Memory Supply Worries Amid PC Market Refresh Initiatives

In this episode of “TBR Talks,” Principal Analyst Angela Lambert and Senior Analyst Ben Carbonneau discuss how supply chain dynamics are shaping the emerging AI PC market. They explore the reasons behind slower-than-expected AI PC adoption and the shift in memory production capacity.

Additionally, the pair looks at how specific companies, including Dell Technologies, HP Inc. and Lenovo, are adjusting to current market conditions and delves into the biggest supply chain unknowns for 2026. 
 
Episode highlights:

• The two major factors converging in the PC hardware space
• Memory capacity and supply chain disruption
• The ways specific companies are adjusting to current market conditions

 
“I think the big question is, what will happen next in that supply chain that the AI servers are consuming too much of, I guess. Will it be storage, for example, or other, just other components that, you know, a computer’s just a tiny server or vice versa, right? So, it’s a lot of the same suppliers and manufacturers. I think that while we’re grappling with memory now, I wouldn’t say that we’re going to go back to normal immediately after that. We’ll probably just be on the lookout for what’s the next domino to fall here is probably the most likely scenario,” said Lambert.

 
Listen and learn with TBR Talks!
 

Submit your Key Intelligence Questions for Patrick and his guests
 
Connect with Patrick on LinkedIn

 
Learn more about TBR at ⁠⁠⁠⁠⁠https://tbri.com/⁠⁠⁠⁠

 
TBR Talks is produced by Technology Business Research, Inc.
Edited by Haley Demers
Music by Burty Sounds via Pixabay
Art by Amanda Hamilton Sy

 

 

‘TBR Talks’ on Demand — Memory Supply Worries Amid PC Market Refresh Initiatives

T

TBR Talks Host Patrick Heffernan: Welcome to TBR Talks: Decoding Strategies and Ecosystems of the Globe’s Top Tech Firms. Where we talk business model disruption in the broad technology ecosystem from management consultancies to systems integrators, hyperscalers to independent software vendors, telecom operators to network and infrastructure vendors, and chip manufacturers to value-added resellers. We’ll be answering some of the key intelligence questions we’ve heard from executives and business unit leaders among the leading professional IT services and telecom vendors. 

I’m Patrick Heffernan, Principal Analyst, and today we’ll be talking about supply chain threatening the rise of AI PC with Angela Lambert, Principal Analyst for TBR’s IT Infrastructure Practice, and Ben Carbonneau, Senior Analyst for TBR’s IT Infrastructure Practice. 

Two major factors converging in the PC hardware space

Angela and Ben, welcome back to TBR Talks. And today we wanted to go into, dive deeper into a special report that you both co-authored recently, just published, about the supply chain, but really the state of the AI PC market. Angela, you want to sort of set it up? What was the special report all about?

Angela Lambert, TBR Principal Analyst: Sure, I’d love to. So, the special report we published is, it’s really, as you said, a combination of two major factors converging in the PC hardware space. And I think maybe it’ll help if I take us back in time about a year. So, over the past year, there was a lot of expectation and anticipation of large growth in the PC market. Microsoft is ending support for Windows 10. That’s supposed to urge a lot of commercial customers to refresh all of their PC fleets, drive a lot of growth. At the same time, there’s the concept of AI PC that entered the market. They have arguably more capability, more efficiency, and kind of new hardware that has yet to be fully understood and adopted. 

What we saw last year was that the PC market did grow fairly well for a very mature market, you know, in the lower mid-single digits. But it definitely was not up to the expectation of what the PC vendors and what other vendors like Microsoft, Intel, and AMD really probably wanted to see because they’ve really been investing significantly in the AI PC. But unfortunately, I think what we saw in engaging with the market is that the value proposition fell flat in concert with tariffs happening, which increased the prices people were paying. And the AI PC, it’s more sophisticated. There’s a more sophisticated hardware, the CPU, NPU in there, it costs more money. And vendors were really relying on the concept of future proofing. Not necessarily that people are ready to use AI today, but you should buy these devices because once they are ready, they’re going to need them. So that message did not really take off at the same time that all of these other kind of cost implications were coming into play. 

So that kind of leads us into 2026, where vendors have made significant investments in the next generation of AI PC. These investments have been going on over the last few years, really leading up to this, but they’ve been really working hard to make this a success. And I think in many ways, AI PC needs to be a success for them. But at the same time now, we saw tariffs last year. We’re seeing a really significant upheaval in the hardware market at large because of huge shifts in memory availability and pricing.

