Lam Research Corporation (LRCX) Earnings Call Transcript & Summary

Welcome. I'm Blayne Curtis, semiconductor analyst at Barclays. Happy to have with us for our next presentation, Lam Research. From the company, we have Doug Bettinger, EVP and CFO. And I pass it to him. He'll cover safe harbor quickly, and then I'll get off to our 30-minute fireside, but off to you, Doug.

Yes, beautiful, Blayne. Thanks for hosting us, by the way. Good to see you today, and good to see everybody else also. Just real quick, I'd encourage everybody to take a look at the safe harbor slide that's up right now. I always like to keep my legal team happy by reminding people that I may make forward-looking statements on the discussion today. I encourage you to take a look at the safe harbor language that provides some level of guidance around what I may or may not talk about. And with that, Blayne, why don't we jump into yourself?

Perfect. Yes. So maybe let's just start off with the most topical point. I mean stocks generally have been kind of sideways and safe for earnings. I think there's this concern about totality that just continues and continues and continues. Semi companies across the board all sound very constructive about where we're at and actually looking for continued growth. So maybe just start with that, like what are people getting wrong here in terms of what you're seeing on demand and the visibility you have going forward?

Yes. Blayne, I continue to think people just worry too much about the cycle. Listen, I've never tell you or anybody on the webcast here, that there's not a cycle in this industry. There is, but the nature of it is very different than it used to be. Quite frankly, as I was getting ready for our discussion here today, I went back and looked at WFE. And over the last 6 years -- sorry, I have got my phone popping up, 5 of the last 6 years have been growth years. That's not a cyclical industry. That's a growth industry. And in fact, how I often describe the industry today is it's a growth cyclical. There's always going to be a little ebb and flow. It's inevitable when you've got an industry with lots of participants and whatnot. But keep that in mind, semis are a growth industry. Semi cap on top of that when you layer some of the things, we're going to talk about today around capital intensity and whatnot. It is a growth plus semis in my opinion. We think right now, and I think it's general consensus in the industry, $85 billion is the investment level this year. And I think as we all sit here today and looking to next year, all of us are saying, looks like a pretty strong year, it looks like a really good year next year. And in fact, you've got continuation of a lot of the things we're seeing this year, frankly. The secular tailwinds around semis, everybody knows what it is. We all get excited about 5G, AI, automotive, all of the things you've heard everybody talk about for the last, I don't know, 6, 9 months, are growing. Semi content is growing on top of that. Capital intensity is growing. Die sizes are growing. Densities in different devices continue to grow. And when you look at that, it just sets up to be what I think is a great industry, Blayne, quite honestly.

You mentioned capital intensity, and this was another way that people use to kind of gauge what's going on. And that's moved up. You're now at some high teens, and we'll see what the semi market does next year, but I think maybe you can talk about why this capital intensity has crept up and what gives you confidence it will stay at these levels?

Yes. Blayne, I think about capital intensity in 2 ways. One is CapEx as a percent of revenue, which I think is entirely what you're asking about. I also think about capital intensity per wafer, right? That is growing every single process node as you look at leading edge. I want to encourage everybody to keep that in mind, remember that because that's important for our industry. Yes, CapEx as a percent of industry revenue is -- maybe at a little bit of a high watermark, Blayne. But frankly, I think that's the wrong metric to be looking at. I think people think about that metric and look at that metric, and I know I do in terms of affordability, right? Can the industry afford to invest at these levels. And frankly, when you think about that as the CFO of a company, that's the wrong metric to be looking at. The right metric to be thinking about because we all invest out of profit dollars, not out of revenue dollars. And frankly, when you look at CapEx as a percent of operating income or as a percent of EBITDA, it's actually at a relatively low level in terms of historical context. And so that is, frankly, the more important metric, and I know not everybody looks at that metric all the time, but I encourage people to go look at that metric, frankly, because that's the affordability question. And honestly, yes, it's a little bit high as a percent of industry revenue because of rising capital intensity and things like the growth of a 3D NAND stack. The shrink in DRAM is getting progressively more and more challenging as you go from one node to the next. Foundry logic, as you walk down the process nodes, it's getting increasingly more expensive. Now the fortunate thing, I've been in the industry a long time, I know you have also, Blayne, as affordability has never been better in the industry. So is this an affordable level of investment? It completely is. Industry profits in the industry have never been better. So keep that in mind. That to me is the most important thing to be thinking about when you ask the question.

Yes. Definitely a lot more dig into. But maybe just looking at 2020, I know you haven't given -- you'll give official guidance next earnings. But maybe just talk about some of the moving pieces there. Others have pointed to foundry logic being up, memory more mix. Any color you can provide on those moving pieces?

Yes, Blayne, our normal cadences in the December call, we'll give you the complete read of what's going on. It looks like a good year. It looks like a strong year, a growth year next year. And I say that looking at all segments of our customer base, foundry logic certainly have a great year. This year is going to have a solid year next -- I think memory is going to have a pretty decent year next year as well. I don't have all the puts and takes in it. But Blayne, when I look at the totality of it, both from a -- listen, we know -- and I would tell you, our visibility has probably never been better simply because lead times are pretty stretched out in the industry right now. And so the customer base needs to be giving us a reasonable visibility into where they're going next year. We know how big the clean space is in every location where there's a new fab. And so when we look at the tops down, meaning what does the market demand look like to us as well as our understanding of, okay, how big are the clean rooms, how many fabs are out there? What are the customers telling us? When we say it looks like a good year or strong year, we say it based on those 2 views of the customer base, frankly, in the market.

