Great. Good morning, everyone. I'd like to get started. My name is Toshiya Hari. I cover the semiconductor space, here in the U.S. Very pleased to have, Doug Bettinger, EVP and CFO from Lam Research today. Before I dive into questions.
First, let me encourage you to take a look at the slide that's behind me and in front of me. I'll read a little bit. I always like to keep my attorneys happy. These are cautionary statements. I may discuss some forward-looking items. I encourage you to take a look at our safe harbor language. You can find it on the investor relations website at lamresearch.com or at sec.gov. By the way, I don't plan on giving you anything new today, so I don't know that I really need to read this, but anyway, that's out of the way, Toshiya, and we can jump into whatever you would like to talk about.
Okay, great. I think typically, these kind of things are 30 minutes-35 minutes, but we have 45 minutes.
Awesome.
I wanted to start off with a softball. So I think you joined Lam in March of 2013.
You're making me feel old.
Yeah, yeah.
But yes, that's true.
Right.
It's now been 11 years, yeah.
I think revenue is up 4x, EPS, 10x.
Depends on the starting point and the ending point.
Yeah.
If you go to the peak in December 2022, it's actually up more than 5x.
Five X.
But anyway—
Okay.
I t's been an amazing time for the company, and we can talk about some of the things for why that has been and why that we will continue into the future, I think.
The stock is up 22x-23x, I think. But I guess, what are you most proud of in terms of what you and the team has accomplished over the past 10+ years? And more importantly, you, Tim, the broader team, what are you focused on? How are you spending your time?
Boy, that is a very broad question. First, I'm very proud at the execution this company has delivered, and hopefully, if you've been an investor with us along the way, you've done real well, and I see a couple of people in the room that I know have been with us the entire time. And you know, I always feel the real sense of obligation to make sure that we're doing the right things for the company long term, and I believe we've done that. Listen, when I joined the company, I remember, you know, I'm a numbers guy. I built a spreadsheet, right?
I understood, at least I believed that I understood all the inflections that were going to happen, 3D NAND, as an example, all the patterning that was in front of the industry in NAND or in DRAM and foundry and logic. I built a little model as best I could put it together, and I'm like: This is a great company. It's uniquely positioned to take advantage of the inflections, the broad fact that things were going 3D. I looked at it and said, "Great opportunity." We have exceeded my most optimistic outlook for what I believed was going to happen. And I'm really proud of what the company has done. We saw these things.
I would suggest to you the timing of bringing Lam and Novellus together in 2012 could not have been more perfectly timed for the changes that have happened in the industry over the last decade plus, right? The fact that we have a materials company and an etch company now as one company, we're better together, right? Because you put the films down the wafer, and then you have to etch those films, and you do it again and again and again. The interrelationship between those things has been critically important for us being as good as we've been and for us performing as well as we've performed, and that five times revenue growth is all about that. I see these things continuing into the future. Now, I can't promise the continued— that number magnitude.
You guys will build your own models, and we all get our models wrong, but I am very optimistic about the positioning of the company. The evolution of architectures going 3D and being materials-enabled continues. Gate-All-Around is on the come line or maybe happening today. That's a 3D structure in Foundry and Logic. Advanced Packaging, HBM, that's a 3D structure in my mind, right? The Chiplets architecture, Die Disaggregation, that's a 3D structure. It's being worked on right now. Eventually, the architecture in DRAM will go 3D. That's still, in my opinion, pretty far away, but it's being invested in and worked on today. And so the evolution of what I see continuing out into the future has us, I believe, again, I'm maybe a little bit biased, as the best-positioned equipment company in the industry because that's what we do.
When you look at, I guess, where the leadership team is spending a lot of time, it depends, truthfully, on where you sit in the company, but we've got brand-new products coming out everywhere to take advantage of these technology changes, Gate-All-Around, Backside Power, the Advanced Packaging, Deep Silicon Etching, Copper Electroplating. You know, we're investing in the technology. As business got soft, as you went into 2023, we positioned the company, I think, to be better when growth resumes. We set up our very large new factory in Malaysia, which is much closer to customers. We invested in a lab in Korea, close to several of our customers. We invested in a lab in Taiwan, close to customers. We continued to invest in the lab outside of Portland, near two of our very large customers.