Patrick: Before we get to that supply chain part of it, the things that you described that prevented the surge that was expected in the PC market were things that are more of a delaying factor than a diminishing factor. So, the high price, the slower adoption of AI PC. Is your sense that while 2025 may not have been the banner year that was expected in PC sales, that maybe 2026 will be? Or are you about to tell me that all the supply chain problems mean that we’re looking to ’27, ’28, ’29 before we see that surge in growth around PC sales?

Angela: Yeah, I think for a few reasons, the surge in growth will be delayed. And there’s- yeah, part of it is the delay because of cost, but COVID changed so many behaviors and patterns in so many ways. And I think buying PCs is another thing that has changed. It really hasn’t come back to normal. Before COVID, it was so predictable, every 3 years, a company was going to refresh their hardware. They’re on a three-year four-year cadence. That’s kind of changed. People are not necessarily committed to that. Maybe they had huge inventories they were digesting, but it’s still just really not clear to the market at large when or if the kind of normal cycle that used to exist will ever come back.

Memory capacity and supply chain disruption

Patrick: Right, and so speaking of COVID, supply chains were also massively disrupted through the pandemic. And is that sort of now playing out again, Ben? And as we look at the chips side of this equation.

Ben Carbonneau, TBR Senior Analyst: I think so. I think during the pandemic, there was a significant tailwind, which was an increase in the total addressable market for PC. And I don’t see that falling. I think what happened there was that there just wound up being a lot more PCs per household because there’s so many people working and learning from home. So, I don’t think that number decreases. But to what Angela was saying, I think the rate at which those PCs refresh is the big thing that’s changed. And I think that’s due some to the cycle and delay of purchases, but also due to the hardware just becoming such that it doesn’t need to be replaced on that same three-year cycle that it needed to be before. 

So, with memory now, what I think I’m seeing or what we’re seeing in the market, I know Lenovo just reported earnings and they were saying that memory costs in the last quarter of 2025 increased 50% year-over-year. And I think that’s- those kind of huge increases are still occurring on a quarterly basis. And what’s happening there is really not a decrease in the manufacturing capacity for the memory that’s going into PCs. It’s more of a reallocation of new capacity. So if you’re a memory manufacturer, which the big three would be Micron, Samsung, and SK Hynix, instead of allocating more production capacity to PC memory, like they would in the past when there would be a PC cycle like this, they’re actually allocating that new production capacity to high bandwidth memory. And that’s going to support more of the AI infrastructure workloads. So that’s a really big, I think it’s a big thing to understand- I think, is that the supply of PC memory isn’t going down from where it was last year. It’s just not growing at the same rate that demand for PCs would be. And as a result, PC prices will go up. And we think that refreshes will be even further delayed.

Will AI PCs take over the market?

Patrick: And do you think sort of the AI PC is something that’s just going to simply take over the entire market that you’re not going to see a refresh of a, I mean, the equivalent of a dumb PC at this point, I mean?

Ben: I think- so I think at some point, the AI PC will take over the market. I think our forecast for the Windows AI PC market by 2030 looks something like a little over 80% of that Windows PCs sold in that period will be AI PCs. I think there’ll be certain markets that are slower to adopt AI PCs. I think we call a dumb PC a traditional PC, but more traditional PCs certainly in emerging markets.

Patrick: When you say markets, do you mean specific kinds of clients, enterprise versus small/ medium? Or do you mean geos or do you mean industries? What do you mean by markets?

Ben: So, I think it’s a little bit of both. I think we see that certain industries, I think it comes down to price sensitivity. So certain industries are more price sensitive than others. For example, retail is always going to be more price sensitive. They’ve always had a longer refresh cycle than a mature enterprise in, say, North America. And then in emerging markets, so Latin America, I think we see, and then outside of China and APAC, there is more price sensitivity. So, in those markets, I would see slower adoption of the AI PC. However, something that Angela was saying, we know that these silicon vendors are working on a third generation, at least for the x86 guys, so Intel and AMD, are working on their third generation AI PC chips. So, as the older generations of AI PC chips become older, they also become cheaper to manufacture, all else equal. And that makes those AI PCs based on those chips more accessible.

Patrick: That’s crazy that we can already be talking about the traditional, older, dumber versions of AI PCs. We haven’t even gotten to AI PCs in everybody’s hands. That’s nuts.

Ben: It is, yeah. I think adoption has really been- it hasn’t been what the OEMs wanted it to be, but it has been, I mean, fairly strong. I know Lenovo was talking about 30% of their unit sales being AI PCs in most of the quarters throughout 2025. So, we do see AI PCs being adopted. But I think, again, what Angela was saying, it’s less on the capabilities and the need for those features today and more about future-proofing your device fleet for what is to come.

