PC Refresh Ahead: AI, Windows 10 Sunset Driving Demand

CRN’s Jennifer Follett spoke with Circana’s Mike Crosby about the anticipated PC refresh cycle and the role the channel will play in bringing AI PCs to market.

[Video Transcript]

Jennifer Follett, Executive Editor, CRN: This is Jennifer Follett with CRN, and I'm here with Mike Crosby of Circana. Mike, so good to see you. How are you doing?

Mike Crosby, Executive Director, Circana: Doing great, Jen, how are you? Thanks for having me.

Follett: Alright, doing alright. There’s a lot of excitement, I would say, as we head into this 2024 year and start to look ahead to the second half. Conventional wisdom says there's going to be a lot of PC refresh activity. How real is that concept?

Crosby: Yeah, and I think we've had this conversation a couple of different times that we believe based on where we saw peak volumes in 2020 and 2021, let's say for Chrome and Windows initially, they typically operate on about a four-year lifecycle, and that's really timing out about now. So we believe really more second half loaded a little bit is where we're likely to start to see the acceleration in the refresh.

We've also got some things that we're up against, like with Windows 10 sunsetting in October of 2025, and larger companies need a little bit more runway. So very quickly, we're edging towards that point of acceleration that we believe is going to be pretty solid.

Follett: In the past, some of these, or I should say the last big PC refresh cycle was a little bit hampered by the state of the supply chain. How are we looking now for that?

Crosby: Yeah, the good news is with supply chain, with unfortunately the byproduct of a of a cooling economy, the positive part of that is we're actually in better shape. Supply-chain-wise, we've seen normalization occur. We've seen rates normalize. We've seen raw materials improve. And certainly just in general, we have inventory levels in line and we see good flow. So from a supply chain, we don't really believe that's going to be anything materially impacting, I mean, barring anything else that happens internationally. But from where we stand today, I think we're set up pretty nicely to have some nice even flow of product.

Follett: You mentioned the Windows 10 sunset. How big a role is that playing in this expected refresh?

Crosby: For B2B it's big. If you look at it, just general estimates are there's still about 70 [percent], a little less than 70 percent of the install base that’s still operating Windows 10 or older. So while we've seen a decent amount migrate over, you're still going to see a pretty substantial migration that's going on. As I said earlier, small businesses, pretty agile. They can move pretty quickly, and in many cases they've already moved because they've bought new product with the updated OS. But if you look at midsize and enterprise, typically they need a little bit longer window. They're making sure and managing through any incompatibilities, any challenges or issues operationally. They want to make sure they've got enough runway before they go hard over and make sure there's not an impact to their business.

So we're going to see that now start to accelerate. And I think what's interesting is with where things are, as we know and we're going to talk about, there's some changes and new products and new technology and innovation that's causing a little bit of a pause to wonder where and how do I go. And I think that's what's going to be interesting because we're also going to get quickly up against it where these large organizations are going to have to start refreshing devices to make sure that they've got the latest and greatest that they need.

Follett: You're talking about the evolution of the AI PC. How are we defining that?

Crosby: It's an interesting question. So what's been fantastic about the term and being communicated in the energy behind it, AI PC, there's a ton of excitement, but I think what we see and understand today, there's less clarity around what that real definition is, not only from NPU or GPU, but specifically is there a minimum requirement, if it's memory, if it's storage, if it's anything else.

So right now, the thinking around the technology, while it's out there and it's exciting and I think people are really grasping the opportunities where it could take us, I think there's still a little bit of a clarity issue and just a clean definition and understanding. What would be maybe an entry-level or a general use device versus a high-end, heavy workload device.

And I think those are a lot of the things that we're talking to a lot of clients about now and just trying to understand and frame and determine what that is. A big part of what you're seeing not only promoted but also what we're also looking at expanded attribution are things like NPU, neural processing unit, that's in some discussions and in some environments working off of a System on a Chip kind of a model. You've got CPU, GPU, an NPU, and that would be one that would be maximizing the benefit of small space and less power usage, etc. But then you have the other camp that might be focused on those things aren't so critical, where GPU can really bring that huge horsepower. So just in those two examples, there's real questions as far as really what that defines.

