 theCUBE presents Dell Technologies World, brought to you by Dell. Welcome back to Dell Tech World 2022. This is theCUBE, alive. My name is Dave Vellante. We're here with our wall-to-wall coverage. This is day two. We actually started last night, the theCUBE after dark. John Furrier's here, Lisa Martin, Dave Nicholson. We're going to talk about Apex, the business value of Apex, flex on demand. Darren Fedorowitz is here. He's the Senior Vice President of Dell Financial Services. And we're joined by a customer and a partner, Judd Barron, is an R&D infrastructure architect at Silicon Labs, and Steve Fezen is the regional VP of Compia Center, Comp-Computer Center. I say that like I'm from Boston. Guys, welcome to theCUBE. Thank you. Darren, take us through what's going on with Apex. You got custom solutions. You know, people are going to ask, is this just a financial gimmick? What is this? No gimmicks, no gimmicks, Dave. So I think when we think about technology, historically, customers purchased. They bought and they owned, and they may have financed it and paid over time, but it was really an ownership model, especially in infrastructure. And Apex is about subscription. So think about Dell Apex as you can either buy or you can subscribe to your technology. And under Apex subscription, we have options for custom-based solutions or an outcome-based. And I know today we're going to talk about flex-on-demand and custom-based solutions. So it's a high level, pay for what you use, when you use it, with a high level of choice and flexibility. All right, Steve, I'm going to ask you to play a little co-host with me, okay? So add some color, color commentary. Judd, tell us a little bit about Silicon Labs. I'm really interested in what your requirements were, your challenges and kind of why you landed on Apex. Sure, Silicon Labs is a semiconductor company. We're headquartered in Austin, Texas, just under a billion dollars a year right now. And at any EDA shop, or people who are doing electronic design automation, it's not just in the semiconductor industry, but we have these HPC farms where we're running millions of jobs a day. And the balance that you have to strike when you're doing capacity planning in one of these environments is we have these things called tape-outs, and that's where we're finishing a project and there's a much higher volume of jobs that we have to run. And you have to decide, do we buy for peak or do we come under that sum amount and say, oh, we're going to buy 80% of what we think the peak is going to be. Overbuy, overbuy for peak. Overbuy for peak normally, right? Correct. The hard one is do you overbuy or do you underbuy? It's always a hard decision. There's a trade-off, right? And so the challenge there is that you'll end up kind of lengthening the time and potentially miss a tape-out window. And there's costs associated with that because you work with the foundry and you kind of schedule based off that tape-out when you're going to deliver the photo mask to them. Anyway, the point is we, in the past, using a traditional like Camp X, we're going to buy a bunch of servers. We tend to undershoot whatever our peaks are because we may have a peak every couple of months during these tape-outs, but sometimes tape-outs slip and so one slips two months, another one comes in a little bit early and now you have multiple tape-outs in the same months and what was going to be a small difference from peak to what you actually purchased ends up being a big peak. And the thing that was interesting to us about Flex on Demand is the ability to have a commit rate that the customer can work with Dell financial services to figure out is it 80% or 60% or whatever and they give us additional servers that we pay just when we're using them. I'm somewhat oversimplifying the process, but we've got to talk about that. But the point is, if I understand it correctly, that infrastructure was dominoing the time to tape-out in a negative way and you've been able to address that more cost-effectively. It can, it has on occasion and so this basically gives us a lever to pull to say, well we can spend some additional OPEX this month and open up this additional capacity. So it's not like bursting to the cloud exactly because you have to have the equipment in your data center already for you to be able to use it but it's under a traditional acquisition model, it's just not a thing that was available to us before and looking at leasing or other types of financing wasn't really attractive previously but the Flex on Demand model when we first heard about it we're like, that's very interesting, tell me more. And we ended up using it in Austin and then we built a whole data center in Asia and did the whole thing on Flex on Demand and got it. Okay Steve, talk a little bit about your role, what's going on at Computer Center and YAPEX, do you give us the background? Yeah, Computer Center is one of the largest global vars on the planet, right? We have a lot of global and international reach but at the end of the day it's about one-on-one customer relationships, talking to them, understanding what their challenges are and we've had a multi-year relationship with Judd, I've known you for a long time and typically that relationship or initially that relationship was about