Great. Good morning, and again, welcome to the second day of JP Morgan's 52nd Annual Technology, Media, and Communications Conference. My name is Harlan Sur. I'm the semiconductor and semiconductor capital equipment analyst for the firm. Very pleased to have Doug Bettinger, Chief Financial Officer of Lam Research, here with us today. Lam is the third-largest semiconductor capital equipment company in the world. Strong leadership in etch, deposition, patterning, with the fast-growing and emerging businesses like next-generation transistor, memory cell architectures, resist processing, and advanced packaging. I'm gonna hand it over to Doug. He's gonna mention Safe Harbor, and he's got some opening comments. Doug, thank you for joining us this morning.
Harlan, thanks for having me. Fifty-second year, I didn't realize this-
Mm-hmm
... been going on that long. It's, it's amazing. That's actually a little bit longer than Lam's been in business. Although, we happily recently celebrated our 40-year anniversary of having been listed on Nasdaq. That was kind of cool to see us on the tower. But anyway, my attorneys like me to read, some Safe Harbor language, so let me do that to keep them happy. And here it goes: Today's discussion may include forward-looking statements that are subject to risks and uncertainties. Actual results may differ materially. Additional information concerning factors that could cause our actual results to differ materially from those forward-looking statements can be found in the risk factor disclosed in our public filings with the SEC, including our most recent 10-K and 10-Q.
So I don't actually plan to announce anything new, so I probably didn't need to read that, but regardless. Anyway, thanks for being here. It's always good to be at this conference, extraordinarily well-attended every year, and judging from the fullness of the room, it is again this year. So real quick, I wanted to make sure everybody saw a press release that we issued this morning, that's incremental new information. So the company was pleased this morning to announce a new buyback authorization from the board of $10 billion, that we plan to execute over an indeterminate timeframe. This announcement is consistent with our ongoing plan to return 75%-100% of free cash flow. So I wanted to make sure everybody saw that.
Then second, we also announced this morning a 10-for-1 forward stock split. Now, some people have looked at me and said: "Why did you do this? This doesn't matter to institutional investors." And obviously, I understand that well. But what it does matter to is the ability of our employees globally to participate in the company's stock ownership plans, including our share purchase plan. And so as the share price had begun to move higher and higher, the ability of some of our more junior employees, as well as our Asia-based employee population, they were challenged to be able to participate. And we believe at Lam, stock ownership is an important part of what we like the employee base to be able to avail themselves to, and that was the motivation for the stock split, nothing more than that. So I want to just put that out there. We can then, Harlan, jump into the business, but I wanted to make sure everybody saw that. So anyway, that's those are my opening remarks, Harlan, and why don't you guide us through the Q&A that I know you've put together?
Yeah, no, it was great to see the re-upping of the share buyback authorization. Just as a reminder, on calendar 2023, the Lam team returned 80% of their free cash flow, $3.8 billion. Start to this calendar year, I think, I think in the March quarter, you guys bought back $1.1 billion. Bought back plus dividend-
That's right
... $1.1 billion. So stepping up from the run rate of calendar 2023, which may be as good timing as we do see, industry trends starting to improve. But, great to see the team upping the buyback authorization.
Yeah, listen, as the CFO of the company, the capital allocation, I believe, is a critical component of the value creation model at the company and a critical way that we have delivered and will continue to deliver shareholder value. So, anyway, let's jump into the business, Harlan.
Yeah, so, you know, after three strong years of semiconductor industry growth, we did enter into a cyclical downturn back half of 2022. In 2023, semi industry revenues were down about 8%. WFE ex-litho was down about 8% last year.
Yeah.
That's according to Gartner. The downturn, particularly in leading-edge memory and foundry logic, was steep, right? NAND and DRAM, peak-to-trough industry revenue declines were about 60%. PC server CPUs were down about 40%-45%, peak to trough. Take us through how all of this impacted WFE spend and Lam last year, and so far this year.
Yeah, let me talk a little bit about 2023 and where I see things in 2024. I guess looking back, last year was a challenging year for the industry, especially the memory component of the industry, and our enablement of our memory customers showed up in our business in a pretty meaningful way. The secular downturn in NAND specifically, I've never seen as dramatic as it was last year. NAND investment was nominally down 75-odd% year-over-year in 2023. DRAM was a little bit better, but also fairly soft, at least with the global customer base. Little bit of spending in China DRAM helped that. So anyway, stepping back, it was a pretty challenging year.