What I think it's going to come down to though, Jen, is just need and use case. And as we get that clarity around that, I think it's going to really, really help. But as I said earlier, I think the nice thing is there's a lot of reinvigoration around—excitement and energy around—innovation, where I think we saw it early-stage kind of post-pandemic, OEMs were so focused on just getting product to market that I think they kind of took the foot off the gas a little bit on the new and innovative technologies.

Follett: So how big a piece of the sales conversation, as far as the channel partners are concerned, should this AI PC piece be? Is this really something to talk about to a specific segment of client? Is it the kind of thing we're going to see on everybody's desktop at some point? It's really just going to be a small group of people that need these things? How should they be out talking about it to their customers?

Crosby: Yeah, it's a great question, and maybe I can frame it in two parts. The first part we could talk about is if you look at the current positioning of the AI PC and if you kind of visualize in your mind that technology adoption cycle where you typically have early innovators and early adopters on the front end before it matures into the broader market, and we're still at that fairly early stage. So the nice thing is you're going to see people that are willing to test and validate technology. And certainly in B2B, you're going to see a couple of areas that are going to really jump out like health care or banking and finance that are really going to show up well there, but again, it's still going to be fairly small.

I do think, though, accelerating into next year and beyond, we're going to start to see a broader deployment of AI PCs. You're also seeing a pretty quickly expanding portfolio across OEMs. So you're seeing a much richer pool of products to be able to pick and choose from. And I think as we continue to see that defined, it's just going to get better and better.

But to your point earlier, I think initially it's going to be fairly isolated. My expectation, I believe it's going to be a little bit less on the horizontal position where it's just Copilot and looking at those types of tools and resources to accelerate productivity and optimize workflows. I think you're going to see some of those verticals like I suggested, health care, banking, finance, really look for the heavy horsepower in really driving things around new drug development, you're going to see finance really work on risk assessment, risk management and monitoring where there's going to be a lot of horsepower required to look over vast amounts of data and move very quickly. So it's going to be an interesting kind of a shift. Is it going to go narrower and more higher-end initially, maybe test and validate? And then as the market broadens, you're going to see it really spread across the lion's share. But I think that's kind of the expectation that we have so far.

Follett: How much hesitancy do we expect to see from the buyer that’s like, “This thing is cool, and I know I'm going to need it, but it's going to develop so quickly that this first round, may be obsolete in four months, right? Let me let me wait and see what comes next.”

Crosby: Exactly right. And see, if you think about that tech adoption curve we were just talking about, that innovator, early adopter, they're willing to try just about anything. And cost is really less of a factor. But there's that high level of interest that I need to be first to market on the latest technology. That's just a fairly small part, though.

But I think to your point, and I've had a number of conversations with some of our biggest and best channel partners across the board, and they're hearing what's been interesting is these mid-sized larger companies that have much larger deployment pools of devices to refresh are kind of struggling with this in-between. And what I mean by that is there's kind of three things that are coming up: One, you have CFOs and CTOs saying, one, yes, they're likely going to be more expensive and am I going to get the return that I need on the added spend? That's one issue that they're questioning and considering. Second one is, because it's so new, to your point earlier, am I going to actually create my own problem a little bit from compatibility issues? Am I going to run into things where I haven't had the opportunity to test and vet and really make sure that everything is operating within my environment the way it needs to be? It's another huge risk. And then the third one is do I have the talent? Do I have the people and the skill set to be able to leverage what that technology is supposed to bring from an ability to optimize my business?

So you've got CFOs and CTOs, as I said, kind of looking on and leaning on the channel to provide recommendations and guidance because they're struggling with it. If I bet wrong and I run into these issues, is it going ultimately to compromise the performance of my business today, or if I go the conservative route and my competition goes forward with it, am I now going to be disadvantaged for some period of time where I'm going to have to work my way through and I'm going to have to play a little bit of catch up? So this is where they're really leaning on the channel heavily in looking for that recommendations and guidance and understanding because, I think, it's just so new.

