Lam Research Corporation ($LRCX)

I'm Vivek Arya from our semiconductor semi-cap equipment team. I'm really delighted to have the team from Lam Research, join us for this fireside and Doug Bettinger, the Chief Financial Officer. And as always, I'll go through my questions, but please feel free to raise your hand if you would like to bring something up. Before we start, Doug, I believe you have a safe...

Yes. I need to start with the safe harbor to keep my legal team happy. Let me give me a couple or a minute here. Today's discussion may include forward-looking that are subject to risks and uncertainties, and actual results may differ materially. Additional information concerning factors that could cause results to differ materially from those forward looking statements [indiscernible] the SEC, including our most recent 10-K and 10-Q. I'm done.

That's the exciting part of the...

I know people are always excited to see how fast I can read that.

So maybe just let's start with kind of a state of the union, right? Since the start of the year, we have seen them consistently raised expectations for wafer fab equipment market this year. So from -- started in the low 30s now it's kind of over 140 right, plus [indiscernible] this year. What do you think have been the incremental drivers of growth since the start of the year?

I think the first thing to understand is the industry, from our point of view, is constrained by clean space availability. I think you're hearing that statement from everybody. And you're absolutely right. We started the year thinking WFE was going to be 135 was a specific number. On the last call, Tim and I updated it to 140, perhaps with a bias to a little bit of upside and what's changed is it's been across all segments of the business. There's just a little bit more clean room. Maybe projects got pulled in a little bit. One of our customers bought a fab and was able to take equipment a little bit sooner. Everybody is trying to figure out in this constrained environment, how to get -- how to squeeze a little bit more out of what's there. And that's exactly where we saw the upside. It wasn't any one specific thing. Just a little bit everywhere. And I'm not saying there's going to be a little bit more, but everybody is working on trying to get a little more up, but that's what showed up in it.

Is there a way to think about what the unconstrained [indiscernible] this year, like if there was enough clean room space.

Everybody wants to know the answer to that question. And frankly, I don't have a specific number for you, except I would tell you, the industry is undersupplied right now. You're seeing it in memory pricing, absolutely in profitability. Foundry is constrained, advanced packaging is super tight. I don't have a specific number for what unconstrained WFE might be. But what I will tell you is this bodes pretty well for what WFE is going to be next year because this is going to roll into next year as projects come more available into '27. I think it's going to be a pretty darn good year in '27 and maybe beyond that But I don't have a specific number for it every [indiscernible]

Got it. Okay. If you look at visibility, right, that seems to be extending out, right, over the next number of years. Is there a way to contact [indiscernible] What is kind of your usual redo visibility and what it is right now.

I think what I would tell you is, right now, given the environment I described, there's very deep conversations occurring with every one of our customers about, okay, what does the next several quarters look like? What do you have? What could push or pull in or out and so forth based on these clean room projects that are underway? And then the conversation almost immediately transitions to, okay, what does that mean for next year? Because at the end of the day, our customers don't want us to be what constrains them, and we certainly don't want to be in that position either, right? So okay, that project looks like it's showing up in '27 in the second half. And so what might that mean for what you head from us? It doesn't necessarily mean I got a purchase order for next year, but the robustness of these conversations is as strong as I've ever seen it, frankly, and all the time I've been in the industry. That's what's happening. And it's pretty easy to understand, okay, that new fab project is scheduled to be completed next year. And so we can also do our own assessment of, okay, the customer is saying this, and we look at it and say, "Yes, okay, that generally makes sense. That's what's happening right now. And then it follows on, obviously, for us, then we're going to go talk to our supply chain partners. We're going to plan our own physical bricks and mortar consistent with -- what is the customer saying they are going to need from us. That's what's happening right now. And like I said, the richness, the conviction in these conversations is as strong as I've ever seen it.

Got it. And the fact that it's the data center end market that is driving...

It's AI.

It's AI versus in the past, many of the growth cycles have been more consumer driven, right, whether it was PCs or phones and such, [indiscernible]

Was it was mobile, yes, all of those things. And now AI is layered on top of it. And frankly, the longevity of this, to me, feels like it's here for quite a good amount of time.

Got it. Okay. Because it's AI where everyone wants to build as soon as possible. Do you think there is an acceleration in how quickly customers are willing to adopt the high end of the stack, so to speak, right, your leading-edge technology, right, versus in prior times.