So that, I believe, is something also we're spending a lot of time thinking about, right? The lab strategy to be close to customers. We believe that is a clear differentiator for Lam in terms of how we think about our R&D footprint. I could ramble on and on. We spend a lot of time, again, depending on where you sit in the organization, focused on what your function needs to do. That's not what this group is interested in. But I guess if I would just pause, technology leadership, getting ready for the inflections we see, and operationally making sure the company is set up to be as good as we possibly can be. Those are the things we're focused on right now.
Awesome. Thank you. So, Doug, before Lam, you were on the device side, so you've seen many cycles over the years.
Sadly, that is true.
Sadly.
Yes.
Any fundamental similarities or differences when you compare and contrast the cycle you're currently in with past cycles, and your ability to sort of foresee, you know, customer projects and what they're working on today versus, say, five years ago, ten years ago? How has that evolved?
I guess I'd say a couple of things. You're right, I've seen a lot of cycles. I started in this industry in the early 1990s, and we always get to the point where we think: Hey, the next cycle is gonna be less than the ones that came before it. I no longer say any of that. You know, the industry has consolidated, that is unquestionably true. And you might think, okay, fewer players, the cycle should be, the amplitude should be less. It's not turned out to be the case. So I always assume there's a cycle that's right around the corner. Invariably, it's all about inventory build, right? And I don't think the industry, in and of itself, does this purposefully. In fact, I know that it doesn't, but we misread demand as an industry, as a collective industry.
And there's a motivation when pricing is really strong, you want to go get as much business as you can get, and you need capacity and investment to do that, and invariably, it just overshoots. And inventory builds, pricing gets weak, profitability degrades, and then ultimately, it always turns, right? For every downturn, there's always the upturn. It is a—i t's universally true in my 30 years in the industry. It always happens. This one, to me, feels maybe a little bit more elongated because we had COVID show up, right? I think a lot of the industry players, I don't know, shame on us, we mistook the demand that was pulled forward during COVID as some aspect of secular growth, and it didn't turn out to be that. I mean, it wasn't purposeful.
I think there was a year where, I don't know, PC units were 350—
Fifty, 60.
Yeah, and that was. We were all sitting at home. Our kids were going to school from home. We all bought new computers, and it pulled forward a couple of years of demand, in my opinion, right? And we're now back to 250, 260, something like that.
Yeah.
So that increased, I think, the overshoot, and maybe this is a little bit longer cycle as a result of that. But you're starting to see it turn, right? Especially in memory, you're starting to see pricing get better. You're beginning to see utilization tick up a little bit, and invariably, those are the early indications that the upturn is not too far down the road.
Got it.
If that helps. I'm just rambling on here.
No, no, it's helpful. Thank you. Twenty twenty-four WFE, on your earnings call, very mild update. I think you're thinking low- to mid-$90 billion range.
That's the new outlook, and so we upticked a little bit. And I think, frankly, what happened is, it wasn't in our market. We realized the litho going into China. We had underestimated it.
Right.
And so as we saw, that print and whatnot, we're like, "Eh, okay, we need to update the numbers." It really wasn't in the etch and deposition addressable market.
Mm-hmm.
That we saw that.
Got it. Then on a year-to-year basis, you know, your expectations or outlook, leading-edge logic foundry versus DRAM, versus NAND, versus mature?
Yeah, so let me unpack it a little bit.
Mm-hmm.
I mean, leading-edge foundry and logic is actually pretty strong this year, and it's the investment in the Gate-All-Around node. You know, everybody has different nodes that they're talking about, but it feels like it's gonna be a very robust node. There's a level of excitement in the industry, especially from the big participant that's playing those wafers, and this is gonna be a pretty robust node. So that's what's showing up there. The trailing edge investment, or I call it mature nodes, or sometimes we'll call it specialty node investment, sort of soft this year. It was very strong last year, maybe partly as a result of the supply chain challenges. You remember, everybody, we couldn't get semiconductors for a while, and a lot of it was off those process nodes. So the investments last year were very strong.
And we're in a little bit of an inventory cycle there. When you listen to all the players in analog, industrial, automotive, and so forth, that's probably well understood by everybody in this room. And so there's a little bit of a pause globally, anyway, in the investment there. Maybe with the exception of China, but China mature node investment is still pretty, pretty robust.
Mm-hmm.