But it's an interesting challenge because everyone is super excited, as you know. And I have probably no less than four or five conversations a week just specifically on AI PC impact, adoption, where we believe it's headed, who's likely to be first, who's likely to lag. So it's a fairly regular set of discussions.

Follett: And leaning on the channel, that concept carries over to the technical side, too, right? Like they might not have the technical expertise themselves in-house, but does their MSP have access to those technical skills to help them roll out these AI PCs?

Crosby: Absolutely. And I think this opens up just a ton of opportunity. Again, that’s what the channel does and does best is they find and they're able to create and craft opportunities around solutions and everything from training and education, installation, maintenance ongoing, all of those things. But also think about even as you look at the level of sophistication by companies, here's where I think small businesses are going to really rely heavily on the channel for that training, education, guidance, maintenance, monitoring, all those different things, all the value-add pieces that really reinforce that equation.

Mid-sized definitely as well. But again, as you get larger, many of these companies also have their own infrastructure internally to be able to manage and operate, etc. But I definitely think there's an augmentation opportunity for the channel in both of those as well, because, again, they're going to be on the front edge of best practice. And I think that's what companies are looking to do, is not learn as they go. They want to make sure that they can articulate what their environment is, good, bad or otherwise, and rely on the channel partner really to be that that consultative solution that can provide, really, “Here's really how this needs to be done,” and in a best practice kind of way.

Follett: You touched on the pricing a little bit earlier. How much of a gap, how much more expensive are these AI PCs expected to be?

Crosby: Yeah, it's a great question. My belief is initially, like in many things that you see, the narrower the vertical, and, again, you're not going to see, in my belief, not a huge momentum early in 2024. You're likely to see premiums paid above that could be as high as 20 [percent] or 30 percent.

Now, that'll be interesting because it'll be [a question of] how long will that be sustainable? And I think once we get the market maturity, we start to see the volume expand and we see that horizontal expansion, then clearly it's going to put downward pressure on pricing, and you're likely to see that premium that maybe is being willing to be paid early probably dissipate a pretty good chunk.

But again, we're going to be in the early stage in this refresh. We believe, as I said, second half, 2024, it'll peak likely in 2026 and then start to normalize again as we come down on the back side of the curve. But it's going to be, again, I think the timing is very, very good. And I definitely think it's an exciting time for companies to really find new ways that they can leverage technology just to grow and expand and to increase their profitability.

Follett: How much of that price premium do you think partners will be able to keep as their own margin? You know, the hardware space is not notoriously great for high margins these days.

Crosby: Well, I mean, and to your point, right? Yes, definitely there's going to be some benefit on the hardware side. But I think where they're going to make it is either on pre-sell or it's going to be on post-sale support, it's going to be config, it's going to be other things that again, the channel is really, really good at. And I think they're going to find that, from a blended opportunity, they'll definitely see some benefit on the hardware end, but to the rest of it where you can blend all the rest of that up, I think it's going to be a nice opportunity.

What's going to be interesting, though, is how does it change by size of business. If you look at it, are small businesses going to be leaning on and saying, “Look, I can't afford to buy it this way now,” but now is this more subscription-based? Now does this become Device as a Service and does this accelerate that, where look, I can have now the latest and greatest first-to-market technology. It's a managed operating expense that I can manage through the period as my business either flexes up or flexes down versus having to have a significant amount of outgo to buy the product, and hopefully that I bought the right product that I can maximize my business. So you might even see some shifts and changes in the way that this is going on.

The other thing that we're hearing, too, is, you know, there's a lot of questions that we hear on why do you need the AI feature set if it’s NPU? Why do you need this locally? Why do you need it natively versus just using the cloud? And there's certainly benefits, and you're going to see both being used. And in fact, if you look at Copilot today, Copilot today, the majority, if not all of it, is done in the cloud. So it's something that certainly is scalable. But as you look at certain verticals that there's security-level requirements involved or there's things that they want to make sure from a standpoint of issues of latency or anything else, there are specific kinds of use cases that are going to require that I need to have this technology natively. At the same time, I think what you're going to see is you're going to see a combination of both, clearly. You're going to see the cloud being utilized in some of the tools that are out there and available that are going to be AI-driven. But you're also going to need to leverage the technology locally to do the things that need to be done either offline and/or within security requirements, etc. So I think you're really going to see a combination of both.

