I was supposed to say that, Tim? Sorry. The mic just came on.
We're going to start. Good morning, good afternoon. I'm Tim Arcuri. I'm the Semi and Semi-Equipment Analyst here at UBS. Very pleased to have Lam Research. We have both Tim Archer, who is the CEO of Lam Research, and we have Doug Bettinger, who is the CFO of Lam Research. So thank you to you both.
Yeah, great to be here. Let me kick us off with a safe harbor just to keep our attorneys happy, and you can see it on the screen, but for those of you on the webcast, today's discussion may include forward-looking statements or subject to risks and uncertainties. Actual results may differ materially, and so forth and so on, so please have a look at the Safe Harbor. I don't think Tim or I will say anything new today, but just in case we do, this is the guidance for you. Okay, Tim.
Perfect. Well, Lam has been my favorite equipment stock for a long time, and the Wafer Fab Equipment market looks like it's going to grow quite a bit the next few years. I think you're kind of pointing to a first half of next year that's going to be flat to modestly up, with growth being better in the back half of next year. Can you just talk about some of the drivers of that acceleration into the back half? Is it mostly due to DRAM and fab readiness issues, advanced logic timing? Maybe you can talk about that.
Sure, Tim. Well, first, thanks for having us. It's great to be here. And it is a very exciting time right now in semiconductors and semiconductor capital equipment. When we look forward to 2026, we talked about the first half being kind of flattish to a little bit up. I mean, that's obviously coming off of what has been looking to be a strong second half of 2025 for us as well.
And really, I think even since our last earnings call, we've continued to see increasing strength in terms of optimism about semiconductors and also the role that technology that Lam provides plays in a lot of these inflections that are coming and how they're enabling this whole AI environment right now. And so I think we come into 2026 feeling very positive about Lam's opportunity.
Yeah, Tim, it's just timing. You know how this business works. It's the totality of everybody's spending plan. It's flat to a little bit up, and then it's just second half weighted is what we see.
I mean, you pointed out clean room. We mentioned that on our last earnings call. We said the reality is you need time to get infrastructure in place. And I think that this demand has been building across the ecosystem. And I think it's going to take some time for some of these elements to come into play. And so I think that bodes well. We've often said we don't really want demand to far exceed ability to supply for too long. We'd like it to be stretched out, manageable, and executable by the entire industry.
And when you look out over the next few years, which do you think, from a relative perspective, if you took NAND and DRAM and foundry? I know if you drive out, we were talking about this yesterday, drive out to see TSMC's campus, you can get pretty optimistic about advanced foundry over the next few years. But what are the relative puts and takes in each of those three markets over the next few years?
You know, I think that, and I'll let Doug make a comment, but I would say what's exciting about semiconductors right now is that all three device segments have a tremendous growth opportunity ahead. They all play together. I mean, you're seeing that come together right now. You've got advanced foundry logic and the role that it plays, obviously well-known role it plays in AI. You've got DRAM being driven by HBM. You've got Enterprise SSD driving a lot of the upside we're seeing in NAND right now. And so I think we're at a moment where there is end demand at the leading edge, especially for all three device segments.
If you go back to what we presented at our investor day last February, Lam's growth story, our SAM expansion is really coming from this change in device architectures, integration approaches that is driving a material upside in Etch & Dep intensity across every device type, across every technology node. So Etch & Dep intensity, Lam's opportunity just keeps getting bigger. And as people try to innovate and drive higher performance across the AI ecosystem, that bodes well for us from a technology inflection, technology evolution perspective.
Let's talk about NAND upgrades. I'm sure you've gotten this question all morning from today. So you gave this number of $40 billion to upgrade the existing installed base to get it to 2xx. I had $45, but they're very close. And we're probably, I'm guessing, about a third of the way through that, maybe a little bit less. But that's not a static number either, right? Because that just gets you to 2xx.
And then, of course, now we're seeing a push to go even higher than 2xx. Can you talk about just the dynamics of what's going on in NAND? Obviously, I think most of the wafer capacity is going to go toward HBM. So it's hard to see a lot of wafer capacity added. So there is pressure. The only way that you're going to be able to add bits is to increase layer count. So can you just talk about that?
Sure. Yeah, so back at our investor day in February, we talked about this $40 billion upgrade number. We said at that time, likely it would play out over several years. And that was to upgrade the installed base. Lam has the largest installed base of any supplier within NAND. And said it would play out over several years because our estimate at that time was that bit demand was likely growing in the mid-teens. Well, obviously, we've seen, especially with drivers like Enterprise SSD, that bit demand is somewhat higher than that now.