And I'm sure you'll ask me about China later. I'm happy to talk about that, but a little bit softer. DRAM is okay, right? So in DRAM, you have a little bit of, maybe more than a little bit. You've got strength with the customer in China, and you've got real strength in high-bandwidth memory, in DDR5. Right? There's a product cycle in DRAM that's happening, and, I'm excited about that, right? If you're looking for green shoots in memory and storage, that's where you see it. High-bandwidth memory is, I think everybody in the room probably knows, is critically enabling for AI compute, right? To feed the parallel, GPU or accelerator compute, you need a lot of low-latency DRAM. And that's done today.
HBM3 is a 8- DRAM stack interconnected with a process called through-silicon via, which frankly, we enable critical portions of that in the deep silicon etching. Pretty much own that, those applications, as well as the connective material is copper, deposited with an electroplating process. That's our SABRE 3D tool. So we have real strength in what's going on there, such that—i n fact, I generalize it, I put this in a bucket that I call advanced packaging. TSV is advanced packaging in my mind, even though there's not really a package there. It's a die stack, if you will. But when you broaden this out and you look at how the industry is evolving, advanced packaging has become a critical enabler, for the entire industry, right? The industry talks about chiplets and again, die disaggregation.
It's not just in DRAM. It's the CoWoS technologies and whatnot, and we, again, the TSV process, I call it the drill and fill. We do the drill, we do the fill is a real area of strength. So a couple of numbers that we put out more recently. I mean, if I was here talking to you three or four years ago, I would have described this portion of our business as hundreds of millions of dollars. And I think a quarter ago, I was suggesting to you, "You know, this is a business I can see in a couple of years being a billion-dollar business for us." And then what Tim, our CEO, said on the last earnings call is: "You know what? It's $1 billion this year." So it is a real area of strength.
We're really differentiated in the TSV stuff that we do. We sell other stuff into it, that we do well with, but the TSV stuff, that's Lam Research.
Got it. We'll definitely come back to that. So NAND WFE going forward, I think there's a fear out there that NAND CapEx could stay lower for longer, and that's sort of been the case.
It's certainly been lower for a while.
For a while. There's this perception that you're NAND- heavy, right? I don't think that's necessarily true, but there is this perception.
Yep. So let me make a comment—
Yeah, no, please.
Then keep going. You know, NAND is really quiet from an investment standpoint. Last year, year-on-year, it declined 75% in terms of investment in WFE. This year, it's only up just a little tiny amount. Yet, I'm still generating $3.8 billion in revenue. So yes, we are very strong in NAND. We enable the NAND structure in the stack, in the etch, in the metallization. I'm very proud of how strong we are there, but don't just assume we are a NAND company, because frankly, right now, there's not much going on in NAND, in the profitability, the revenue levels of the company, still pretty good.
You're doing fine. Yep. The through-cycle expectations as you think about NAND, again, cyclically, we get it, you're in a downturn, but through cycle, your views today versus three, five years ago, have they evolved, changed?
I guess what I observe is, you know, before this current cycle, I think the consensus view in the industry was that the sustainable bit demand was in the mid-thirties.
Mm-hmm.
I hear my customers now saying mid-20s. Maybe that's because we have a recency bias of a down cycle, or maybe something has changed. I'm puzzled by that, actually, 'cause when I think about it, from a demand standpoint, in my mind, nothing has really changed except to the positive side, which is AI all of a sudden has become a real demand driver. And frankly, the data generated out of these large language models need to get stored somewhere, right? The model needs to be stored. And to my way of thinking, when I look at these AI servers, there's 3x-4x the NAND content in an AI server versus an enterprise server. So I don't know, maybe there's a little bit of a recency bias or maybe something is different. I'm not sure.
If you simply get a 5% uptick in the percent of servers that are AI, and you think the baseline, demand is mid-20s, you easily get to 30. So anyway, I think we all need to just kind of wait and see how this is all gonna play out. I'm not suggesting my customers are incorrect. I would never do that. They know this market way better than I do, but I scratch my head. I'm a little bit puzzled by it, I guess, is all I would suggest. And so clearly, we're at the bottom of a downturn. We're now 2 years into it in terms of the investment in equipment. I'm beginning to see utilization tick up just a little bit. I talked about that. Now, that's not just a NAND comment. It's a broad industry comment.
We saw a little bit of uptick in consumption of spares—
Yep.
W hich for us, is the first thing that shows up in my business. And listen, I'm not gonna give you numbers about next year, but I think we're, we're set up for things to be better next year.