Follett: You've spent a fair amount of time looking at the different market segments, both size of business and vertical markets. Where are you expecting to see the sweet spot for these AI PCs, particularly in the early days?

Crosby: I think from the standpoint that if you look at it from, let's go this way first, we'll go small, midsize, enterprise first, then we can go to vertical.

If you look at from size of business, even though we know that small businesses are extremely agile, they can move very quickly, one of the challenges that small businesses have continued to have in the current environment is clearly inflation, cost of capital, labor, being able to maintain and retain talent for their business. So what they're really looking for from a small business [perspective] is predictability, and I think there typically, especially as we start to climb out and inflation continues to improve and everything else, you're going to see small businesses get in, but they're likely going to get in a little bit softer and a little bit slower. So I think you're going to see fairly low adoption initially on small businesses, mainly because I think they're still just trying to manage through the current level of challenge and trying to find that predictability. But then I think when you see it mainstream a little bit, I think you're going to see a nice acceleration on small business. And I think a lot of that's going to be in things, simple things, repeatable, redundant things, that are needed to ultimately where you can leverage an optimization, where you can take the cost out, you can make it more efficient and more effective is likely the areas where you're going to see that occur within small business.

The real sweet spot, I think though, is going to be that midsize and that enterprise customer. They have the ability again, from an operating flexibility. They've got the capability to do that. They can take a little bit longer view. I think once they get past the compatibility and the test and validate stage, I think you're going to see a much more accelerated onboarding of these technologies, and you're going to see this being maximized.

Now, hopefully we're not going to see just a significant kind of job loss that's going to be associated with AI, because we're already seeing some of it in tech right now, engineering, software design, etc., coders. Hopefully we're not going to see that in some material way that moves forward, but likely you're going to see some of that erode and some of that revamp on the hiring, because you're going to see businesses operate more efficiently and more effectively.

Definitely, if you look at it from that perspective, I think with mid-sized businesses, sales and marketing, you're going to definitely see it in engineering and supply chain. You're going to see it in areas where really you can find efficiencies in economies of scale. And I think that's the goal. That you'll be able to do more with less, ultimately, I think is some of the benefit that we're looking at.

But if you look on the largest companies, and these tend to be, I think again, they've got a longer view, they've got deeper pockets, they've got the ability to take a long view on these things. I think that's once we get past the incompatibility risk issues, you're going to see that broaden fairly quickly. And I think within these largest of companies, we're seeing a lot of innovation around that heavy workload, like within, as I mentioned earlier, verticals like banking and finance and health care, because they're really going to take advantage of the horsepower that AI can bring. But also the ability to optimize the rest of their business overall. That's going to be everything from supply chain and how much efficiencies and forecasting and planning, and it's going to be transportation and planning and forecasting again on that aspect of it. New product development is going to help accelerate where they can bring products to market in a faster, more efficient way, predictability financially.

So all of these aspects that I think are really just going to make companies operate significantly more efficient, and hopefully it gives the ability for them to be even a little bit more agile because they have the ability to use some of this from a predictive sense to try anticipate where some of the things are going within the broader market. I think that's some of the big things that that's likely going to help.

But I would say again, just to recap, I think small is going to be a little bit slower. I think mid-sized are going to be a little bit faster, maybe moderate. And same with enterprise, and it's really going to be tied again to once we can get over those initial questions of making sure I'm not creating my own problem. Do we have high levels of compatibility? Things were operating well and it's going to integrate within this.

Now, as I mentioned earlier, though, timing is kind of the essence, especially for the mid-sized and enterprise, because as we said, they don't necessarily have the luxury of waiting a long period of time because these are devices within their install base that are timing out that need to be refreshed. And with that Windows refresh, that's coming migration to [Windows] 11, you're up against it. You need to make sure that we've got that operating and operating effectively within that window of time where Windows 10 will still be supported to make sure that I've got these other products that are on and moving forward.