And that's, I think, going to encourage people to upgrade a little bit faster than we might have seen at that investor day. That's partly because upgrades themselves, they are the fastest and most economical way for a customer to get to those higher layer counts, higher performance devices that are needed for those Enterprise SSD applications. So when we look at it, there is a conundrum that some of our customers, I believe, have, and they've talked about, which is, again, clean room space.
You've got to make a decision. What at that moment in time is most important? You're adding HBM. You're adding NAND. The NAND upgrades provide a really nice opportunity to upgrade in place and get to those higher performance bits. So I think that we'll continue to see that happen as sort of first priority. But eventually, you have to start to add capacity. As we said on our last call, there's some clean room space that's going to need to be brought online to be able to accommodate that, but I think the most important comment you made was, this isn't static.
We said $40 billion in February, but once you get to $200, you've got to get to $300, and then you've got to get to $400, and that's simply because these leading edge applications require higher bandwidth, faster read-write speeds. They require higher densities to create the enterprise solutions that are needed today, and so this doesn't end with one upgrade cycle. It will immediately morph into the next upgrade cycle, into the next upgrade cycle, and in each of those, Lam is the winner.
Yeah, and I think maybe you can also double-click on just it quite literally is the most fertile ground of any pool of money being spent in the industry. It is the most fertile ground for any single supplier is your capture rate of those NAND upgrade dollars. I think it's about $0.40 on the dollar. You've not given a number. I don't know. Does that.
All we've said, Tim, is when upgrades happen, two-thirds of our SAM is what's exposed to that. And we have very high share. So yeah, we love the upgrade process. And by the way, you also get some new equipment that shows up in there, right? You've got to upgrade to Moly for the QLC stuff initially. And then you always get a few constraint tools that need to get purchased. So it's not just upgrading. There's also some new equipment that goes into that number.
Yeah. I think it's available on our website. But we published about a year ago this article on the path to 1,000 layers. Now, obviously, that's still a bit ways out there. But it talks about all of these different applications that ultimately come in. Doug just mentioned Moly for higher performance, lower resistance within the structure itself. That's necessary for the read-write speeds.
But from a density perspective, we end up, once you end up in the 200, 300, 400 layer devices, you're starting to stack tiers on top of tiers. And what we've said is that device manufacturing is becoming increasingly complex, especially in this vertical scaling. And vertical scaling, the verticalization of everything in semiconductors today, whether it's more layers in NAND or it's advanced packaging across foundry logic or DRAM, even in NAND, the cell bonded to array, all of these things increase Etch & Deposition intensity. And that's why we think that Lam has a very unique secular growth story within our SAM, which is sort of unmatched amongst equipment suppliers today.
Yeah. I mean, so much of your overall WFE share depends on NAND because that's your highest share market. But if I just take the different pools, I take the foundry logic pool, and I take DRAM, you're gaining share in both of those pools too.
That's right.
So a lot of people think of you as like the NAND guy, but you're not just a NAND guy. And so can you actually talk about that? Because some of the other companies in the space haven't been able to gain share in those pools the way that you have. How have you been able to do that? When you sort of think back, when you targeted these applications, how did you do it?
Yeah. Well, I would say it's the culmination of a strategy that we embarked on probably six years ago to somewhat better balance the company across all three device segments. We saw the inflections that were coming. 3D NAND was a huge watershed moment for the company because 3D NAND is so Etch & Dep intensive. But when we looked six or seven years ago at the inflections that would be coming and the verticalization of foundry logic, the verticalization of DRAM, we saw the same opportunities.
And so we've spent the last five or six years developing products specifically to target smaller, taller features within foundry logic and DRAM, as well as newer materials that address challenges like RC, basically the speed and power aspect of devices. And I would just say it's fantastic execution by the company on those new products. And we've talked about a lot of those wins this year, but it's things that you might not have thought about for Lam in the past. Low-k ALD within foundry logic, our dry photoresist process within both DRAM and foundry logic.
These are like new areas for us. We said that Lam has the opportunity through these inflections across all three device segments to expand our SAM from the low 30s to the high 30s by the end of this decade. And that's a material jump in terms of the opportunity for us. And because of the progress we've already made on those products, we said we would win more than 50% share of that newly created SAM for Lam. And you ask why, simply because almost all of the new technology inflections are either 3D transitions, so the verticalization, or they're material changes. And those are both very, very good for an Etch & D eposition-focused company.