Where demand is showing up from these AI compute -- I mean these accelerator chips are very, very big die maybe not the most advanced known, but pretty close. And so that generally means it's the most advanced equipment. It's the newest capability required to deliver these very advanced feature, very small, high aspect ratio features and things that you look at. It's high bandwidth memory, it's a lot of DRAM needed in the compute infrastructure. And it's NAND is also being to show up, right? You got to store these tokens, not to generate them. And so it's showing up everywhere. And it's primarily showing up at a pretty advanced process node, which generally needs our most advanced equipment.

Okay. In terms of Lam-specific outlook, right, Lam has been a market share gainer. I think in the past, you said that the goal was to go from kind of the mid-30s, [indiscernible] towards the high 30s. Where is Lam in that transition? And do you think your SAM kind of expense?

It absolutely has been and will continue to be. I mean that's the unique story about our company. everybody benefits in equipment when WFE raises. That's great. Rising tide lifts everybody. The unique position we sit in and maybe we're a little bit lucky, but also we're executing extremely well. You look across the totality of advanced process nodes in foundry and logic, in DRAM, in NAND, things are inflecting in the third dimension, right? High-bandwidth memory is a 3D structure, 6 [indiscernible], is a 3D structure. 3D DRAM is being worked on is an even more 3D structure. The NAND stack keeps growing. That's a 3D structure. Gate-all-around is an advanced 3D structure. Advanced packaging is a 3D structure. When things inflect in the third dimension, etch and deposition intensity grows that's all we do. And so 1.5 years ago, we began describing this growth in our addressable market. And what we said back then in early '25, we did our last Investor Day then we were coming off a year where we address low 30% of overall WFE. So call it, $0.32 of every dollar spent on WFE was spent on etch and deposition. We said we see a path over the next several years for that to grow to the high 30s. That's what you were just describing to that. As we sit here today, it's in the mid-30s, right? 1.5 years later, after we said low 30s were already in the mid-30s because of the evolution of these architectures, this is going to continue. There's a couple of cool data points, I think, from my point of view, and a cool data points, where we said in foundry from 5-nanometer at, our sand per wafer will double because of the intensity of creating this very complicated CFAT structure. Gate-all-around is a step towards that. backside power is another step towards that. Those are just examples of what I'm describing. So we sit in a unique position within the equipment sector in that everything we do is growing. That's wonderful. On top of that, the strength of our product portfolio has never been stronger from my point of view from the executive team's point of view. And so as we see this growing SAM, based on our products that are coming out like a car like Vantex, like the [indiscernible], like Halo, we call it, we believe we're going to win half of this growing SAM, right? So low 30s going to high 30s. We believe because of what we're able to deliver, we're going to do really well in winning that growing SAM. That's the unique story about Lam Research. Everybody and equipment is going to do well over the next several years. We're going to do even better. We outperformed WFE last year. We're going to outperform it this year. we're going to outperform it for the next several years based on what I see. And what I've been pointing out actually, I think people that know the company know that NAND flash is our strongest end market. That is growing a good amount, but growing slower than the other leading process nodes. And yet we're still outperforming with NAND still sort of on the come line. So actually great about where we're positioned about what we've invested in, about how we're executing. We're going to keep delivering on that.

Got it. I think you peaked at my next question, so let me ask it again with the [indiscernible] I can see. So despite NAND having a relatively slower year, right, versus DRAM foundry logic packaging. What do you think is helping Lam outgrow I can only imagine it's mostly share gains. But people do associate land more with NAND than with DRAM. So despite NAND being a software are, what do you think is helping Lam outgrow the market? And what if NAND is software next year also, can you still outgrow the margin.

Listen, everything I just rattled on about all of these 3D architectures is what's happening. And the strength of the product portfolio is what is enabling us to win a good amount of that. One thing I'll point out, I guess people think of us as [indiscernible] company or the memory company from an equipment standpoint. Last quarter 54% of our systems revenue was in foundry. The quarter before, 59% was in foundry. So yes, I love our memory strength. I love our memory customers immensely. But we've done extremely well in advanced foundry and logic because of investments we made 3 or 4 years ago. We saw some of these inflections coming. We knew we were going to be really well positioned to win and we've delivered on it, right? It's gate-all-around. It's backside power that really hasn't meaningfully showed up right yet. It will -- it's advanced packaging, where we do the through silicon via steps, I call it the drill and fill. We etch the silicon to create the space for the interconnect. And then we have a wonderfully strong electroplating business. That's the film, the drill in the film. We own a good amount of that across the totality of the industry. HBM, [indiscernible], over all of those things, we enable with the TSB. And we do other things in advanced packaging, but we're just really well positioned.