How specific companies are adjusting in the current market

Patrick: Right, yeah, that makes a ton of sense. And so, you mentioned Lenovo a couple times. Angela, are there other companies, again, TBR, we look at the companies. So, are there other companies in the special report or more broadly in the space that you’re keeping an eye on for the changes that they may have to go through in order to adjust to what we’ve been talking about?

Angela: Absolutely. So, we, of course, look- the report does show a little bit of our data on the big three, right? So, Dell, HP Inc., and Lenovo. We also look at the chip vendors as well. So, Ben hinted at this when he said the X86 guys, but what he’s really saying there is, there’s more change even within the chip space here than there’s been in probably the whole time I’ve been covering the devices market. So, Qualcomm is a relatively new entrant in this space. They are long time servers of the smartphone chip market, but they’ve partnered with Microsoft to try to bring some new things into this space on an ARM architecture. So, there is, besides just the PC vendors trying to aggressively, you know, win and maintain share as they always do amongst each other. There’s also some changes in terms of the competition of what the longtime leader Intel has seen in this space, not just from AMD, who’s been their primary competition, but kind of a new category as well. And this is, it’s definitely a big time for Intel. They are taking on a new strategy of bringing their PC chip manufacturing to the US. And it’s really important for them for that new production to be successful. So, there’s a lot of dynamics at play behind the scenes here.

Patrick: So, a couple of years ago, Intel was definitely struggling. And were sort of the company that everyone will point to for how to go wrong in the chip space and the manufacturing space and just the tech space all the way around. Has it changed for them? I mean, are we looking, is Intel well positioned for the changes that you’re expecting across the market?

Ben: I think their strategy is well positioned. And I think the stock price kind of tells the story there and what we’ve seen in the change over the last year in the price. Where really Intel, I think, when they came out with their foundry business, which was essentially making chips for external customers, I think that didn’t accelerate in the way that they wanted it to. And it was really misses on certain process nodes and being delayed. But I think right now, we’ve kind of pivoted more back to, I wouldn’t say completely back to the old Intel ways, but more to kind of favoring the production of chips to be put in their own devices as a proof point to external customers. And I think that was a really strong move to what Angela was saying earlier. Their 18A process is what the compute node in their third generation AI PC processor will be based on. So, all that manufacturing coming to the US I think is really big. And I think that will draw in external customers. And I think that’s why you’ve seen kind of, they’ve iterated, they’ve called it IDM 1.0, IDM 2.0. They’ve talked about these strategies in their earnings calls. But I think this strategy is probably the strongest, kind of going back to Intel’s roots. being able to actually execute on the 18A process, which I think they’ve proved that they have. And I think this third generation of AI PC chip will further demonstrate that. So, I think it looks good for Intel.

The biggest supply chain unknowns

Patrick: That’s good. Well, that’s encouraging. And earlier you said that when talking about the supply side of things, it’s not like the physical supply is in trouble. It’s more the demand is shifting in terms of where it’s going. And then Angela, you mentioned tariffs as one of the complications in the supply chain discussion. Going forward in 2026, are there- is tariffs the biggest unknown when it comes to what can happen in the supply chain? Or are there other things that we should be anticipating? Whether it’s manufacturing in the US not standing up fast enough, whether it’s demand falling off a cliff, like what are the other things that might disrupt the supply chain? Or is tariffs sort of the biggest unknown at the moment?

Angela: I’d say tariffs, I think that buyers and companies have wrapped their heads around the tariffs to the extent that they can at this point. And it was a lot of uncertainty at this time last year within the PC space. Now, with the just skyrocketing price of memory, I think that’s the biggest thing we’ll be seeing over the next few months through this entire year. Maybe not better next year either. But I think what, to me, the question would be is, how will the AI server market continue to impact PC? Because really, this is all being driven by that insane demand for AI servers. So, memory is kind of the first major domino to fall here, where now, as Ben was saying, if the component pricing goes up 50% for Lenovo, that’s an expensive piece of the build. That makes very material increases to the end customer. I think the big question is, what will happen next in that supply chain that the AI servers are consuming too much of, I guess. Will it be like storage, for example, or other, just other components that, you know, a computer’s just a tiny server or vice versa, right?

Patrick: Right.

Angela: So, it’s a lot of the same suppliers and manufacturers. So, I think that while we’re grappling with memory now, I wouldn’t say that we’re going to go back to normal immediately after that. We’ll probably just be on the lookout for what’s the next domino to fall here is probably the most likely scenario.