Yeah, I mean, if you think about your overall WFE shares, 12 and a half, 13, 14, depending on the year, I mean, that's a significantly higher number than that if you think of your incremental share of this new SAM, way, way higher than what your total share is of the WFE market. Can we talk about China for a few minutes? Just like everybody else and all your peers, you're guiding China down next year. Of course, that's what all the companies thought a year ago too, which is not an indictment against you.
And the year before that, as you were pointing out to me yesterday.
Yes. It just is the way that it's turned out. But China always seems to surprise to the upside. We've had this, there was the BIS affiliates rule that is costing you $600 million roughly next year.
But, Tim, that was a revenue statement, not a WFE statement, right? Because the WFE is still the year. When we talk about it trending down, we're talking about WFE.
Yeah, correct. Correct. But I guess part of that is as those companies are allowed to take tools, that is something that you'd think that much, if not more, gets added back because they only have a year's reprieve. So they're going to try to get as much from you as they can in a year. So it seems to me like there's a lot more upside drivers for China WFE next year than not. So can you talk about why you see it down?
Maybe I'll start, and then, Tim, you can add on. It's a numerator-denominator impact as well, right? We're talking about a percentage. You know how strong everything else is. We've been talking about it, strength of AI, the investments leading edge foundry and logic, leading DRAM, NAND. That's a global statement, right? So understand, we believe that grows. And I know you do too. And just when we look at the totality of the spending in China, it looks like it's going to soften a little bit.
The last couple of years, you're right, we've gotten it wrong. I think everybody has. You're hearing all of us describe a view that it's going to trend down. So that's what we all see. What's happened the last couple of years is one of the customers in China, or maybe a couple, ended up spending more than they suggested at the beginning of the year. We just don't see that happening in 2026. We could be wrong. But that's just we're describing what we're hearing from our customers, Tim. That's all.
And can you talk about this? So this is more of a revenue question. You took $200 million out of December for the BIS affiliates rule and $600 million out of next year. And we were talking about this yesterday. But now that this rule is reversed, you probably gave away those slots for December, maybe even for March.
But that $600 million at least comes back. And it probably comes back more toward the back half of next year, I would think. And so that's the first question is sort of what's the timing of that to come back? And then two, why would it not come back at a number that could be significantly higher than that $600 million? Because they only have a year's reprieve. So they're going to rush to get as many tools as they can as fast as they can.
Might it be more than $600? Maybe. It's spread through the year, Tim. I mean, it's not simply back halfway. They're just going to be all the way through the year, I think, best we can tell. And might they try to spend more because they're worried about the rule going away? Maybe. It's too soon for us to give you color on that right now.
Yeah, I guess I would just say the only thing I could add is that when we look at China, we have a great team there that supports the customers that we can sell to very, very, very well. We have great products that are targeted towards those trailing edge applications. And when you have great people and great products, I mean, you win. And so some of our performance in China is also just share gains as well. And it's something that, again, we're applying just, like I say, great people, great products, and doing well.
But overall, I mean, we spend a lot of time talking about China. And I understand it's important in the short term. But I'll bring you back to the fact that Lam's real growth story is all about leading edge. It's about the progress we've made in foundry logic, DRAM, the enabling role we play in higher layer NAND. I think the one thing that I don't know if it gets enough attention, but I mean, the fantastic job we've done at advanced packaging.
I mean, we've flipped the script on leadership in advanced packaging as a result of our etch tools, our copper plating tools, the role we play in things like wafer shaping or stress management, the dielectric deposition applications that help enable. We announced one earlier this year, late last year, Inter-die Gap Fill. These kinds of applications and the role that Lam plays, I mean, it's become a game changer for us from a revenue perspective. We haven't quite updated the numbers by themselves at this point. I mean, there's some competitive aspect to that. But we had already said, I believe, last year, over $1 billion.
Over $1 billion.
Over a billion dollars. We gave an update that Gate-All-Around plus advanced packaging would be well over $3 billion this year, and I think just from what you see going on in foundry logic and in DRAM, advanced packaging is just a growth driver for Lam going forward. And it didn't really exist if you just went back five, six years ago when we sort of embarked on this strategy of doing well in foundry logic and DRAM. Advanced packaging, it's a foundry logic and DRAM primarily focused area, and that's why it really bumped our performance in those two segments and much better balanced the company today than when we started.
Great. I wanted to ask about three things that are pretty unique to Lam. One is the move to 4F² in DRAM. The gates get a lot more vertical. Anytime anything gets more vertical, that's good for.
Good for us.
Second edge company. That's one. Transition to Moly number two and Dry Resist number three. So I could spend, I mean, we could spend the rest of the time on those. But maybe you could touch on each of those. And you could just talk about why those are so important and why they're unique to Lam.