Makes sense. Gate-All-Around, on your recent earnings call, you called out $1 billion in revenue, give or take, for calendar 2024.
Yep.
I was hoping you could spend a couple minutes talking about that and maybe compare and contrast the GAA transition with the FinFET transition.
Yep.
How impactful is that? How are you positioned competitively?
Yes, let me talk about it. When I look at Gate-All-Around, in my mind, FinFET was a bit of a 3D structure, but Gate-All-Around is a true 3D structure. So a couple of numbers that you referenced, we'll talk about. Listen, when you look at, like, the creating those sheets, there's a incremental ALD steps, there's incremental selective etch steps, and when we look at that in our addressable market, we see opportunities. Our addressable market grows by $1 billion for every 100,000 wafer starts of capacity that the industry puts in place. That's just around the formation of the transistor. So obviously, we're excited about that. It plays to the strength. It's back to what I was talking about earlier, these 3D structures showing up, playing to the strength of Lam. Gate-All-Around is a great example of that.
And then, yes, you're right, what Tim talked about on our call a couple of weeks ago was for the Gate-All-Around node in total, the revenue this year for Lam, not just creating the gates, but the total node, $1 billion in revenue. We like to talk about billion-dollar numbers. It's just nice, big, round numbers. But we're very excited about what's going on. We're extraordinarily well positioned, and you're gonna hear us talking more and more about this as as time unfolds.
Okay. To your point about a billion-dollar opportunities here and there, advanced packaging, I think you talked about that.
Yep.
Talked about strength in HBM. We had Micron here with us this morning. They're thinking 50%, give or take, kind of CAGR in terms of HBM bits. Can you talk a little bit about that opportunity for you guys?
Yeah, listen, I'm extremely optimistic about where this is going into the future. Like I said, the baseline was $200 million, now over $1 billion, growing into next year. I'm not ready to give you a number for next year, but this is a clear way the industry is moving forward. I see clear roadmaps. AI is not going away. The excitement that everybody has about the percent of servers that are going to be AI-type servers will continue to consume more and more of the low-latency DRAM and need more and more of the drill and fill that we do, right? These 8-die stacks are going to 12 or more beyond that. So as far as I can see going forward, this is a clear growth driver for us in terms of equipment.
Got it. China comes up in a ton of conversations, and I'm sure you get—
Almost every one I have.
Yeah, exactly. So post the export restrictions in late 2022—
Mm-hmm.
Your business, your peers' businesses, declined—
Yep.
Given the restrictions.
I should pause there for a minute.
Yep, please.
When that came out, you may remember, and I'll remind you, because it's been a while since we talked about this, we lost $2+ billion of business. Just boom, gone. It was from restriction of one customer that has not come back, and it was because of restriction in certain technology nodes. That hasn't changed.
Yep.
So anyway, that was the baseline. Don't forget about that, because that plays into, okay, what's going on today?
Yes. Yes. And then, you know, from that base, you've performed really well. What's driven the reacceleration in China, and more importantly, how should we think about the forward outlook, whether it be DRAM, whether it be Foundry, some of the newer IDMs that are popping up?
Yep. Listen, I would describe to you, there is a broad set of customers in China, a broad set investing in areas where they can invest. And when I look at it, broadly speaking, it's in industrial, it's in automotive, it's in analog, it's in power, it's in microcontrollers. It's not one or the other. I believe what I observe in China, at least if I step back, is the customers I see in China are trying to supply the demand in China with supply in China, and they're making some level of progress. And so a lot of the meetings I've been having since earnings is I sense a concern that this is gonna go away. I don't see that. I do not see that. So I've got account teams that call on these customers.
The conversations we have, there's roadmaps that are years in duration. The intention is very clear. In fact, this year ended up being a little bit stronger. We had been suggesting coming into the year, "Hey, we think China sustains," and in fact, what we said on the last most recent call is, actually, it's gonna grow a little bit this year. I do see right now, when I look at China, it's a little bit first half-weighted this year, but don't misinterpret that to be like, it's going away. It's not. And I'm not ready to tell you, "Hey, what do I think 2025 looks like?" But I—i t's gonna continue to be pretty good, is my opinion.
Okay. Okay, that's great. And on the regulatory front, you know, I'm not gonna ask you to predict what's gonna happen or not gonna happen, because probably difficult. But I'm curious, you know, how has your connectivity with the government improved or evolved over the past couple of years?
