From BFA Semiconducr team. I am really delighted and honored to have the team from Lam Research join us this morning, and Doug Bettinger, the CFO of Lam Research. We will start with a quick safe harbor statement from Doug, and then we'll get into the Q&A. Please feel free to raise your hand during the session if you'd like to bring something up. With that, back over to you.
Hey, listen, I encourage you to take a look at our safe harbor language. It's on the website. You can see it here in the room. I am not going to read everything on that slide, but it's important that you understand to the extent I make forward-looking statements. I don't plan to tell you anything that the company hasn't said before, so you're not going to get anything new in here today. Anyway, just keep this in mind as we go through.
All right, let's try and make this exciting.
Let's try to make it exciting.
Yes. Thank you, Doug. Maybe let's start with the state of the union as you see it, right? A lot of macro and micro cross-currents. What I would really like, Doug, if you could speak to the analyst day that you held recently. One thing that really impressed me from the analyst day was this prospect towards higher gross margins, right? We always pay a lot of attention when companies talk about the upside in gross margins. Maybe give us how you see the demand environment right now, and then maybe some.
Everything I could actually talk for the entire 30 minutes. I'll touch a few of them, and then you can redirect me, and we can dive in. I guess first, let's start with this year. When we look at overall spending this year, we think wafer fab equipment is roughly $100 billion, up from mid-$90 billion last year. So think about, I don't know, mid-single-digit growth this year. If I unpack that a little bit for you, leading-edge foundry is quite strong. DRAM is quite strong, driven by DDR5 and HBM, high bandwidth memory. I don't think that's a surprise to anybody. Both of those things are driven heavily by what's going on with AI compute, parallel compute, these great big accelerators and whatnot. I'm sure you want to ask about that. Logic, so-so, right? If I separate foundry from logic, more strength on the foundry side.
NAND, I think everybody likes to hear from us with NAND because they know we are critically enabling to the NAND industry. NAND is going through an upgrade year. I think actually for the next several years, that's largely what you're going to see. When you unpack that, we're coming off two years of very modest amount of spending in NAND. By no means am I suggesting to you it's back trending towards peak investment levels. Nowhere close to that. The simple fact that this set of our customers is upgrading what's there, and there's reasons for that, technology reasons related to QLC and enterprise SSDs and so forth, and I'm sure you'll ask me more about that. We see that happening this year. That also is contributing to the growth.
That's an outstanding situation for us because when you go through upgrades, we are the equipment that gets upgraded. That's what we see going on this year, Vivek. You asked a little bit about the investor day. There's a wonderful longer-term story here. You can unpack this with more specific questions. When we look at over the next several years, maybe trending to the end of the decade, what you have going on at a very high level is the evolution of 3D device architectures. Things are going 3D. NAND did this 10 years ago. You're now very much seeing it in foundry and logic with the gate all-around structure and advanced packaging, CoWoS. These are 3D structures. When things go 3D, you need to create these interconnects. You need to deposit materials to create the architecture. This is all about etch and deposition.
When you think about, frankly, WFE is going to be a nice place to live, right? We live in a good neighborhood. Within that, our view of etch and depth share of that spending moves from the low 30 range to the high 30 range. Our markets are getting bigger because of the evolution of architectures. We're in a good place. Compounding that for us specifically, you know we're coming off several years of pretty big growth in R&D, which has positioned us to have probably the most competitive product portfolio in the history of the company. I think, Ram, our CEO said that on the last earnings call, if I remember right?
When we look at this growing market in terms of etch and depth, we believe we are going to take at least half of that growth because of the strength of the product portfolio. This is driven by things like the Aera product that we announced, the Halo product that we announced. We're positioned extremely well. When you look at the growth prospects, I think it's actually quite compelling. Anyway, I'm rambling on now, so you can redirect me on all of that stuff. There was a lot to unpack. I didn't talk about margins, but there's a wonderful margin story here. Also, maybe I'll ramble on for another couple of minutes. You know, I always do a ton of homework when I get ready for earnings.
What I did this most recent time is I went back and looked the last time the company touched $5 billion in revenue. It was the September 2022 quarter. We just guided for the first time back to a $5 billion quarter in June. Here's an interesting comparison. In that quarter, same revenue level, gross margin was 46, September 2022. The March quarter we just printed, gross margin was 49%, and we just guided June to 49.5%. Part of this comes from the fact that we have a very strong product portfolio, so that's good. More importantly, though, and those of you that follow the company for a while will remember, we suggested to you, hey, we're going to adopt this close to customer strategy. We started talking about that in early 2023.