On the memory side, Doug, if we were to think of a scenario where, let's say, if memory pricing goes down next year, for whatever reason, do you think your memory customers would still be as incentivized to put money in all these fabs, build all these clean rooms? Or do you think their thinking is very sensitive to the pricing of memory?

I think everybody is being very purposeful about investments that they're making. And honestly, if you think about everything around AI, there are many things that are constraining the ability of everybody in the industry to supply true end demand be it power, be it data center, be it leading-edge foundry being high bandwidth memory, all of these things right now, the demand is beyond what supply is able to support. And so I see everybody lagging their way into trying to step that forward to a certain extent, but also with a clear eye on profitability. That's important. And so as long as everybody is as profitable as they are, they're going to invest, and they are investing.

Got it. And one thing we have seen is several of these memory companies starting to do more long-term with their customers. [indiscernible]

Good for everybody.

Yes, it kind of helps extend that visibility. So has the nature of their discussions with their suppliers, i.e., have they changed because now they are on the hook to provide and assure that assurance of capacity for the next few years?

I think that's probably a part of what's behind these very rich, robust longer-term conversations that I was describing earlier in our talk, Vivek. I mean, like I said, the confidence in these conversations is very, very high. And this is probably certainly part of it.

Got it. On DRAM, how is your content changing from 3 HBM 3 to 4? And I saw yesterday, Samsung talked about as well. So how is your kind of content evolving along these different HBM generations?

Yes. As the stack grows and as the dimensionality of the holes we're drilling gets tied the need for equipment grows. Obviously, right, if you go from 8 to 12 to 16 die you're putting together for the same construct, you need more equipment. . . And again, we own the TSV. We do most of the TSV for everybody in the industry. And so that's a big part of what's showing up. And then if you layer on these diet bigger, the trade ratio changes. You've got the process noted and of itself going from one bidder or on so on gamma and beyond. Our SAM grows just in the process itself and then you layer HBM on top of it.

Got it. Does your opportunity change, Doug, if memory goes from kind of more conversions and upgrades towards more greenfield, if you could kind of walk us through, does it make any difference to them.

You're really asking a NAND question now because largely what we see happening in NAND are conversions, right? The installed base is converting from 1 process node to the next is what's happening right now. I'll take you back to 1.5 years ago, we described the view that the industry would need to spend $40 billion over several years to go through these conversions. On our last call, Tim updated that statement to be, okay, that $40 billion, we believe, will be largely complete by the end of next year, so it's happening sooner. And so at the point that you get through, okay, things have been converted, you're going to need capacity at it. We're happy with conversions. We get a bigger share of spending. When wafer capacity gets added, the customer base will need to spend more our share is still quite strong. So we're almost agnostic between 1 or the other. We're here to support customers in whatever makes sense for them. right now, that's largely oriented around emergence. But at some point, you're going to need a little bit more wafer capacity.

Do you think the allocation of new clean room to NAND kind of lags because most memory companies are a lot more profitable in DRAM. So it makes more sense for them to allocate [indiscernible]

That's what I observed happening whether people are consciously saying this or not. Right now, clean room is a constraining item. Three of my customers do both NAND and DRAM. And to the extent that there's cleaner constraining things, what are you going to invest in? When you're going to invest in where the highest profit opportunity is. And right now, that's DRAM. NAND is getting closely caught up, though, and profitability is quite attractive in NAND right now. So and will accelerate at some point but DRAM is getting the priority is what I observe happening.

Got it. The fact that over time, more clean rooms if let's say, there is a lag with which clean rooms get allocated to NAND, does that extend your growth cycle kind of further out, right, versus right, some of your peers who might be more DRAM exposed? Is that...

No, I think so. But listen, everything is tight right now. And so that's going to extend the investment profile until these constraints start loosening up, which I just don't see happening in the near term.

Okay. And then within the clean room that customers have, do you see them upgrade their tools faster? And I'm talking more DRAM rather than NAND.

Listen, there's always upgrades that happen -- that's a high return way to get the next-generation tool capability is if there's an upgrade path for a tool, almost always the first priority is going to be do the upgrade before buying new equipment.