We went through a period of restructuring the company, trying to get the company set up to be better when growth resumes, and I think we've done, or last year, we did all the right things to benefit from whenever that growth does show back up. Now, roll that into you kinda led into, all right, what, what's going on in 2024? 2024, WFE is a little bit stronger, although in memory, it's still fairly muted, especially in NAND. It is a little bit more investment, but only a very little bit. I think, though, Harlan, and you mentioned this to me as we were sitting down here, some of our memory customers have been here at the conference talking about what they're seeing in their business.
I guess when I look at what we see going on there and what we see relative to pricing, you're starting to see things getting a little bit better, right? Pricing's getting a little bit better, profitability is getting a little bit better. I've been describing it as maybe some green shoots showing up, but really not much in terms of real investment occurring, maybe with one exception around high bandwidth memory, and I know you're gonna ask me about that later, so I won't go deeply into it. But yeah, we see WFE this year in low- to mid-90s, up incrementally from last year, although still a fairly modest amount of investment on the memory side of things.
Perfect. Yeah, you know, we had Western Digital present yesterday. They presented a profile of continued price improvements and supply tightness in NAND pretty much through the second half of this calendar year, extending into calendar 2025. We had the Micron team here today saying exactly the same thing about DRAM and NAND. And so from my perspective, what the memory players are seeing today and the strong sort of focus on profitability and supply discipline is being reflected in your comments on WFE in calendar 2024.
That's right.
But the demand trends and the supply dynamics heading into next year seems to indicate, you know, the prospects for a stronger WFE spending year, and Micron did take up their CapEx today to the upper end of their range. So maybe we're at the start of something here. But, you know, as you step into the second half of this year, it does feel like advanced node momentum is starting to accelerate. I talked about Micron and WD.
Sure.
But in foundry and logic, we've got some big technology transitions happening now, right? 3-nanometer wafer starts are picking up pretty significantly because of all, of all the AI and networking SoCs targeted for accelerated compute and AI. You've got, you know, Intel and TSMC putting in the groundwork for their next generation 2-nanometer technologies, and design starts at 2-nanometer appear quite strong, right? So given your broad share position in the industry, your visibility, you know, customer discussions, you know, what is, what does the outlook look like for calendar 2025 WFE, maybe from a qualitative perspective?
Yeah, it's too soon for me to give you any numbers on things, but I agree with everything that you described. And maybe just unpacking a little bit, yeah, I think that you know, talking to investors as I have been today and over the since earnings, I think the view is '25 is going to be a decent year, and I agree with that. You're absolutely right in that you know, the entire world is excited about artificial intelligence, and I am as well. And the stuff you described, Harlan, around the Gate-All-Around node showing up, we're all talking about that, around backside power showing up, around advanced packaging, around high bandwidth memory, all leads into the enablement of AI, which is a real thing.
And so, when we look at that, and maybe I'll just mention a few things, as you begin to look at leading-edge foundry and logic, you're beginning to see real 3D structures, and Gate-All-Around is a great example of that. When we look at that node this year, we've described an expectation that revenue for that first Gate-All-Around node for us this year will be a billion dollars in revenue, and obviously growing into next year.
We're excited about this because when you look at the architecture there, because it is a 3D architecture, the intensity of etch and deposition in terms of creating those sheets, the gates, there's a decent amount of selective etching required there, a decent amount of our ALD tools enabling that, such that when we look at what this looks like to us, we've put numbers around it saying that for every 100,000 wafer starts of capacity the industry puts in place, the incremental opportunity for us is a billion dollars. Now, we won't get 100% of that market share, but that, just to have you understand a little bit about it, describes why business is gonna be incrementally better next year. Capital intensity is going up, and it's etch and deposition intensive.
So maybe I'll pause there and let you redirect me, Harlan, but yeah, we're excited about all, all the transitions that we see right now in the industry.
Yeah, and we'll get into more details on some of those dynamics. But if you look back historically, which is quite interesting, you know, wafer equipment spending intensity bottomed in that sort of 2013 timeframe, right at about 9%. That was right at the cusp of transition from 2D to 3D NAND, and, you know, transi- and conventional transistor architecture to FinFET architecture, right?