Yeah, sure. So I mean, as we said, anything that is getting smaller, taller, it's great for us because what we specialize in in the etch space is basically etching very small, very deep features. And 4F² is basically a new device architecture in DRAM that creates very vertical, very small features that require a precision etch. And it's both precision from the size of the pattern, but also the selectivity and control you have for the depth of that etch.
We introduced a new tool recently called Acara with direct drive. And again, forever, Lam has been pretty much the world leader in conductor etch. But even for us, this was a pretty big breakthrough, which was the first time we've been able to now control without a mechanical, what's considered to be a mechanical match. We use solid-state drivers to match the plasma to the chamber. And this allows us to control the plasma conditions 500x faster than had been previously possible.
And why this is so important is, as the features become smaller, you can sort of imagine if you're pulsing this plasma and you're trying to change the chemistry and the species inside that plasma, the faster you can get control of the plasma and the power, the better you can control that etch. And so it's become very, very important in Gate-All-Around as well, I mean, basically in logic where you're trying to create very small pattern features. And then it will become very important in 4F². 4F² is still a couple of years out, but it's one in which already D2R decisions are being made today.
We feel really good about our position. Very similarly, Moly, I mean, on the NAND side today, and eventually it'll be on foundry logic and maybe eventually on DRAM as well. But Moly is a tool where it's a metallization change. Lam's been the leader in tungsten metallization up to this point. We're very strong in ALD metals. And so as this inflection comes to reduce the line resistance, help you stack taller features, Lam's taken a leadership position there.
We've basically been the company that has been winning the initial positions within the Moly transition. And then Dry Resist, kind of brand new innovation. It's not often that Lam has come up with something completely new to the industry. When you introduce something brand new, it takes a while to get people sort of on board with that. And so we are a few years late, I think, to our original projection of revenue. We did recently announce this year that we're in high volume production now at a major DRAM maker.
And that's a really important step because DRAM runs a lot of wafers. And that basically will prove out the Dry Resist application. We announced a very important partnership with JSR, which they are a leading provider of photoresists. And we're going to work with them on not only precursors for the Dry Resist, but also precursors for ALD and atomic layer etching. And I think that, again, bringing together equipment suppliers like Lam with these unique capabilities and then material suppliers, this is kind of, these are very important partnerships for us going forward. And we look to do that across the ecosystem.
I mean, if I think about some of the numbers that you've given for Dry Resist, they seem a bit, I mean, it's great. But those numbers seem a bit low to me. If I look at how big the track market is, I mean, it's huge. And you're basically replacing wet photoresist application. So why can't we look at the size of the track market as a guide for how big that could be for your business? I mean, the market's significantly higher than the $1.5 billion number you've given over the next five years.
Yeah. Maybe just a matter of time, Tim. Let us hit the $1.5 billion first. And then, I mean, it is kind of one of those cases where, look, change comes hard in this industry too. I mean, some of these tools have been in place. And that's been the technique for decades and decades and decades. And so that's why I say sometimes that switchover takes time. But what's nice is when it switches over, you don't go back. I mean, you've made that choice to go to Dry Resist for the improved pattern fidelity, the improved productivity you get from EUV.
And I'm satisfied if customers just layer by layer, generation by generation, keep adopting. If it comes faster, I'll be even happier. But what I'm really encouraged by is the fact that you're starting to see that switch over. You're starting to see the commitment in high volume production, and the way things work in this industry is once a few start it and you're seeing success, it sort of starts to snowball, so let's see where it plays out, but we feel really good right now about the momentum in that space, and we'll update you next year and the following years about how big that market's getting.
Doug, I want to ask you about one of your favorite topics, which is CSPG, which is your service business.
My favorite part of the company from a business standpoint. I always say that.
So before I ask about the overall business, one more innovation you've driven is this use of cobots in your service business.
Cobots are a part of what we characterize as advanced service. There's a bunch of equipment intelligence-enabled service offerings that we bring. But cobots are a key part of it for sure, Tim.
So, can we just talk about A, how that's helping CSPG? And really, the bigger picture question is just to break down the parts of CSPG and talk about what you think the long-term growth rate is for that business.
Yeah, let me unpack CSPG in total. And then I'll come to the advanced service. And I'll let Tim maybe talk about the cobots.
He's got to let the cobots. It's my job.