collaborating, working hand in hand, kind of figuring out what the solutions were that best fit their environment to solve the issues they need and it was typically a procurement, a purchase-based relationship and it worked well for a long time but when Judd posed the challenge to us about kind of more pay-as-you-go subscription-based modeling for how he wanted to acquire in the future, we just, we huddled with the Dell team collectively and talked about what we could offer and how we could solve the problem. APEX is a really nice brand today but this was two and a half years ago, okay? So we were a little early on putting it together, I feel good that we were able to put that type of solution together for Judd and it's working today, working wonderful today and it was good for, it's good for, whenever it's good for the customer, the manufacturer and the partner all together, it's a wonderful solution. So you took a little risk, but it worked out and you helped reduce that risk. Yeah, that was probably the infancy as we were growing our as a service. Think of this, you know, there's a lot of big words out there, Dave, right? As a service, utility, cloud, it doesn't matter what it is, it's super cloud. It's super cloud, it doesn't really matter. This is really, Judd was talking about a really important element, which is around flexibility, choice, there's uncertainty oftentimes in an environment but they want to control. They still want to have a level of control and leveraging partnerships, being able to deliver flexibility and choice. Don't worry about the words, don't worry about cloud, utility as a service. We end up solving the customer need, right? And when we talk about flux on demand, I'll give you a little bit deeper into flux on demand. So when we think about flux on demand, it really is about understanding the customer needs and our capability and Judd referenced this, determining what a baseline is. So if you think about your own utility bill, right? You go home and even if you're on vacation for a month, I'm sure you went on vacation for a month, right? One month at a time, hitting a Y, yeah, I know. But if you leave, your utility bill, even if you don't turn on a Y, you still get a utility bill. It's your baseline. So we determine a baseline with our customers, with computer center, to understand in your environment, you're going to use this minimum amount and that becomes your baseline. And that baseline can go as low as 25% and up to 80%. In a server environment, it usually is typically in the 70, 80% and then we determine what is going to be optimal based on that 25 or above. We charge based on the usage on a day-to-day basis, averaged by a month. And if you go up one month during your peak, you get charged at that peak. If you then a couple months are lower, then you're going to pay only for the usage. And so for a customer that's growing, has variability or seasonality, this is a great model because they can still control their environment either within their own domain or in a colo. They also have the capability to pick anything within the Dell ISG catalog, any product, configure it to meet their environment, be able to work with a trusted partner like computer center, that it's a Dell solution based on a partner relationship and delivers choice and flexibility on the catalog of anything Dell sells within your control of how you can configure it. So it gives this ability to say, instead of buying and instead of paying a predictable payment, i.e. a financing, I'm going to pay for use. If I turn on my light switch more, if it's during the summer in Texas where I am, the AC is a lot higher, so your utilities go up. And if you are a much lower because you're on vacation in Hawaii, maybe you've been on vacation in Hawaii for a month, you're going to have a much lower and you're going to hit your baseline. So it gives flexibility, choice, and it gives the control back to the customer. I see the whole ISG portfolio. So you're like the tip of the spear for future Apex. We absolutely are the tip and that's why, Steve referenced a couple of years ago as we were still in our infancy growing, listening to our customers, listening to our partners. We've evolved to become a more robust program. 35 countries today. So we can cover 35 countries over the globe, all ISG products that are sold with a high level of flexibility in it. And it's Judd and feedback over time that we've continued to evolve this program. So Judd, if I understood correctly, the business impact to you was better predictability. You didn't have to over buy or under buy and take all that risk. Is that right, maybe you could quantify, did you ever quantify that? Or what can you tell us about the business impact? Yeah, sure. So I mean traditionally we will base our capacity demands on a complex calculation that effectively just boils down to number of engineers, like headcount, and kind of personas within that. And we figure out, okay, well, how many compute cores do we need? And then we say, okay, well, how many tape outs are we doing and when are those tape outs gonna land and try to figure out which months are gonna be the hot months? And the design teams have to kind of vary their tape out schedules so that they don't pile up all into like July or something. And