Yes.
And 9% spending intensity bottom, 2013, 9%. Last year, a down year, it increased to 16%-17% intensity, right? So another way to think about that is that over the past 10 years, including last year, wafer equipment spending has grown at a 12% CAGR versus semi industry revenue growth at a 6% CAGR. Over that period of time, Lam has grown its total revenues at a 14% CAGR, right? So clearly, taking share. Given the complexity challenges, technology inflections, the need for more installed capacity to support some of these strong demand trends, do you anticipate WFE and Lam continuing to outgrow semiconductor industry revenues? And maybe what, within that, like, what are going to be the biggest drivers?
Yeah, Harlan, you know, when I abstract this and look at what I see happening, your observation is right. WFE is growing at a faster rate than overall semi revenue, that's because of capital intensity growing. Now, what I always point out to people though, and this is important, is, yeah, as a percent of revenue, capital intensity's gone up. Capital intensity per wafer put in place has gone up. But if you look at capital intensity as a percent of the industry's-
Mm-hmm
... operating profit, it actually has not gone up. So that's important. This is a metric I look at all the time through cycle, 'cause you got to normalize for, okay, where are you at in the memory cycle? You have to ask yourself, "Is this affordable?" And if you just look at revenue, you might say, "Hmm, maybe not," but industry profitability has never been better, and so that, that's an important metric I always look at, and I just thought I'd mention that. And then again, abstracting, when you look at what's going on, you know, the 50-plus year history of the semiconductor industry was largely all about just optical shrinking. Things got smaller and closer together, and that stopped happening, I don't know, 10 years or more ago. But Moore's Law didn't slow down in terms of enabling incremental performance.
It's just happened in different ways, architecturally, and in the manufacturing processes. The benefit relative to the share gain trajectory relative to share of wallet for us has been 3D architecture showing up. It's obvious in NAND flash that went from a planar to a 3D structure. We benefited significantly from that, in that our addressable market doubled, right? It's etch and deposition intensive. You have similar things if you kind of step back and think about, okay, in foundry and logic, we went from a planar gate to FinFET, now going to gate all around. That's an evolution of 3D structures, more etch and deposition intensive. We've also benefited in there from patterning being etch and deposition intensive.
One thing, I'm really excited about right now, and it shows up in high bandwidth memory, but it's the 3D architectures-
Mm-hmm
... around advanced packaging. Reduced to be CoWoS and other things, high bandwidth memory. Again, you're starting to see some of these very large accelerator die, especially GPUs, approaching the reticle limits of how large the die can be. And so to continue to drive transistor performance, advanced packaging has been a critically important 3D structure that shows up, again, benefiting us because of the intensity of etch deposition, specifically in something called deep silicon etching and electroplating. We largely enable a lot of that architecture. And so when I try to generalize it, I just see all of these things continue to evolve in terms of materials, in terms of architecture.
We're only just beginning to see the industry working on a DRAM structure that will eventually be more 3D-like, for similar reasons that you saw it in these other places, to continue to drive performance. And so, I don't know, in business, sometimes you got to be a little bit lucky, and we've been a little bit lucky at Lam in that we're an etch and deposition company. 3D structures need more and more etch and deposition equipment. You know, success in business, you got to be a little bit lucky, but you also got to be good, and you got to be able to execute, and you have to anticipate these things and step into it, and I think we've done a nice job with that as well, and we intend to continue to do that as we look into the future.
We're in the right place at the right time in a great industry, is how I think about it.
Yeah, you know, and then from that perspective, you know, given the unprecedented memory downturn that we are emerging from, some strong share gains in advanced foundry and logic by the Lam team. To your point, whether it's foundry or logic or memory, I mean, everything is moving to a 3D-like architecture, right?
Yeah.
Interestingly enough, the mix of your business has evolved over the past few years, right? From 20%-30% foundry logic mix, to now 45%-50%+ foundry logic mix. Much more balanced end market exposure.
Yep.
Given the share gains in logic, the strong exposure to some of the 3D-like architectures, like gate-all-around, backside power distribution, does the team believe that it can sustain a more balanced mix, even if memory WFE comes back strong starting next year?
Listen, you got to unpack this a little bit. We've made very nice progress in foundry and logic, right? We've talked about share doubling nominally node over node at a large logic customer.