As the company has grown the top line, we've begun to gravitate that incremental volume to that close to the customer location, which a lot of the customers in this industry are Asia-based. You are absolutely seeing the benefit of that in gross margin today and in operating margin because of the proactive activity we undertook back in 2023. I'm very proud of what the company has been able to do. This is very much, we told you we were going to do this, and we did it. We are doing it. Anyway.
Got it. No, absolutely. While we are on my favorite topic of gross margins, which is 49.5%, I think you mentioned that that actually includes the effect of some of the tariff issues that were. Since that tariff discussion seems to have cooled off somewhat, is 50%?
Cooled off, and then the 232 steel and aluminum came. Listen, we just need to get the rules finalized so we know how to manage in this environment. That is what I am hoping for. Let us just sort through it, and then we will figure out how to run the company in the most efficient way we possibly can. Absolutely, you are right. That 49.5%, to the best of our ability to predict the impact of tariffs, contemplates that.
Got it. While we are on that topic, we'll come to the longer-term fundamentals. Last Friday, there was some news about the U.S. Department of Commerce, you know, perhaps widening the scope of restrictions for certain China entities that have over 50% ownership by sanctioned firms. Is that something that has come across your desk, Doug, in terms of how it impacts your business with China or?
To the best of my knowledge, it doesn't have any incremental impact.
Okay, makes sense. On the China issue, yes, you know China WFE is lower this year, right, based on everything we have heard, down 20%-25%.
Our description is as a % of total revenue or a of total WFE, whatever you want to look at, it's kind of the same thing. China will be down this year from last year.
Right, from like whatever, 40%-ish to 30%-ish.
That's right. Last quarter, our revenue, 31% was China-based.
Got it. You think it is now kind of predictable for as much as any China business can be predictable? Is it fairly predictable for the next several quarters, or are there still upside, downside risks, assuming whatever the tariff situation is or restriction situation stays as is?
Yeah, I mean, listen, to the best of our ability to understand everything customers are doing, it isn't any different in China versus anywhere else in the world. The way we manage the customer relationship, we've got account teams in each region that are responsible for every customer that we sell equipment to, right? They'll describe, here's my plans, here's what we intend to do. Now, people can change their plans always. That does happen. But you know somebody at Lam every day, every week, I don't know with what frequency, pretty frequently is talking to every one of our customers and understanding at least how they represent to us, here's what they plan to do. No different in China versus Korea, Taiwan, or Japan, or the United States, or anywhere else.
Got it. Then looking longer-term, Doug, you mentioned NAND upgrades as a very strong growth driver for Lam over the next several years. I think there was a $40 billion addressable opportunity.
Yeah, what we said at the investor day was, when we look at where we believe the upgrade path in NAND to evolve, we believe spending will be roughly $40 billion, and we've described it over the next several years to go through this upgrade cycle.
Right. So far, we have only seen, I think, one really large customer go through those upgrades. When do you think that kind of broadens out when you see a lot more customers going after NAND?
I think this year it's going to be quite broad. It's not going to be one specific customer. In fact, it isn't right now. It's broad industry adoption. Maybe let me describe what's going on in NAND because everybody's interested to hear our point of view on this. When you look at the installed base in the NAND customer base, two-thirds of the bit capacity is at a layer count that is below 200 layers. Why does that matter? Why are we talking about layer count? It's not layer count per se, but the equipment capability at that sub-200 layer count isn't able to actually manufacture a QLC device. When you look at where the market is going right now, enterprise SSDs are one of the growth drivers in NAND because of AI compute and all of the stuff going on there.
If you're one of my customers intending to compete in the SSD space, you almost have to start upgrading to a layer count above 200 so you can manufacture QLC so you can be relevant in the enterprise SSD space. That's what's going on. It's not bits for bits' sake. This is where the market is going in one very important growth segment of the end customer base. That is very much what we see going on. Upgrades to begin to enable QLC and at layer counts above 200, that's where we see things happening. It is very much all upgrading the installed base. The wonderful part for us anyway about upgrades of the installed base, when that spending occurs, our SAM is two-thirds of the spending. Two-thirds. It is a big opportunity. Why is that?
It's because we are the constraint tools for the most part in a NAND fab. It's that stack. It's that deep high aspect ratio etch on through. It's the tungsten metalization that ultimately will begin to evolve to molybdenum. Let's just call it molly because it's really hard to say molybdenum. Molly, right? You are going to begin to see new equipment sales that enable molly. That's what's happening underneath the covers of that $40 billion over the next several years, Vivek. That's what we see going on.