So I think at the end of the day, there's an urgency on the part of these larger businesses to really [say], “Let's get through the detail and understand clearly risk and opportunity and do we have what we need,” and again, this is where I think the channel can really, really provide a lot of guidance, I think, to these types of companies and it's all good. It's all good for everybody, I think, in that situation.

Follett: What about guidance as far as privacy concerns and security issues? Where should the channel be looking in those areas as they're advising these clients?

Crosby: I think, again, those are really huge. These are these are those nice areas of specialization. And what we've started to see already with the channel, as we know over the years, it used to be where they did a lot of the core things, but you're also now seeing a lot of partners really developing some really nice vertical kind of value propositions to companies and customers that not only improves their margins overall because they're effectively able to drive higher margins, but also they drive a lot of loyalty within those things. And I think it really builds a nice strength in their customer base, and it becomes a very loyal customer base because the partner is an integral part of really how these companies operate.

So security is significant. It's not going to be less. It's going to continue to be even more. And I think in making sure that the channel really has as a high degree of education and knowledge and understanding of these latest technologies, it really puts them ahead to really be that consultant, be that be that advisor to these businesses that think they know what they need, but I think they really rely on the hand on the rail of the partner to really understand, “OK, I have confidence now that this is really where I need to go and what I need.”

Follett: Is there a way yet to size the amount of services drag we're going to see with this, how many other services partners are going to be able to tack on here?

Crosby: I think it's early, but we are working on it now internally. And I think here as we come up, we might be able to come back with a follow with you and then be able to share a little bit. But we think it's going to be significant, and we think it’s going to be lasting. It's not going to be something that's just an initial hit and then we'll see that kind of fade a little bit. We're going to see acceleration, and there's so much we don't know already about AI today and what's been developed, let alone what will be developed as we see the technology really start to mature and understand the nuances of all these environments.

Watch the acceleration on new app development and new technologies that's going to be coming. I think it's going to be exciting to be a part of it. But I think it also where you're out on the front end of that, there's going to be a lot of change and again, a lot of reliance on the channel to really be able to provide that level of understanding and knowledge because I think companies are getting out on the wing, so to speak, relative to their knowledge and understanding of some of these technologies, and they need to maximize their competitive advantage. At the same time, when you're out in front that far, you know, you run some risk and you really need to have confidence in the partners that you're working with that can provide you with that.

Follett: So all in all, we're saying this thing's going to start kicking off at the end of 2024, particularly with some early adopters and mid-size and enterprise customers, really start to develop nicely in 2025, and then peak and say, what, mainstream in 2026?

Crosby: Yeah. And when I say peak, that's really what we believe is the lion's share of that four-year period of refresh. You're likely even to see a little bit of a secondary wave because if you look on the Mac side, which is also interesting, we look at Mac, and that tends to operate on about a five-year lifecycle, not four, and their volumes didn't peak until a little bit later than, say, Chrome and Windows. So you're likely to see kind of a camel hump in this refresh. Initially, it's going to be more Chrome and Windows, but then you're going to start to see Mac start to accelerate as well. And I know that in the press as of recently, you've seen a lot of communication around [Apple’s] neural engine and AI, and there's more announcements that are going to be coming.

So again, I think you're seeing the heat really come up, you know, relative to competition. And I think, again, it's going to be an exciting time. But look, for the second half to be that initial part, as I mentioned before, I think a couple of times, Jen, is that the real catalyst, I think for me, to be is going to be the Fed taking their first action. If we look at, if we start to see a quarter point cut on interest rates, that's going to be dropping the flag relative to, I think, midsize and enterprise really starting to lead that initial lot.

So we're watching that closely. We're still hoping that June, July is that timing of where it is. But I think watch that closely, and then watch the acceleration and then we'll see where we go from there.

Follett: Oh, we'll be watching, Mike. We will be watching. Thank you so much for spending some time with me and sharing your insights on the AI PC market. Really appreciate it.

Crosby: Great, love to be here. Thank you again.