Mm-hmm.
That is happening. Some of the benefits of evolution of the manufacturing process around things like patterning, and we're extremely strong in the patterning module, and that's been a key enabler for device evolution in foundry and logic. We've benefited from that. And so when you look and unpack at an application level, Lam has made very nice progress in terms of gaining larger footprint in foundry and logic. That doesn't obviate the fact that we've got very strong memory footprint, and that's not changing. But I feel really good about the investments we're making in terms of gaining in foundry and logic. We've already talked about gate all around. Backside power is another example. The metallization scheme of moving power to the backside significantly benefits our plating business again.
Our Sabre tool is gonna drive opportunities there. We're excited about that. We're investing for that. It's been a part of the reason why R&D investment is growing this year. When I look at the backside power, again, we like throwing big numbers around, but these are big numbers. Again, similar to gate all around, backside power, for every 100,000 starts again, we see an incremental opportunity of a billion dollar again. We like billion-dollar numbers. So we're positioned well for that. I'm excited about where we're going there. I'm optimistic that next year is gonna be a decently stronger investment year in memory. So relative to mix percentages, Harlan, there's a numerator and denominator moving around there.
I don't know which is gonna do more or the other, but frankly, the momentum the company has in every component of the industry is actually good. That, to me, is the most important thing.
You know, one of the, one of the large demand drivers that wasn't around or pervasive even 18-24 months ago was advanced packaging, right? Whether it's high bandwidth memory for DRAM or 2.5D, 3D packaging for some of these very advanced logic, AI-based SoCs. And so all of a sudden, we're seeing a lot of adoption of these advanced packaging architectures and a lot of spending around it, right? And again, perfect example, HBM is just a 3D architecture. It's like a perfect fit for Lam's core competencies, right? And from that perspective on advanced packaging, the team said on the last earnings call that you're gonna drive a billion dollars in advanced packaging revenues this year, which includes HBM.
Right.
I believe this is about 2 - 2.5x growth versus last year. I believe you have 90% share in HBM through-silicon via etch and copper fill for those TSVs. What other areas does Lam have a strong share position in, either HBM or 2.5D packaging?
Yeah, thanks for... All your data is correct, Harlan. Yeah, we enable a lot of what's going on in advanced packaging. So maybe let me take you back. You know, if I was here talking to you two years ago, I would have been describing this as a business that was hundreds of millions of hours in terms of the size. Fast-forward to late last year, I began to suggest, and Tim began to suggest, "Hey, this is a business we can see approaching a billion dollars in a few years." On the last call, Tim actually said: "You know what? We see this as a billion-dollar business this year." Unpacking that, actually, a large piece of what's growing is high bandwidth memory. We see that tripling year-over-year, so we're super excited about that.
We sell lots of different tools into that, but the positions of real strength you mentioned, Harlan, I tongue in cheek call it the drill and fill, right?
Mm.
For the through-silicon via, you need to etch spaces, and then you need to deposit the conductive material. We pretty much own those applications everywhere. And looking at that, like I said, it's a real tailwind. And listen, as you think forward over the next several years, I see this continuing, the strength in this continuing. Everybody, like I said, you're starting to see some of these large die approaching radical limits. You need to keep driving performance forward. Everybody's talking about die disaggregation, and so forth. That all means advanced packaging. So we're excited about this. You know, I think, you know, the bright spot I see in memory right now is high bandwidth memory.
You need more and more of this to feed the parallel compute engine, and we're right in the middle of all of this stuff. So anyway, it's a key aspect of the growth of the company. There's some things we just do extraordinarily well in here, again, in the drill and fill. And this isn't gonna slow down for the next several years, as far as I can tell.
Yeah, and the one thing I'll mention, because we had Micron here earlier today, you know, their view of the DRAM industry is, I think DRAM bit demand growth in the sort of low double digits type of range on a go-forward basis. They told us today that HBM bit demand growth over the next few years is looking like a 50% CAGR. So advanced packaging, I think, is gonna be a strong tailwind for the infrastructure and the support.
Yeah, I agree.
You know-
You know, also, when you look at it, Harlan, and you know this, but I'll just remind the audience, this is a much bigger die, right?
Mm.