Yeah.
Maybe, Doug, not so much yourself, but I'm sure, Tim—
Yep.
Y ou see a lot. I think your customers tend to have a louder megaphone, when it comes to interacting with them. But do you think the government fully understands and appreciates the criticality that you guys sort of, you know, the critical nature of your business and your position?
Yeah. No, I believe they absolutely do, right? The focus on what's going on in semiconductors, government understands what's going on. I, I guess if I would turn the clock back five years ago, we had a very modest presence in Washington. We now have an amazing government affairs team that actually spends a lot of time doing our best to educate the government about what's going on, who are we, where, where do we do what we do, how do we do what we do, what is the impact of, like, the jobs that we have, where, where do we—a nd by the way, we are a great American company, right? Most of the R&D we do is in the West Coast of the United States. Company has always been sort of a huge California footprint. Novellus had a huge Oregon footprint. That's where we are, right?
That's where we do what we do from a R&D standpoint. And so, listen, the government's smart. They spend time with us and others in the industry. We've ramped the team up, who has great capability. And again, in my humble opinion, I think we've got the best government affairs team in the equipment space. And we're doing everything we can do to educate everybody that'll listen to us about who we are, what we do, how we do it, why we're important to what is happening in the United States economy.
Okay, great. Multi-part question on localization and implications for the market and you guys. Impact on global WFE demand, near and long term. Your views probably haven't changed, but I'll ask it anyways.
Yep.
How you think about your R&D, manufacturing, sales footprint going forward? And I guess the competitive landscape, any implications from everything that's going on across various governments?
Okay. I'll probably forget some, but so redirect me. I'll just—
Right
I'll ramble on. Listen, this is a global industry. The majority of the fabs in the industry are in the Asia region. We do most of the R&D that we do in the United States, although, like I mentioned a little bit earlier, our lab footprint has grown over the last several years. We've built a big lab in Korea, Taiwan. We've got engineering capability in India, and Bangalore. So the reason we've done that is proximity to customer. We want to be close to the customer because so much of how this industry does what it does is collaborative with the customer. The customer is working on a roadmap for N+2 or N+3.
They're not quite sure how to create the features that they want to create or do what they want to do in terms of formation of the architecture of the transistor. They need to work with us to try to sort through it. So we have a belief that being close to the customer makes us quicker to solve problems, quicker to turn information around, quicker to get their wafer, run it, bring it back to them. I believe actually that's a differentiator for Lam, uniquely differentiator for Lam. So there's that. We are a global company from a manufacturing standpoint. We manufacture our equipment in California, in Oregon, in Austria, in Taiwan, in Korea, and increasingly, you'll see us ramping up a factory in Malaysia. I think that's been important because our customers are global, right?
We wanna be near them with manufacturing as much as possible also because you can get stuff to them quicker. I'm excited about, as you know, whenever business growth does resume, the factory we built in Malaysia will enable us to be very efficient, relative to maybe what we've done in the past. And I say that because first, obviously, in Malaysia, it's closer to where the Asia-based fabs are, and so we can get equipment there quicker. We fly things in and out of the factory, and to get things to our customers, oftentimes we're putting it in the belly of an aircraft. Being closer means we get it there quicker, and frankly, we spend less money on freight and logistics.
I believe also, our footprint in Malaysia is gonna differentiate us over time, there as well, again, because proximity to customer is such a critically important item in manufacturing, not just in R&D. You asked me something?
WFE and impact from localization.
Yeah, no, I get that question a lot. Listen, to your earlier question, I think the governments understand what's going on in this industry. Absolutely, they do. I think every region of the world has woken up to the criticality of the semiconductor industry, and equipment is a key portion of that. You know, you got the U.S. CHIPS Act money. There's money in Japan. The Chinese have money. The Europeans have money, right? Everybody is trying to get more, like, customers, perhaps, than the equipment, but they want fabs in their region.
And so that's got an impact, certainly on where WFE shows up, in what region of the world, where the fabs are, where we're shipping equipment to. I sometimes have people think, Okay, the U.S. CHIPS Act, they're gonna deploy $38 billion into building manufacturing capability. That's upside to WFE. Don't run there quite that fast, because it's important, and, I find this to be universally true when I get confused about, Okay, the cycle has changed timing of something. You have to step back and ask yourself about, where is the demand for semiconductor output? And is any of this government money changing the end demand for semiconductors? To my simplistic way of thinking, the answer to that question is no. And so does the $38 billion CHIPS Act create $38 billion in upside to WFE? It doesn't.