Four things in CSPG: spares, service, equipment upgrades, which is doing really, really well this year, and then the Reliant product line, which is our older equipment that we sell. When you look at the profile of this, the great thing about this is our equipment almost never goes away. So the opportunity to do more of this because chamber count grows every year, that's part of what drives CSPG. Spare parts intensity goes up as more advanced tools come out as well.
But what we're really excited about is what you initially asked about: cobots and some of the advanced service offering. What we're able to do for customers here is predictable, repeatable, consistent service to the tools. We can match chambers better. We can enable yield enhancement, utilization improvement. This is a unique thing that we've brought to the customers that they're actually really excited about.
And I'll let Tim talk about how he hears customers pulling on this. But this is part, Tim, of how we grow this business. And we give you the 2028 model. We said CSPG will be 1.5x as big as it was last year. And then by that $1 trillion model, it'll be double. Part of how we drive dollars beyond just the growth in chambers is with advanced services. I don't know if you want to talk about what that means.
Yeah, no, that's great. Thanks. I think that if you go back, we talked about this point of key strategic shifts we made. In 2019, we launched our new Sense.i platform. And part of the whole idea with Sense.i, when we launched it, we hadn't changed our platform for etch for 20 years. It's kind of like if it's not broke, why change it? But what we recognized was that the trend was going to be for a platform that could collect tremendously more data for every wafer run, collect tremendously more data off of all of the components that were running in the tool.
And in 2019, while we put all those sensors in place, I mean, we're still in the nascent stage of being able to actually utilize all that data. I mean, what we're excited about now is, I mean, look, that's what AI, that's this whole AI revolution is fundamentally how do you use data to come up with new insights. And so our equipment intelligence, our tools are getting ready.
And so I think that's going to be a big driver for, like Doug said, really shortening the time to troubleshoot tools, to install them and match them in these big fabs, in fact, to match them across fabs that happen to be on different continents, which we're seeing increasingly happen, including right here in Arizona. And so the EI piece we're feeling really good about. But we're super excited about the innovation in cobots. And initially, we introduced cobots with this idea that if you could do robotic maintenance on the tools, it would be great for people who, like right now, there's a workforce shortage.
We talk about the fact that there just aren't enough engineers if all these fabs are going to be built in all these different places around the world. So we kind of initially thought about it from that perspective. It could replace some of the labor that our engineers don't really like doing and our customers' engineers don't like doing. But what we've actually now found is probably the even greater value proposition is the precision with which the maintenance gets done.
And you kind of think about if you're going to try to build something at the 2 nanometer or 1 nanometer level where every angstrom matters, well, having the maintenance be done exactly the same every single time and not relying on every single engineer being trained to do it exactly the same, that cobot does it every time exactly the same. And so we've seen the first time right after maintenance be dramatically better. Some customers have reported some yield improvement because they don't get wafer to wafer variability like they used to between different chambers, different machines.
And so we just see it as something that almost now goes hand in hand with this desire to accelerate node changes and technology evolution. And so you might think about it as CSPG. But I think that having cobots is actually going to make our systems also that much more attractive for a customer who's thinking about building these mega fabs and thinking, if I buy tools that have cobots, how much easier is that going to make my mass production ramp? And that's time to ramp, time to mature yield. Those are things that make money for our customers. And it's a real priority for them. So cobots, I think, is going to be a big thing for Lam and for the industry in general.
And maybe just last question, Doug. Can you talk about gross margin? I know gross margin depends a lot on mix, really. And you're fabless, really. So you don't have a lot of fixed cost.
Very small fixed cost here, yeah.
But can you talk about where you think margins can go? I know one of your peers talked about raising prices. You haven't really talked about that, whether you've done that or not. I don't know. But you're pricing to value, I think, pretty consistently.
Yeah.
Can you talk about that? Just gross margin, how much upside is there?
Yeah. I mean, the way I want people to think about the financial performance of the company is the model we put out in February is still the right way to think about it, which suggests 50% gross margin in the 2028 time frame and greater than 50% by the time we hit this magical trillion-dollar industry. We just put the plus next to the 50. Some of the things that drive improvement in gross margin are some of the new tools that Tim's talked about, right?
They're better technical performance. So we're going to get better gross margin, we believe. The close to customer strategy of ramping the Asia factory network, well, a lot of it is already in the P&L. There's still a little bit left to go. So that will also lead to some level of improvement. And yes, pricing is always something w e're doing our best to get fairly paid for the value we're delivering to the customer. That's always in the mix, and you're always negotiating that, and so that's all part of how we are going to deliver that financial model.
Great. Well, we've run out of time. So thank you to you both.
Thanks for having us.
Thank you, Tim.
Thanks.