I think we’re all seeing it. It’s going to go to the application, it’s going to go to the phone, it’s going to go to the PC, it’s going to go to the smaller edge data centers. There’s just too much data. No way you’re going to move it all in between. And Lenovo’s building out a portfolio of edge ready, cloud ready, but kind of it’s AI ready. Talk about how you’ve gone about deciding what goes into that portfolio, what solutions you’re leading with, how is that being built out? Because I think you’re trying to solve the problem of what do customers need right now with AI and how do we make it really simple?
Robert Daigle: Absolutely. AI is going to be everywhere, and our mantra is bring your compute to where your data is and where your data’s being generated. So bring your AI to the data, don’t bring your data to the compute so you don’t have to pipe everything back to the data center or cloud. It really doesn’t make sense. So we have to put purpose-built AI systems along the spectrum of edge, from pocket to edge to core data centers, and even into the cloud. At Lenovo, we build over 70 AI ready systems across our ThinkEdge, ThinkAgile and ThinkSystem portfolio and then also purpose-built AI systems. We actually have a systems purpose-built for generative AI workloads. We’ve got systems that are purpose-built in our edge portfolio for computer vision at the edge. So there’s going to be a mix of having systems that are AI ready and then also systems that are purpose built for AI workloads.
Patrick Moorhead: Yeah, it’s interesting. Back of the envelope, right? Degenerative AI, is it 10 x more, 5 x more compute that’s required? Well, it depends. It does just depend. But we do know it takes more horsepower and with more horsepower means heat. Blake, you had talked about that. I’m curious on the edge, what are some of the challenges that you’re finding that you’re having to solve with AI? And in fact, I think you brought out a brand new Think Edge platform recently, second generation of a platform. Maybe talk about the types of problems that that’s solving.
Blake Kerrigan: So maybe I’ll just talk about the two different products. So the first flagship product within the ThinkEdge portfolio was the SE 350 V one. It was way ahead of its time. It really kind of was the first system that people started centering enterprise workloads outside of the data center with the same availability, security, reliability. What we’ve done in this next version was we’ve basically created two different systems, one of which is a short depth one U, the other of which is a short depth two U. And what we’ve been able to do with both systems is while we have definitely a need for consolidation on the 350 V two, so we’ve optimized it for both storage and interconnectivity reliability. So you can get extremely compact, three node clusters running something like VMware edge compute stack to do enterprise workloads. But say in the day of tomorrow where you want to start implementing AI, you can also add another node to that solution that allows you to throw two NVIDIA A2s into the same chassis.
So a lot of the different areas that we’re seeing, some of which I’ve already mentioned, obviously retail. Not only are we trying to create new efficiencies within the store itself, but we’re also trying to help customers augment their customer experience. How do we bring the customer back into the store? And you do that by better understanding more about your customer, who they are, what their demographic is and where they spend their time. These types of solutions, they actually create more and more revenue opportunity for maybe a retailer per patron because we can do targeted promotion, deals, loyalties and rewards. So that’s one example of where you could take one of those systems today and then maybe scale as you go with implementing something like true scale along the way that really helps the customer right size today, but also have that vision for where they want to be in the future.
Patrick Moorhead: So heat, space and maybe as a service.
Blake Kerrigan: That’s right.
Patrick Moorhead: Excellent.
Robert Daigle: Also increasing the GPU acceleration performance, we doubled the amount of GPUs in that system and we’re bringing in the latest NVIDIA L fours. So now you’re talking to something that’s four to five times a GPU performance of what we had in the previous generation.
Patrick Moorhead: I love it.
Robert Daigle: It’s really impressive.
Patrick Moorhead: I love it.
Daniel Newman: I really like the retail application. It’s a great example of where you’ll see a lot of compute horsepower, software. You’re going to see application specific for industries, and you’re going to see generative application opportunities because consumers are going to want what they’re going to want and they’re going to want a better experience in the store to come closer to what we get when we shop at home. So we have a long way to go.
Lenovo’s been trying to really build out its reference with its AI innovators program, big investments in that area. I believe you started it out back in 2022. Talk a little bit about that program and how is it doing? Are you achieving what you set out to achieve?