Got it. Is there a simple way to quantify, Doug, how much of the progress have you made towards that $40 billion or might make this year?
Early days. Only just beginning.
Like the first quarter or so?
Yeah. First inning of a 90-game game. Call it that. Just starting.
Got it.
What I would point out, if you guys pay attention to our revenue, you saw strength in NAND last quarter. We have suggested to you in that June quarter guide that the highest percentage growth is going to come from NAND again. Highest % growth, maybe not dollar growth. Dollar growth is probably going to be foundry. It is happening.
Right. Does this historically turn out to be lumpy, Doug? Like any kind of infrastructure project, like is it like they will do these upgrades for one or two quarters and they might pause just because there's only a handful of customers that have the ability to do that?
Yeah, that could happen. Nothing in this industry grows every quarter, every single quarter, or stays flat every quarter. Yeah, there will be some variability for sure, just like every part of the business.
Right, but it's fairly predictable, at least for the next handful of quarters as far as you can see it.
As far as I can see.
Right. Got it. Okay. How does this benefit accrue to your services business, right? The customer service and business growth?
Yeah, so let me go through CSBG, if you'll allow me.
Please.
It's not quite the question asked, but let me unpack that and then come back to your question. One of the wonderful parts of the business model at Lam is what we call the customer support business group. There are four components of that: spare parts, service, upgrades, and then what we call the reliant product line. Great part of the business model for the simple reason that our equipment, it runs forever. It almost never goes away. When you think about this and we give these numbers at the end of every calendar year, chamber count grows every year. The opportunity to sell gets bigger every year, spare, service, and upgrades. Anyway, great part of the business model. On average, after we sell the tool, we generate more revenue after we sell it than when we sell the tool itself.
That surprises people sometimes when I tell them that. It's because our tools run for decades. I'm rambling on. Let me now come back to your question. The upgrade part of CSBG is what we see benefiting in a pretty significant way this year because of what I described going on in NAND. Offsetting that this year, though, the Reliant product line is somewhat softer because it was quite strong in China spending last year. It's part of why we describe, hey, we think China's down a little bit this year. Anyway, CSBG in total, we've described as likely flat-ish this year, but real strength in the upgrade piece of it.
Makes sense. On the CSBG side, you know long-term, do you think as you kind of come out of this China, is that a double-digit growth business? You know, one of your competitors has kind of described their services business in terms of subscription-type contracts. Is your CSBG business constructed that way? Is there a part of it that is very regarding it?
Absolutely. There is a part of it that way. More importantly, the predictability of spares consumption, whether it is under a contract or not, is quite high. If the fabs run, you are consuming spare parts, right? Stuff wears out. I compare it to your automobile, right? You have to change your brakes every once in a while. You have to put new tires on the car because stuff wears out. The same is true with our equipment. Excuse me. Anyway, I forget where I was going with that, but that is the nature of what is happening there.
How much of CSBG do you think is that kind of subscription or recurring?
We haven't put numbers on it, but the biggest individual component in CSBG is spares. Like I said, whether it's under a contract or not, it is completely recurring.
Got it. One other question, Doug, that comes up on NAND is, when can the NAND market get back to prior peaks if one really large customer in China is restricted from the market?
Yeah, I don't know the answer to that question, Vivek, right? Us describing this $40 billion over the next several years doesn't assume it gets back to that $22 billion peak investment level. There was a ton of investment that year in NAND. In the horizon I can see right now, we probably aren't going to get back to that investment level. However, for us to get back to nearly peak NAND revenue, we don't need to get back to that level because of that two-thirds of the SAM when upgrades happen coming to us. I think that's important to understand.
Got it. Okay. On the leading edge side, you know over the last number of years, what we saw was consumer market had very frequent upgrade cycles, right? That is why you saw the leading edge foundry always invest in leading edge capacity.
When you say consumer, you're talking about mobile phones.
Yeah, smartphones, right? PCs. Those were really driving a lot of the really bleeding edge, right? Now what we are seeing is a slowdown in those consumer markets. If I look at the nodes that the AI companies are adopting, they're actually N minus one or N minus two. What does that say to you about leading edge demand and that as a growth driver?
Leading edge demand is still quite strong, right? When I listen to one of my big customers talk about how compelling two-nanometer looks, the first gate all-around node, they seem quite optimistic about power performance, all of that stuff taken together. Yeah, it is still driving a good amount of investment.