The case for AI PC may come from security benefits

Patrick: Right. That’s not- and so as encouraging as what you said about Intel was, that’s the opposite of encouraging if we’re thinking about the dominoes that are going to fall. Just to wrap it up with something personal. So, are either of you using an AI-enabled PC right now? Have you tested some, played around with some? When do you think you are going to get your hands on one full-time?

Angela: Oh, boy. We have been lucky enough to see some really cool devices in our travels. We haven’t been lucky enough to receive them here at the office, but it has been very cool to see some of those devices. And yeah, I certainly look forward to it and getting that to myself.

Patrick: Is AI slowing down your current PC? Is that part of the way that you’ll get a faster refresh? Because in order to use all the AI tools that are out there, you actually need a PC that can handle it. I’m saying that’s a very loaded question because I know my own laptop is slowing down.

Angela: Right, we’re all we’re all hungry for the new laptop.

Patrick: Yes.

Angela: You know, it depends on your use case. Admittedly, a lot of what I do with AI today is cloud driven. I think that’s true for most people. But what the devices can also provide and what I think Ben and I feel is going to be the super strong AI PC use case is going to be security. It’s not necessarily going to be, you know, you could be a data scientist who has some models on your computer, you’re running your private AI-driven analysis that’s compute intensive. For most of us office workers, we’re going to see it in kind of behind the scenes performance enhancements, but security is such a big threat, just the unwitting incidents that happen with end users. That’s a huge threat to every company. I think we’re going to see over the next couple of years some really impressive use cases come out of that, where that, I think, might be the thing that turns the tides on selling companies. It’s not maybe going to be a crazy change in productivity. It’s going to be a really great advancement in protecting the company’s data.

Patrick: And so does sovereign AI and sovereign data come into this as well, where, because one of the things we’re seeing increasingly is this idea that data needs to stay in country, that data needs to stay particularly as to go back to tariffs, but almost all the companies we’ve been talking about are US companies. There’s a sense in Europe and even in Asia that they need to keep the data there and keep AI there. So, do AI PCs play into that as well, where that’s part of the equation, especially from the security standpoint?

Ben: I think AI PCs play into that a little bit. I think, though, what we’ve seen, at least with sovereign AI, or what I’ve seen, is that there’s been some impressive private cloud solutions that have been built for companies, and just by leveraging your private cloud server cluster, you’re gonna get a lot more compute out of that than you would from any NPU for on-device AI, so I think I guess the wrap would be that the AI PC, I see the NPU on the AI PC SoC really driving improvements kind of behind the scenes to what Angela was saying. Maybe some security integrations from Microsoft along the way that get pushed into operating system updates. But really more of a, right now, I think a battery life savings situation. And for sovereign AI and data privacy, I really see that mostly being driven by the architecture of private cloud solutions.

Final thoughts

Patrick: Excellent. So, let’s end with a prediction. How soon until the three of us are sitting here doing an episode of TBR Talks that you guys are looking at your AI PCs in your laps as we’re sitting here? How soon is that coming?

Angela: I’m going to optimistically say, let’s say six months from now.

Patrick: All right, mid 2026. Okay, Ben?

Ben: Mid 2026 would be nice. I don’t know. I guess it depends on the-

Angela: We’ll see how those prices go up.

Ben: Yeah, we’ll see the prices. We’ll see what our IT administrator does. I would love to be working on an AI PC right now.

Patrick: All right. Well, I’ll make sure he listens to the podcast. Excellent. Thank you, Angela. Thank you, Ben. It’s a lot of fun. Thanks. 

Angela: Thanks, Patrick. 

Ben: Thank you. 

Patrick: Tune in next week for another episode of TBR Talks. Don’t forget to send us your key intelligence questions on business strategy, ecosystems, and management consulting through the form in the show notes below. Visit tbri.com to learn how we help tech companies, large and small, answer these questions with the research, data, and analysis that my guests bring to this conversation every week. 

Once again, I’m your host, Patrick Heffernan, Principal Analyst at TBR. Thanks for joining us and see you next week.

TBR Talks: Decoding Strategies and Ecosystems of the Globe’s Top Tech Firms

Join TBR Principal Analyst Patrick Heffernan weekly for conversations on disruptions in the broader technology ecosystem and answers to key intelligence questions TBR analysts hear from executives and business unit leaders among top IT professional services firms, IT vendors, and telecom vendors and operators.

“TBR Talks” is available on all major podcast platforms. Subscribe today!