Okay. What do you think has helped Lam do so well in foundry logic, you mentioned over half the business was in foundry the last 2 quarters. Is it all share gains? Is it just that the pace with which technology is rolling out? You mentioned the change in transistor [indiscernible] and transistor forms, what is the [indiscernible]

Yes. It's back to what I tried to describe earlier, [indiscernible] structure going to gate all around. And if you look at the little pictures of this, there's these nano sheets that need to get created that get created by depositing material and then etching it sometimes selectively etching it. So it's a different way of etching. You're doing it sideways that etch and deposition. So our SAM grows, the opportunity to sell more equipment grows because of the technology inflections that show up. So that's 1 statement, right? There's ALD steps in there, there's selective etches in there. there's always conducted dielectric etch in there. It's just -- it plays to the strength of what we do. And the product portfolio is very strong. we are the unequivocal leader in conductor etch in this industry, unequivocal by out lot. And so when you see that showing up, that helps. And then you layer on top of that advanced packaging. Again, back to the TSV, these are 3D structures. It's just what we do.

Got it. One other very key part of your business is the customer service business group, right, CSBG.

In many ways, it's my favorite part of the business model is the customer support business group.

Right. And now over $2 billion, right, on a quarterly...

2.1 last quarter.

Do you think the growth rate -- when do you think the growth rate there starts to converge with your tool business? Or it is -- that's just a natural consequence of this is kind of a long tail kind of secular, right, embedded base business as opposed to something that's exposed to new tools?

Yes. So let me unpack it a little bit for those that might be new to the story. The customer support business group we call it CSBG, 4 components of that business, it's upgrades, it's service it's spare parts. Spare parts is a pretty large component of it. And so when you think through why is it so strong right now, utilization in the industry is very high. right? It's basically running at 100%. So if you think about spares and service the spares and service is modulated by the number of chambers in the field as well as utilization. Utilization is 100% can get higher. So in the March quarter, that's why it grew so much. 100%, you can't grow beyond 100%. So those 2 things in the near -- the more near term are going to grow, but not as much as you saw in the March quarter. So that's why I tried to pull people back a little bit relative to what's going on. The place we're innovating in this is what we call advanced service. It's using equipment intelligence and cobots to deliver service in a different, more predictable way frankly, customers like quite a lot. So that portion of -- when you think about service, that drives incremental growth on top of everything else. But I love this part of the business model. And in fact, a lot of these meetings I've been doing today and last week, people don't ask about this part of the business. It's 1/3 of the business. Very profitable, very recurring, right? Fabs are always running, which means they consume spares and service. So like I said, in many ways, to me, as the CFO of the company, this is my favorite part of the business model. It just keeps -- it keeps running. People are often surprised to hear that our tools literally will run for decades. -- the Reliant product line I didn't talk about, but this is the part of the business where we're selling tools that have been around for a long time. It used to be refurbished tools today. There's almost no refurbishes, Nobody has given given up equipment, but it's selling older model equipment into fabs. That's a really good part of the business model because the tool was designed a long time requires not a lot of R&D. We're just building the same tool that we built 10 years ago in selling it. So you see that in the analog space in power in CMOS image sensors, a lot of the business that we have in the China region is the Reliant product line, great part of the business. So anyway, when you put this all together, it's quite profitable. It's quite cash generative and it's very recurring.

Got it. I know we have had this discussion before, but do you think you would ever feel comfortable giving like a backlog, right? Because if this business is to be rated as, let's say, a SaaS, right, subscription-type business, then people would also love to know what is the backlog.

We don't.

Okay. I asked it answered.

Yes, I think to think everybody when you ask me about backlog, you want to know the visibility we have in business. I already described what's going on there, right? [indiscernible] constrained. Industry is going to grow again next year, right? I mean the nature of the conversations showing up are very robust. You don't need backlog to hear me describe that. And I'm a pretty conservative guide generally speaking. I'm talking as optimistically as you've ever heard me if you've listened to me talk for a while because I've not seen it as rich as it is right now with these conversations moving. Right.

One thing that I think Lam has perhaps said or some of your peers have said is that there is a potential for just fab equipment to grow faster in '27 than in; 26. Is that still a reasonable expectation do you feel?

I think '27 is going to be a pretty good growth year, again, because we're constrained this year. All of the unmet demand we're rolling next year. clean rooms will become more available if you look at totality of projects in the industry, should be a pretty good year.

And do you think the mix changes in any way? Does it favor any one of the areas.

I think you're going to see continued investment in leading-edge foundry. You're going to see continued investment in DRAM, and you're going to see NAND begin to catch up a little bit to those other 2.

Okay. One other question that, Doug, as you've seen come up is, are semi caps extracting as much value as every other player in the ecosystem. And when we have definitely seen gross margins that used to be in the mid-40s have now come up right towards. [indiscernible] Do you think that's a durable trend and how much more upside is kind of left in this margin expansion journey.