Now, will it create some upside? Of course, it will, right? Because if you're building fabs in more locations than fewer, there will be incremental equipment needed, because you need some level of redundancy in each location. But don't run to the conclusion that, oh, $38 billion of incremental government money is $38 billion in incremental WFE. That's not the case.
Got it. That's super clear. On to competition, one question on these incumbents and the other one, China. One of your nearest competitors, they've been pretty vocal about cryo etch and how that could drive share gains for them in the 4XX, 5XX layer, 3D NAND era. I think Tim did a great job on the recent call, sort of addressing that view, but if you can sort of, yeah, walk us through what you're seeing there, that'd be helpful.
Yeah, listen, this particular competitor, if you will, had been suggesting a couple of years ago that they were gonna win business from us this year. Didn't happen. That's what Tim told you on our most recent earnings call. And now they're talking about, "Eh, a couple of years from now, we're gonna win business." Yeah. Okay. Talk is talk, be that as it may. Listen, what I would tell you is there's only one company that's got cryo etch in production in the industry, is Lam Research. We've got 1,000 cryo chambers in the installed base.
You get good at doing things in this industry by doing it, and so that's a statement. We've got 6,500 etch chambers in the installed base in NAND. The incumbency is really important. The best solution for the customer is always: Can you upgrade what's already there? And so listen, are they gonna keep trying to win business? Of course, they are. When you have a position as large as we have, everybody's looking at it, trying to figure out how to get a portion of it. But all I can tell you is we defended everything this year.
Okay. That's very clear. In China, I know you've been competing with these guys for years. The Chinese equipment companies. The Chinese, yeah, the Chinese equipment companies, for years, if not decades. You know, some of the public companies, obviously, we, we have access to financials, and they are growing quite fast. They are off a low base. Naura, the AMEC of the world, do you increasingly see them inside China? Most likely not outside of China, but inside of China, do you see them more often and, you know, their capabilities today and how they compare with you guys versus five years ago, 10 years ago?
No, it's a great question. Yeah, they're growing off a small base in areas where I call them sort of non-critical applications, generally speaking. Listen, in my experience, in, again, my 30 years in the industry, almost always, best technology wins. Almost always.
And you know, when you're a company like Lam, who's been around for 44 years, leader in etch for most of that time, Novellus was around almost as long. We're really good at doing what we do. The incumbency factor makes it really difficult to catch up with us if you're trying to take something we already have. I think, if I'm being honest, though, there are certain customers and certain technology nodes we can't sell to in China any longer. And so if you're one of those customers that can't get our equipment any longer, what choice do you have but to buy what you can buy and do your best to make it work as well as you can make it work? I think that has a lot to do with what you're seeing relative to growth rates of the China equipment companies.
Technically, we're getting better every year. You got to be better this year than you were last year, and better next year than you were the year before. We're always gonna do that. We're always gonna move the goalposts forward. You know, you got to go where the puck is headed, and we're the ones driving the puck up the ice, and they're gonna keep chasing and skating hard. And are they gonna keep growing? Sure, they will. Yeah. And in China, with customers in China, there's a huge incentive to buy Chinese equipment if they can make it work, and yet my growth in China has been actually quite, quite strong. So, anyway, that's how I see that.
Yeah. No, I appreciate that. I'm gonna pause here and see if we have any questions from the audience. There's one in the back.
So, you mentioned that Chinese have money, and they want fabs in their region. Do you see any benefit from their $48 billion semiconductors fund? And also, second, do you anticipate any impact from the Section 301 tariff? Thank you.
I guess on the tariff, nothing I know of right now. Things could change, but nothing I'm aware of right now. And I kind of think of the new fund money in China, similar to what I see across governments all over the place, right? It's an intention to continue to have the industry within the country that they're in. It's a continuation of what's already been there, to be honest.
Yep, I got one right here.
Does having fabs just effectively redistributed around the world, because that's the way it was many, many years ago—
Yes.
Does that make the industry more cyclical or not? Does it have any impact?