Got it. How is your.
Listen, the other thing I would say is, yeah, phones are not growing like they used to for sure, but they are growing a little bit. That is important. More important is there is still a content story in mobile, which is when you look at the average DRAM content, it is going up every year. When you look at the average NAND content, it is going up every year. The high-end SKU might not always be going up, but it is still mixing to a higher level of content. That still is an important thing to be thinking about, especially in DRAM and NAND.
Nice. You mentioned in the foundry logic, logic has obviously come down for all the well-known reasons. That customer has often described that they have a lot of tools. I think they use the term bubble wrap, that they have a lot of tools flying around. Do you think that that's a risk factor for the industry over the next several quarters, or is that pretty much contemplated in how you think about it?
It's pretty much contemplated.
Okay. That was easy. DRAM, how do you see that business growing this year and what are the key drivers for Lam?
It's all about high bandwidth memory. It's all about, first, what's going on? If you look at DRAM, WFE year on year, last year to this year, it's flat-ish, but the composition of it's different, right? You had a customer in China spending a good amount last year, likely spending not much this year. The rest of the industry growing spending. What's going on there is, at least architecturally, a couple of things to think about. DDR4 going to DDR5. Some of these new server CPUs need DDR5. So there's a process upgrade just in and of itself. Layered on top of that, AI compute pulling a lot of high bandwidth memory. That's an important part of what we see going on. We do important steps in that through silicon via the TSV process. But for the most part, we own all of it.
I call it the drill and fill, kind of a tongue-in-cheek way to describe it. We do the silicon etching to create space for the interconnect, and then we do the electroplating to put the copper interconnect down. Very strong part of what's going on with our business right now in HBM. I love what's happening there. You have got a process node migration. You have got the move to more and more HBM, all three of the global customers aspiring to compete for that and spending consistent with that outlook.
Right. Does your content change as the industry goes towards HBM4?
If you think through, absolutely, right? Because you go from an 8 die stack to 12, potentially to, not potentially, to 16 is what's being contemplated. That's more capacity.
Got it. Is it linear content growth for you, or is there?
Not necessarily. No, there's a yield calculation that you have to contemplate when you go through this. The die gets bigger, right? For HBM, it's a little more complicated than just ratcheting things up. It's not quite that easy to sort through. Each customer is in a different spot of the learning curve as well.
Okay. Doug, I had just one or two questions on the competition side. First is we get all this third-party data from Gartner and whatnot. I think last year they had Lam gaining share in deposition, but in the etch side, it came down somewhat. How would, and your Japanese competitor gained some share. I think there were parts of the market where you could not address. How do you look at your market share positioning in your core areas right now?
Yeah, actually, it's very strong in etch and deposition. I don't pay a lot of attention to Gartner, Vivek, if I'm honest, but I would guess likely what is showing up in whatever math they're putting together, two things. One, there's a mix component to it, right? Our strongest end market is NAND. Last year, NAND was spending almost nothing. We sold a lot of etch tools into NAND. It's not that somebody took a position from us. That did not happen. Our strongest end market wasn't spending much last year. That's one thing. Two, we're restricted from a bunch of customers in China, and the rules in Japan are different. The Japanese are not restricted from that. I'm sure that's part of it as well. I'm speculating. I haven't even looked at the Gartner data, but I'm guessing that's what's showing up.
Okay. How do you think, Doug, over the next several years, competition from domestic China vendors? Because what we have seen, and I know this is apples and oranges, but what we saw, for example, in AI or more on the software side, just given the constraints, there was a lot of innovation that happens there. How should we think about that aspect when it comes to the equipment industry? Where is their state right now, and where do you see it going, and how are you positioned?
Listen, I think largely what you're seeing is they're growing a lot. They're growing a lot because there's a whole segment of business we can no longer compete for, right? There's end-use and end-user restrictions in China, and we're not there. We can't be there, right? We're restricted. If you think through the customers that are in those categories, what are they doing? They're buying the equipment they can buy. A lot of that is China.
Right. How would you assess where domestic China is versus where the Western peers are from a competitive perspective?
Technologically, we're way ahead. I mean, here's the thing to understand. This year, we're spending roughly $2 billion in R&D. We've been around for 45 years. I've got the best etch engineers in the world trying to do things that's never been done in the world before. We don't know how to do it, right? Customer needs a different result showing up on the wafer that they're challenging us to figure it out. It's hard. I can't overemphasize how hard it actually is to do what we're trying to do. I'm not dismissive in any way about the emerging competition coming from China, but the way you deal with things is you just keep running fast. You just keep innovating from a productivity and a technology standpoint. Almost always, in my experience in this industry, if you do that and do it well, you win. That's what we're doing.