Robert Daigle: Yeah, it’s been a really successful program since we launched it. We have over 45 companies in our innovators program covering nearly 200 plus enterprise grade AI solutions. Really, the premise behind it is when we started going to market in the enterprise to bring AI capabilities, whether it’s infrastructure or management systems or even applications, really customers were faced with it. They had to stitch this together themselves. They had to piecemeal this together. They were working with multiple vendors. Maybe it was us for the infrastructure, maybe someone else for the management, manageability layer, someone else for the application. That becomes really complex to bring together an AI solution.
So what we’re doing is we’re actually working with these software vendors that have really great AI capabilities, really great AI IP that they’ve developed, and we’re validating that on a full stack solution. So we can show up to a customer and say, here’s everything you need to go do in store analytics. Here’s everything you need to do loss prevention in a retail store. Or manufacturing, here is the full stack to do visual inspection on your assembly lines. It’s really powerful. That has been resonating so well with our customers. And we’ve seen it just take off.
Patrick Moorhead: So Dan and I like to joke that since November we’ve been spending 90% of our time on AI, and it really started off this big, what does it mean to consumer, right? Search engines, things like that. And then it started to pivot to more enterprise SaaS, and then it’s now moving to enterprise infrastructure and on-prem software. But the number one question I get when I’m talking to an enterprise is, hey, this consumer stuff is neat, but until you can help it not drift, and help me find a way to securely use my proprietary data to make it better, it’s really not interesting to me. So I’ll give this one to you, Blake. What is Lenovo doing with the AI plus storage plus data to help solve some of those challenges or answer those questions and just make it simple?
Blake Kerrigan: Maybe I’ll play off a little bit of what Robert said. So one of the things that I think our customers struggle with the most is number one is they have to get to that proof of value. So they’re trying to take a piece of technology and a piece of software and create some residual business outcome. If you can prove that on a bench, it’s one thing. And even if you can deploy to maybe 10 different stores, that may even be what success looks like. But the difference is that doing that in a data center where you have one location with 1000 systems versus 1000 locations and maybe two to three systems running, the things you run into is this issue with scale, right? And it also gets into managing the full stack. So yes, at the application and orchestration layer, there’s things that we’re going to want to do, lifecycle just like we do in the data center. But how do you get from bare metal to full stack at 1000 different edge locations?
So we recognized this very early on in our journey, and one of the things that we’ve been focused on is building a platform, Lenovo open cloud automation tool set that allows a customer to take a system, ship it to a location, securely authenticate it and bring up a full stack and be able to manage remotely from the cloud which applications they’re going to deploy. But then also to your point is what happens when they need to change something? Maybe they want to scale from one ISV that does one particular use case in the store, in the hospital that’s maybe like loss prevention and scale to something that’s more customer experiential centering. So those are a couple of examples of the value that we’re bringing not just in the box, but also around it and to our customers.
Daniel Newman: So something that’s really a red-hot topic that’s not so technical, but it’s a bit more ephemeral is the responsibility, the governance, the ethics around AI. As we deploy this stuff it’s great that we can, but the question is becoming often if we should. When you talk about industries, we’re talking about everything from PII data. We’re talking about wanting our systems of record to be able to reference using things that look like recommenders and filter systems that maybe work for product, but we don’t want it to work when it’s personal data for people. What is Lenovo’s sort of thought? You’re not necessarily building the apps, but you are kind of an ecosystem for picking and choosing. Do you have a position around the ethical and responsible AI that you could share?
Robert Daigle: It’s a timely topic, especially with more and more generative AI capabilities coming in. Really what we’re seeing in this new paradigm of AI, when everyone’s talking about generative AI, really what’s happening on a practical level is we’re democratizing AI. AI over the past five years has taken a lot of effort, a lot of lift. Not that there’s no value in it, but there’s been a lot of lift. If you want to do computer vision, you may have to fine tune a model and have a data scientist actually work on that. With generative AI, you have these large language models that can take more generalized questions and give responses, and then we put it on the internet and give everyone access to it. So we’ve seen this democratization of AI, which is great, but then it brings up a lot of ethical questions when you don’t have a PhD data scientist that’s trained in how to implement AI in a responsible and ethical way, it brings up a lot of questions for how you manage that and how you deal with that.
And I think there’s a couple ways that we’re looking at responsibility and ethics when it comes to artificial intelligence. Last year, we launched our own responsible AI committee in Lenovo. So we’re reviewing any AI solution, our capability that we’re consuming ourselves using internally at Lenovo. Or, if we’re recommending it to a customer, it goes through this governance process and for good reason, because we want to make sure that we are not putting our customers or even their customers and their patrons or employees in harm’s way with these AI systems.