From a cyclicality standpoint, I don't think so. It makes it a little bit less efficient. Honestly, I think, you know, that was my comment about, "Hey, is this money incremental to WFE?" A little bit, because you're gonna need incremental redundancy tools and whatnot. I don't really, in my mind, think it adds or detracts to cyclicality. These are still very large facilities. It's still largely like one customer. They're just building a fab in two different locations or three different locations. That's not different than the industry's ever been, so I could be wrong about that, but to the way I think about it, I don't know that it adds anything to cyclicality. Good question, though.
I'll keep going. There is no fireside chat with Doug Bettinger without CSBG.
I love CSBG. Thanks for squeezing it in.
So great business. However, the Reliant business has weighed on growth.
It has, yep.
Hopefully, that's, you know, approaching a bottom. You've been growing this business, you know, comfortably double digits through cycle. I know there are four parts to the business, but if you could walk us through how you're thinking about the individual parts and longer term, the growth rate there.
Yeah, let me unpack it. So, the acronym CSBG, Customer Support Business Group at, at Lam, and what it is, is the business that supports the installed base for the most part, except Reliant. I'll get to that in a minute. There's four components in CSBG for, for us. There's spare parts, service, equipment upgrades, and then what we call the Reliant product line, which by and large, is what shows up in the mature node spending profile.
This is a great part of the business, a great part of the business model, frankly, and I'll remind people, or if you're new to the story, if you look at an average piece of equipment that we sell, people are sometimes surprised when I tell them the revenue we generate after we sell the tool on average is at least as much as when we sell the tool itself. Our equipment runs for decades. We sell it to a customer, it then has an upgrade path, it consumes spare parts, it needs service. At some point, perhaps it gets redeployed from one fab to another. That shows up in the service line. It used to be the case that sometimes equipment would come back, we would repurchase it, refurbish it. That's not available anymore.
I mean, nobody is letting their equipment go any longer, but that is the Reliant product line. If I unpack this for you a little bit, just to understand it, again, revenue runs for a very long time. It's a very profitable part of the business, and I'll describe it. It's really the majority of the R&D in terms of designing a tool is done when the tool is developed, and then as you sell spares and service, it's not that there isn't any R&D, there is R&D, but it's a much more modest amount of spending. And so that results in pretty attractive operating margins in this business and actually quite nice cash flow, free cash flow generation. That's important, thinking about the return of capital and the dividend level of dividends.
CSBG comfortably supports the dividend, and comfortably supports our ability to grow the dividend over time. I love this part of the business model. We give a chamber count number at the end of every calendar year. It grows every single year. I used to say CSBG is a business that should grow every year. Well, this year it's not, because what you referenced there, the mature node spending relative to the Reliant business unit is a little bit soft this year, and utilization is a little bit soft. So those two things create a little bit of a headwind. I've suggested CSBG this year likely is flat-ish, is the word I've used. But still a business that over several years in duration will again be growing every year, in my opinion.
The biggest component of the business today is spares. If I was talking to you a year ago, I would've said that there's two big components, spares and Reliant. Reliant is a little bit soft this year, and perhaps largely why it's flat-ish this year, is there isn't huge strength in the Reliant product line this year. But it's still a great product line. I still, like I said, I love this part of the business model. The things we're excited about, relative to service, I would just mention briefly, is, you know, service in this industry historically has been, show up with your engineers and perform a task, do a maintenance. It's almost like if you think about your automobile, bring it in, change the oil, give it back.
The customer can change their own oil oftentimes, so that's never been a great part of the business, the pure service. But increasingly what we're doing, if you wanna get excited about AI and data-enabled activity, is we're using the data coming off the tool to deliver different kinds of service, predictive service, where we can improve performance metrics of the tool that deliver real value for the customer. As I look at that and think about where we're trying to drive this part of the business, that is really good business, and increasingly, that's becoming a bigger and bigger percentage of the service component of CSBG.
Got it. And the chamber count out in the field, being the KPI makes a ton of sense. I think at the Investor Day back in 2020—
Yep.
You guys talked about the revenue per chamber opportunity growing.
Growing faster than the chamber count.
Right, and that obviously is still—
But now Reliant has skewed that a little bit.
Right. Right.
So frankly, I need to construct a new metric and, and help people understand it. That is still what the company is doing. If you kind of pull the Reliant piece out, that grew a ton and then is a little bit soft this year, average dollar per the installed base is growing, because of this advanced service stuff.
Yep.
Increasingly.