Nominally, twice as large. And so it requires more fab capacity just to manufacture the die. The fact that it's a larger die also means it's gonna have a lower yield. For the same level of defectivity, a die that's twice as big has a lower die yield, so that requires more equipment. And then, like I said, the drill and fill and the other stuff layers on top of that. So yeah, I'm very excited about this.
Do we have any questions from the audience, before I move on? If you do have any questions, raise your hands, and we'll get a mic over to you as quickly as possible. You know, you already talked about some of the transitions and opportunities that you're taking advantage of on some of the advanced foundry and logic, the move to next generation, 2-nanometer nodes. You did a great job of quantifying the team's opportunity. I think there are 3 big inflections. Number one is, gate all around, right?
Yep.
a billion dollar per $100,000 wafer starts, backside power, a billion dollar opportunity per $100,000 wafer starts.
Yes.
But the third inflection is continued expansion of EUV lithography critical layers, obviously, with the eventual move to high-NA EUV. But I mean, here, the team has come up with some pretty innovative capabilities around, you know, EUV, dry resist processing. Help us understand. We haven't had an update from the team-
Yep
... in quite some time. So give us an update on that, the opportunity, and maybe the timing of adoption.
Yeah. So we've been talking about a disruptive technology we call dry photoresist coming in in partnership with ASML to help enable EUV. I guess I'd call it throughput manufacturing. Yeah, this is a segment of the business we're not exposed to today, right?
Mm-hmm.
This is the track that we're attempting to disrupt technology that's been around for decades, frankly. And what I would describe to you is we've announced a couple, PTOR positions. I would describe to you every one of our customer that uses EUV is evaluating our hardware in the lab. Technically, they see the benefit of it. It's more efficient in terms of absorbing the photons because it's deposited with a dry process as opposed to a wet chemistry process. I describe this as a business that cumulatively, over a 5-year timeframe, is a $1.5 billion opportunity growing through the back half of those years. And eventually, this, I believe, becomes another billion-dollar business, but still a little bit further away. We're very excited about it, like I said. Customer evaluation seems quite good.
It's gonna be adopted, we believe foundry and logic, as well as DRAM. The momentum is beginning to show up in a more meaningful fashion, Harlan.
Now that we're seeing a strong recovery in NAND supply and demand trends, we are seeing your flash customers starting to drive technology inflections, right? Technology migrations, primarily the move to 200+ layers.
Yep.
Very good for Lam, as you have a strong leadership position in etch and dep, NAND layer formation. But the move to 200+ layers also introduces a lot of other complexities, right? You've got bonded array technology, you've got more aggressive lateral scaling. Like, how are all these incremental technology components benefiting the team?
Yeah. No, Harlan, it's a good observation. Frankly, we're right in the middle of all of this stuff. To your point, that alternating film stack, we deposit that. The high aspect ratio etching, we etch those holes. The tungsten fill, we deposit the tungsten pretty much across all of the industry. So looking at the evolution of the stuff and eventually, you're gonna start seeing, the customer base stacking, these layers, and we are right in the middle of understanding all the stresses there. We've brought a tool to market we call Vector DT that actually helps manage some of the backside stress. And so, simply because we're right in the middle of all of where the NAND customer base is, because we enable all of those critical applications, we see things coming before others.
We anticipate things that need to come out, and I think that Vector DT backside stress tool is a great example of that. We saw the challenge the industry was gonna have, and we anticipated it and brought this capability to market.
We, the 2023 WFE market share rankings were just published. Obviously, Lam continued to maintain a very strong number one position in etch, number two, strong number two position in deposition. What was interesting, though, is that the two large China domestic equipment companies saw their businesses grow almost 50% in 2022, and, I don't know, 65%-70% last year, when total wafer equipment spending ex litho declined by about 8%, right? In dry etch, one of your competitors moved into the number four position. Your other China competitor moved into the number five position. In deposition, one of your China competitors moved into the number five position.
Mm-hmm.
So, how does Lam view this threat from the China domestic guys?
Yeah, it's something we obviously pay close attention to, Harlan, and I would suggest to you a lot of what the numbers you're seeing are largely a function of the fact that U.S. equipment companies are restricted from shipping to certain customers as well as certain technology nodes, and the customers there need to buy whatever they can buy. That actually is a big part of what's going on. But yeah, the growth there, and obviously, at the end of the day, the investment in the China semi industry, they buy China if it's available, even if it's maybe not as good as from the leaders in the industry. So that's what you're seeing show up there, Harlan. But listen, at the end of the day, in my experience in this industry, the best technology wins.