Patrick Moorhead: Yeah, it’s interesting. Huge challenge. Particularly when you operating over in hundreds of companies, it’s like how do you possibly be the traffic cop for all that content? It’s got to be tough. Hey, I’m going to take this great strategic question that Dan asked. I’m going to ask Blake a very straight up, more tactical question, which is why should customers use ThinkEdge? What’s the advantage? There’s a lot of options from OEMs, ODMs, and everything in between. Quite frankly, even the CSPs are getting into the edge game as well because they’re trying to figure out, wait, after 14 years, the public cloud, why is 75% of the data not in the hyper scale data center?
Blake Kerrigan: At the heart of it’s why should somebody choose Lenovo? Like I mentioned, we started our journey about six years ago. To date, we have around eight platforms that we’ve built from the ground up. So we’re talking about nothing leveraged from third party. These aren’t tier three and four different piece partying systems together. So we’ve basically started from the ground up. Now, we’ve also been at it for a little while. So even beyond the six years we’ve been deploying infrastructure, yes, in the back of stores, back of hospitals, but all the way down to where patient data gets collected. You think about the Think client sitting behind the screen at your doctor’s appointment. These are all Lenovo deployments. So we’ve obviously got a lot of experience here. But when you really get down to it, you take a look at the system and yeah, it can look like a black box until you start pulling back some of the layers.
A couple of good examples are physical and digital security. So from a physical security standpoint, when we deploy edge servers, we actually have tamper and motion detection sensors in the system that can actually be factory configured to basically wipe drives should they be detached from a wall or taken too far away from a specific geolocation. Another example is actual tamper detection with the actual physical ports in the system, which is actually unique. So same thing, we can infer what a specific malicious behavior might be, and we can lock a system down or even wipe drives. Another area, and it still amazes me that today in my role that I would actually be contemplating things like acoustics. But one of the things that we found out when we were trying to do some highly complicated AI algorithms at the edge was that, hey, if you put a standard two U rack server with four GPUs…
Patrick Moorhead: It’s really loud.
Blake Kerrigan: Extremely loud and so much so it’s harmful. So if you’re in the back of a store or near a, maybe it could be a janitorial closet, you’re combating with what could be a couple of people having to have a conversation in a hospital room next to a system. So we’ve actually optimized acoustics for the system. That, and I think we continue to push the boundaries of how much storage and compute and acceleration can we put into these systems. And I got to tell you, if you haven’t come over and seen the pedestals, I mean you really start to, it really comes to fruition when you put this system next to one of its data center siblings and it really becomes more real, right?
Patrick Moorhead: Yeah. I mean, it’s apparent to me that Lenovo is making huge investments in the edge and listen, as analysts, we need to be very careful in how we say things for obvious reasons. But your team had said, hey, we have the broadest portfolio of Edge. I’m like, prove it. I want to see, show me the check boxes. How are you measuring that? I’m like, okay, yep. You have the most comprehensive edge portfolio out there by this metric. And I think that says two things. First of all, I see it as Lenovo’s commitment to the edge, but also the reality that the edge is kind of the wild west of form factors. You can go from the edge where they have raised tile flooring, which is kind of like a data center on the edge.
You can go through a fast food drive-through where I can see that server. It’s screwed into the wall right next to the person taking your order to a bank or a trust department where it’s underneath, and they’re looking face-to-face at the customer. So the variability on the edge is pretty crazy. So there we go.
Anyways, gentlemen, thank you so much for coming. I feel like we’ve literally been to the edge and back of AI.
Daniel Newman: Absolutely. I think we should do IoT next. When you want to talk about complicated form factors.
Patrick Moorhead: Well, let’s go another 15 minutes.
Daniel Newman: Let’s go 30. Let’s do it all day. You guys don’t have meetings, do you?
Robert Daigle: This is just intermission for us. We’re ready to go.
Daniel Newman: Exactly. Robert, Blake, thank you both so much for joining us here on The Six Five.
Blake Kerrigan: My pleasure.
Robert Daigle: Thanks for having us.
Daniel Newman: All right, everybody, tune in for all our episodes here at VMware Explorer 2023 Las Vegas. For Patrick Moorhead and myself, it was fun having you here. Subscribe to all our shows, but we got to go now. We’ll see y’all later.