Yes. Listen, I think we're very focused on right now and making sure we're getting fairly paid for the value we're delivering. And that's not a new phenomenon. That's always been the case with our business. We're always working on that. And your observation because it got the [indiscernible] right? For the first decade I was at the company, our gross margin was pretty consistently in the middle 40s. And then we moved into the high 40s, and we're now touching mid -- or excuse me, low 50s. And we've got inflationary headwinds we're dealing with when you got oil above $100 a barrel, right? We fly stuff all over the world, you've got to manage all of those type things that show up both for us directly as well within our supply chain. We're dealing with a little bit of a headwind in customer mix, right? Some of the most profitable customers aren't growing nearly as much as the biggest customers who tend to get a little more favorable pricing a little bit anyway. And so as you think through that, I went out of my way on the last call to say, listen, as we go through the year, we see these headwinds, we're working on efficiencies. We're working on pricing. And so I said keep gross margin right where it is. In that 50-ish percent, 50.5% is what we just guided to, we're managing the headwinds that we see in the short term through a variety of things, including pricing.

Historically, WFE and semi CapEx have kind of grown in that 8%, 10% range. But here we are in front of multiple years of 25%, right, [indiscernible] [indiscernible] How is Lam prepared, right, from an operational perspective, from a capacity perspective? Where are you seeing the constraints in your ability to kind of execute to that growth?

Listen, I think, again, also on our last call, we went out of our way to say, hey, we're building a second manufacturing facility in Malaysia. So that's part of this. We anticipated what was happening, and we're getting ready for it. Like I said, we are spending a lot of time with customers working on the demand signal they're providing to us such that we can then make sure that propagates back through our own supply chain. And so I don't want to say it's working flawlessly perfectly when you build this complicated a product as we build, there's always things that you're having to expedite and work on and it costs money to expedite things. And so we're managing all of that pretty effectively. It's not to say that it's easy. It's absolutely requiring a ton of time, a ton of effort to make sure every single component in the bill of materials shows up when it is required and needed, and we're having to expedite lots of different stuff, but we're managing it pretty well. And I know my team listens to me when I talk at these things, I want to thank the global operations organization at Lam Research. I know everybody that works there is working super hard on all of this stuff and it's enabling execution that is extremely good from us.

Got it. Hypothetically, if next year WFE grows 35, 40-plus percent, what would Lam need to do differently today?

We're doing everything we can do to make sure we are not going to be the constraining item in the industry based on the demand signal we're getting from our customers. So we're actioning everything that I think we need to do to be ready for what customers are telling the [indiscernible] and I'm not going to say that growth number you said is the right growth number. It's going to be a good year next year, but I'm not going to endorse that number necessarily.

Got it. China, should 1 just assume it kind of stays in this 25%, 30% exposure, and that's just the easiest way to...

I think so right now I mean the way we've described the wafer fab equipment spending in China it's flattish to slightly up this year from last year. And as a result, that's a percent of overall WFE and as a percent of research is revenue, it will decline because everything else outside is growing faster than that. So that's the way I think about China. It's not going away. It's just pretty steady.

Got it. On just kind of capital allocation. So I think the semi cap industry has done a remarkable job, although not always appreciated as much for kind of buybacks and returns, right, retiring -- actually retiring a bit on right of outstanding shares. How do you think about capital allocation? And is there a scope to actually target even higher dividend yields? Or you think buybacks are still a better way to use your cash?

Great question. Listen, our plans are to return 85% of free cash flow to shareholders. In the last several years, we've returned more than that. Underneath the covers of that, our intention is to grow the dividend on an annual basis. I think the last 3 years, we've grown at annually 15% we'll grow the dividend again this year. I'm not going to apologize for the fact that the dividend yield has declined because the share price has gone up. That's a good problem for all of us. I do look at that yield. I do benchmark it to everybody else in the industry. I want to make sure we're competitive with that. But then to get to the 85%, we supplement that annually growing dividend with share buyback. And that's still the plans of the company.

Any -- I know M&A stuff in the industry, but any place where you think there is a scope for kind of tuck-in acquisitions to...

I don't know. Over the years, we've done some small tuck-ins. And in fact, we did a panel packaging tuck-in that I'm super happy that we did a few years ago. But that's all that's left in the industry. The big stuff is in the rearview here, large-scale M&A, I think, is in the past.

Thank you so much, Doug.

Appreciate it.