And the best technology needs a certain level of investment. It needs continued evolution. The barriers to entry in terms of competing for leading-edge capability are quite high, and we intend to make this a moving target. We will continue to be really good at doing what we do. You know, we're spending substantial amounts in R&D to keep moving the technology forward. So that's how you compete, right? Best technology wins almost always-
Yeah
... in this industry, and that's how we're gonna continue to do what we do.
Any questions from the audience? Let's talk about services. 33%-35% of your overall sales, nice annuity-like business, typically grows even during downturns.
Typically.
Last year and this year, you know, a bit different because of the unprecedented downturn across your customer base, discipline on taking down utilizations. On top of that, part of your services business was impacted by sort of last year's equipment restrictions. Despite all of this, though, your services business, CSBG, has grown at a 14% CAGR over the past 5 years, above your target of 10%-12%. Team is continuing to expand its installed base and value-added services on that installed base. So putting it all together... How should we think about the growth of CSBG over the coming few years?
Yeah, thanks for asking about CSBG, Harlan. A lot of the meetings that I end up doing with investors, people never ask me about it. It's a great part of the business model, and maybe let me unpack it a little bit for you in the last few minutes here. The Customer Support Business Group at Lam has four components to it: spares, service, upgrades, and then what we call the Reliant product line. And when you look at it, the benefit of this business is that chamber count grows every single year. It grew last year in a fairly soft WFE environment, 90,000 chambers. People are often surprised when I mention, "Hey, our equipment runs for decades." It just gets redeployed.
It can move from one application to another, it can be upgraded, and so forth, such that an average tool that we sell over the life of the tool generates more revenue than when the tool is initially sold. That's not every tool, but on average. So that's important to understand. It's a key part of the business model. Today, I would tell you the largest component of CSBG is the spares business. It's just part of what you need to do. You need to replace parts in the equipment to keep it productive, to keep yield high, to keep utilization high. So that's a key part of things, and as we begin to see utilizations improve in the memory portion of the customer base, that's where it's gonna show up first, spares.
And I mentioned that on our call a couple of weeks ago, that we did see a little bit of an uptick in the spares business. Now, if you'd been listening to me talk over the last year, I would've been describing, "Hey, there's two big components of CSBG." I started mentioning the Reliant product line. Think about Reliant enabling the specialty nodes, the more mature nodes in the industry, and investment in 2023 was actually very, very strong there. I think it's pretty well understood that investment in some of the trailing edge or some of the mature node capacity this year is a little bit soft. And so Reliant is down to a certain extent this year. And then the business is rounded out by the upgrades and service component in and of itself.
That's CSBG. It's a great business. It's a key part of how we do what we do. The great part about CSBG is it doesn't require the R&D intensity of developing a brand-new etch tool, as an example. There is R&D there, but the intensity of it is less, such that when you look at the operating margin profile, it's pretty good in CSBG. So it's a great part of the business model. Don't forget about them.
Last time you updated us on your financial target model was a few years ago, but I think it's still very relevant, 33%-34% operating margins. You've been successful in driving to your revenue targets that you set forth, but not quite able to get to sustain sort of 33.5% operating margins. Help us bridge the gap from current business profile to achieving the op margin targets, and over what period of time?
Yeah, Harlan, thanks for asking. The numbers you just mentioned are still the objective of the company, and we have touched 33%, just not on a consistent basis-
Right.
... which I, I think is kinda what you said. Yeah, I mean, part of it, it has a little bit to do to revenue growth. Part of it has to do with the fact that we came through some inflationary periods, as we were going through the COVID environment, such that gross margin was a little bit muted. That was a, an aspect, of how we got there, was expansion in gross margin. Now, we're managing that through the footprint in Malaysia, and when revenue growth does resume, Harlan, I expect we continue to, to drive, gross margin consistently higher. Yeah, I mean, that's, that's been the biggest delta relative to the model from, from my standpoint.
Great. Well, we appreciate the participation today, Doug. Good to see the re-upping of the buyback and the confidence in the business. Thank you.
Great, Harlan. Thanks for having us.