Got it. The reason I'm kind of focused on that question, Doug, is that often when we look at WFE intensity over long periods of time, we kind of assume that it's going to stay at this mid-teen space. But when you have such a large customer in China, right, where WFE intensity has been so much higher than average, if they continue.
A large set of customers.
Large set of, exactly. Large set of. As a block, their WFE intensity has been so much higher than average. As they continue to come down, can there be enough strength in everything outside of China to kind of keep that WFE intensity going?
I think so. I mean, when you look at the evolution of these 3D architectures that I've been describing, it costs more. The complexity is growing. Absolutely, if you unpack the set of customers in China, a lot of smaller customers, new, trying to understand how to do stuff for the first time. They are early in the learning cycle. They are spending and then learning how to do stuff. That is an inefficient spend, generally speaking, if you categorize it that way. Maybe a couple of interesting data points. Let me do one. When we look at foundry as an example, you have got this evolution from 5 nanometer to 2 to 18A to 14A and beyond, and the evolution of a couple of gate all-around nodes to eventually something called a CFET. Complexity in that situation is growing.
For the magnitude of doubling per wafer in our addressable market, if you look at 5, and we did a bunch of interesting stuff at the investor day, I encourage everybody to go take a look at it because we worked really hard on that stuff. I think there's really insightful analytics that were put together. In this case, from the 5 nanometer node to that CFET node, etch and dep intensity per wafer doubles. That drives a growth in WFE, obviously. You have similar things as the NAND layer count grows and as DRAM goes from 6F to 4F squared to eventually a 3D architecture. You have similar type things. That's what's going to drive WFE.
Right. So lithointensity goes down in that time frame?
Lithointensity goes down, generally speaking, with 3D architectures. You saw it in NAND. I think you may continue to see it here. Listen, EUV, high-end EUV is going to be critically enabling to the industry. It is. I'm not suggesting anything different than that. When you look at these 3D architectures, you do not need litho to create gate all-around necessarily. You need etch and deposition, advanced packaging. I would tell you the same thing. You need the drill and fill. You need TSV. Anyway, I cannot necessarily speak to exactly what is going on in their market. I can speak really well to what is going on in etch and dep. And I know that is growing for sure.
Got it. How would you, in the minute or so we have left, how would you encourage investors to look at China contribution to this market over a long period of time? Right now, it's 30% of the industry. Is there a level where there is kind of a natural resting place or it's hard to, because every year we'll come to Q4, there are always restrictions that people hear about. What is this industry kind of prepared for in terms of China contributions over a longer period of time?
I think five years from now, we're not going to be having this conversation about China, right? These new set of customers, my phone's ringing, sorry, are going to get better. Maybe some consolidation will happen. It'll just fall into the background of this global industry. Nobody ever asked me about what's going on in Korea. I got two really big customers in Korea. What's going on in Taiwan? You got a really big foundry customer in Taiwan. Why are we fixated on China? Because they're new. And several years from now, they will no longer be new, and they'll either get good at doing what they're doing or maybe they consolidate together. I'm not exactly sure. But I think five years from now, we're probably not having this conversation.
Yeah. Okay. And then finally, Doug, what has been very impressive is Lam's ability to generate a lot of free cash flow even when your prior core market came down significantly. So like mid-20s, I mean, it's better than most of the analog companies we cover at much lower multiples.
This is a great business. We generate a ton of free cash flow. Curiously, I mean, people are sometimes surprised by this. When business turned on in 2023, we were generating record cash flow. Why? Because working capital comes down. AR comes down. Inventory generally comes down, and that generates cash. You're right. The metric that you just described is kind of how we see the business.
Right. So buybacks is still the bigger kind of variable area, or do you see a place where dividend and dividend yields can go up? So semiconductors are actually seen as good dividend yielders also.
Yeah. Listen, the way I think about free cash flow, we're going to return at least 85% free cash flow. Over my tenure at the company, it's been over 100%. We're going to grow the dividend on an annual basis. We benchmark it with the entirety of the industry to make sure we're competitive. We top that off with a buyback. I think that's likely what we continue doing.
Got it. Stick with that. Great. Thank you so much, Doug. Really appreciate it.
Thanks for having me.
Thanks.