Yep. Okay, makes sense. Gross margin, one of the metrics we're all fixated with, near-term and long-term. By design, I think you guys have built this model where you have little variability, which is great. As we think about the forward, what are some of the opportunities that could potentially drive margins higher for you guys?
I guess I'd have you think about a couple of things. First is the new products we're bringing to market. We believe there's increasing differentiation. So generally speaking, when you have that, profitability is better. So that's one. Two, is something I referenced earlier, as we were talking. There's an efficiency in terms of having manufacturing closer to the customer with the Malaysia factory, the Taiwan factory, the Korea factory, that over time, whenever growth does kind of show back up, the efficiency of the company should be better as we go forward.
Understood. Capital allocation, I think you guys had a big announcement last week. You had a stock split as well.
We did.
You're the trendsetter. One large GPU company followed suit.
It was the next day.
After you guys. But, I think you've been super clear with your capital allocation priorities, but if you can kinda—
Yeah
Remind us.
Let me unpack it for you a little bit. Yeah, last week, we announced two things. We were coming to the end of our previous buyback authorization, so the board reauthorized a billion-dollar buyback program, with an indeterminate timeframe. So, I think that's a statement on our commitment to returning free cash flow. It doesn't change any of my outlook relative to what we're gonna do or what we're planning to do, which has been for the last few years, to return 75%-100% of free cash flow. I think over the last, again, my decade at the company, we returned more than 100% of free cash flow. There are a variety of reasons that, looking at the clock, we don't have time to unpack.
But, yeah, I mean, our first priority for cash at the company is invest in the business, right? Invest in R&D, invest in labs, invest in factory, invest in what the business needs. But even after that, we generate a lot of free cash flow. In fact, it was a very strong free cash flow year last year. And so we do that through a combination of the buyback, which has been, historically anyway, where a lot of the capital return has been. But we put our first dividend in place in 2014, and we have pretty much grown it every year since then. I know that's what most investors wanna see, a commitment to continually growing the dividend, and that's what we plan to do.
And again, that CSBG, CSBG cash generation comfortably supports the dividend and growth in the dividend over time. And so, yeah, that's again, what we have been doing, what we plan to continue to do. It's a key part of the value creation of the company, right? Share count is coming down, because we're using the cash to make that happen. It has been, during my time at the company, a great use of cash. You know, situationally, you look back and say, "Look how smart we were." But look how smart we were. The average buyback price since I've been at the company, I believe, is below $200. Stock's, I don't know, yesterday it was $950. I don't know where it is today. I haven't looked. But it's been a key aspect of the value creation of the company.
And then just to comment on the stock split, I didn't do it for you investors. I know you guys don't care about share price per unit, but my employees do. And what I was hearing increasingly internally at the company is Doug, Tim, you know, the shares are too expensive for us to participate in the level we want to participate in, especially from some of the global employees, as we're ramping in regions of the world where, you know, compensation isn't as high as it is in the United States. And so over time, as I heard this more and more and more and looked at the participation rates of the share purchase plan, I realized we needed to do something.
And to my way of thinking, to Tim's way of thinking, the best thing we could do to help employees participate more meaningfully was to split the stock. The employees are excited about this, by the way. I will tell you that. And, you know, I interact with employees. Many people have already thanked me for doing this. So again, I know you guys look at this and say, "Doug, you know, this, you didn't need to do this." I know, but the employees needed it. We have a strong belief at the company that employee share ownership is a key portion of, like, keeping the employees motivated, keeping them focused on trying to increase the value of the company, and I want that from the top of the company to people at the most junior levels in the company.
So that was a large part of the motivation for doing what we did there.
Makes sense. Last 90 seconds, anything that we didn't touch that we should have, or any key messages from you guys?
No, I think we covered it pretty well. I mean, it was a, it was a pretty comprehensive set of things. Listen, you know, business is decent right now. The memory cycle hasn't, like, hit its recovery point, but I see early indications that I call them green shoots, right? Pricing is getting better, profitability is getting somewhat better. In DRAM, you have a clear product cycle happening with DDR5 and High Bandwidth Memory is a real thing. AI is a real thing, the magnitude of which I think we all need to wait and see how this plays out. NAND is still pretty soft, but I feel like we're looking into a year next year that should be better there.
Makes sense. Awesome. Thank you so much. Really appreciate the time.
Awesome. Thanks for having me, and thank you everyone, for for showing up today. Appreciate it.