then there's not enough compute capacity. So with something like Flex on Demand where I can turn additional capacity on in our HBC farm, we just go in and make some changes to the LSF configuration and say, hey, now you've got these extra nodes available. We don't really have to worry about that as much. In fact, last year we ended up with one month where for us it was unusual, we had five tape outs all land within two weeks of one another. And they all finished, which in previous years before we had deployed that would not have been the outcome, things we would have had multiple tape outs delayed and that's a seven figure impact for each one of those commits that we miss with the Foundries. So it's a big deal. Yeah, that's real dollars. It is. And you know what else? As Joe was going through this, we all know there's supply chain constraints and this solves a lot of supply constraints because, Joe, if you would be purchasing today, you'd be buying, you're looking ahead and you're actually having to purchase today where if you go into an apex flex on demand, you don't have that full commitment of having to purchase but you can get ahead of the supply chain. So you can be looking six months in advance, you can be doing capacity planning. And I'm sure you're doing that, leveraging like what's my future and not be worried about, I have this huge burden up front. Yeah, and I mean, we have two levers right now. One is we have this extra capacity there. I can pick up the phone and call our Dell rep and say, hey, I'm going to modify my commit rate. And so now that's the new baseline I can use all day every day. And we still have some burst ability. And then separately, we can say, we want to expand the contract or basically acquire more hardware for additional burst or additional commit. Both of those things are options. We only had the, we had to go buy it and we need to know when we have to have it available. So you kind of back into this ordering schedule for like a traditional capex purchase. So Steve, obviously Silicon Labs is leaning in. Are you seeing any other patterns in your customer base where this is being applied? What can you share with us there? I believe this is a fairly horizontal solution. Any customer can really utilize it. I mean, traditionally people would buy for two and three years worth of capacity and slowly consume it over time, but you paid up front, right? That's how it, that's kind of how it worked because they didn't want to go back to the well year after year after year. So, you know, and I think, I think, if anything the cloud, the hyperscalers has taught the world some things, taught the industry some things, you know, in a perfect world, customers would like to consume and pay for what they use, you know, and in the increments that they use it as much as possible, as closely aligned to that as they could get. And what I see, what I see in this, you know, because I kind of put, in my role, I'm putting solutions and customers and bringing those together, right? And complimenting that with services of our own. But what I see over time, that almost all the manufacturers, and Dell does a wonderful job, but almost all the manufacturers will be delivering technology on a subscription basis. So the more I learn, the more I know, the more I understand about how to deliver those and provide those to customers, the better off we are. Because it aligns with business value, and that's what we're seeing, Judd. Steve made an interesting comment in there, he's talking about the cloud, and for us, there's always pressure to say, hey, you know, can we burst in the cloud? And for EDA workloads, every time we look at this, it's a data problem. It's not a computing problem for us. EDA workloads tend to generate a lot of data, and you know, there's a lot of tools, there's just a bunch of stuff that you have to have available to run those jobs, and so you have to look at that very carefully. The company that I work for, Silicon Labs, has been around for a long time, and we have a lot of development effort that's been put into automating and simplifying things for our design engineering, and trying to manipulate that and make it to where we can burst just certain jobs out to the cloud efficiently and cost-effectively, hasn't really resonated for us, but the flex-on-demand thing gave us the ability to kind of achieve some of that burst ability, I mean, not to the same level of scale, of course, but we can do that at our own speed, in our own data centers, with our own data, and we don't have to worry about trying to peel an onion and put something new together to make it cloud-friendly. It's substantially similar. We got to go, but Darren, bring us home. Yeah, I think when we think about Dell, it's about listening to our customers and our partners, which we continue to do, we continue to evolve our products, and Apex is around choice and flexibility in delivering to customers an option to pay for what they use. It's a great solution, appreciate the time. Guys, great conversation. Thanks so much for coming on theCUBE. All right, thank you. All right, and thank you for watching. This is Dave Vellante from theCUBE. We'll be back with more wall-to-wall coverage. John Furrier will be back. Lisa Martin, Dave Nicholson, you're watching theCUBE.