As there's been a lot of trends over the last couple of years, some good, some bad. One issue that everybody has had to face is just the supply constraints, and that's been an increasing headwind for you. Clearly, the semi market was navigating that maybe a little bit sooner. You have a footprint in Malaysia that had issues. I mean it's one thing after another I'm assuming. But can you maybe address what are the issues you're still facing? And kind of what's been the impact on the P&L from all this?

Yes. Let me walk through, Blayne, a little bit. I mean as we came through the year, business strengthened, right, things strengthened. I think that's a true statement across the whole industry, frankly. And so as we went through that, I think we did a pretty nice job in terms of managing our own internal capability, call it that, right, our own factory footprint, frankly. We built a lot of our tools in the West Coast of the U.S., both California and Oregon. We've expanded space there. We announced a new facility in Oregon actually where we do some of our wet deposition, are going to do more of our wet deposition. We were fortunate actually in that, we were building a new factory in Malaysia. We pulled that schedule in actually. I know everybody is kind of -- think about Malaysia as creative constraints and whatnot. Frankly, for us, it's been some measure of upside from our internal capability because we had a facility we were building that we pulled forward, frankly. We've got manufacturing locations in Korea and in Taiwan. We've bought out a joint venture partnership that we had in Taiwan so that we could accelerate what was going on there. So when I look at our whole factory footprint, I think we've done quite a nice job, and that's not where the constraints are, right? We've been able to hire people, train people, ramp those facilities. We're now working deeper in the supply chain. Our challenges have become -- and it's not necessarily just with our direct suppliers, it's several layers back in the supply chain as new constraints have popped up, right? And one thing I've been pointing out to people and some people kind of talk about this is a constraint that has emerged, in fact, we talked about it on the earnings call, our semiconductors. Semiconductors are not constraining equipment. The semi industry needs more equipment, and I need more semis to deliver more equipment, which is kind of crazy, right? And so that's an example of -- we're fairly deep in the supply chain. We've never directly procured semiconductors, but frankly, we are right now to help mitigate some of the challenges our suppliers are having. And there's other examples like that. So it's sort of -- it's game of whac-a-mole. You go alleviate one of the things that's creating challenges for you, and then you go to the next one and the next one. I would tell you, I feel like we're making incremental progress. I feel like we've done as good a job as anybody in the industry, frankly, in terms of managing this. But there's challenges that are still there. They probably are going to be there for a while. It's getting incrementally better with time. And it's part of the reason as I look into next year, there's some level of unmet demand. And maybe part of the reason I feel so good about telling, hey, it looks like a pretty strong year next year because we've got pretty good visibility as a result of the situation we're in.

My next fireside is actually with the Head of instrumentation of ADI. So I'll put it a word for you. Maybe just on the logic side, right? So one -- there's been kind of 2 pieces of this. You've seen this leading-edge heat up as Intel tries to push further. You clearly have seen end markets that are gobbling up the leading edge, be more robust when you think of things like the data center and the PC market, but then you have the trailing edge that has also had its own drivers, right? So maybe let's start with the trailing as first. Is there a way to think about how big that has become as a part of that $85 billion you're seeing?

Yes, I'll give you the color that I've shared in the past, and I haven't quantified it, Blayne. But we've been saying for a while, we view the -- I call it, the specialty part of the wafer fab equipment investment. Think about lagging lithographic nodes, if you will, maybe 28-nanometer and above. We believe actually that grows faster than overall WFE -- the overall growth in WFE. And we've been saying this for a while, and I still continue to say it today. So overall, that grows faster than WFE in total. And WFE in total is growing. Obviously, having a very strong year this year. When I look at it, Blayne, actually more wafers being added at the lagging edge, but because of the differential in capital intensity, more dollars at the leading edge is the way to think about it. I haven't put a hard number on it, but you can kind of think about it that way. And when you look at what is driving this stuff, it's a broad set of demand drivers, right? It's industrial, it's automotive, it's IoT, it's analogue, it's power, it's RF. It's fairly broad and that's a lot of what is driving the -- I call it again, specialty. And we talked about -- it's our reliant product line, where we go to market for those specialty nodes, 11 consecutive record quarters of revenue in that product line. And so that, I think, is a real statement of what's going on there. In fact, today, we issued a press release about a new tool, a deep silicon etch tool, we call it Syndion GP that is targeted at the most power process nodes, if you will. That will enable 200 and 300-millimeter flexibility in the tool. So that's kind of a big deal because we're starting to see some transitions from 200 to 300-millimeter wafers in that space. So yes, that's an important part of the business, Blayne.

I think you already kind of answer this, but I just refine the point. I mean, the increase in spending at the leading edge, continued process migrations on memory, all that should just trickle-down equipment to the trailing edge. But do you think when you take -- look at the whole market in some, you still need to see this trailing edge spend for incremental capacity continue.

Yes, I do, yes. Just -- there's a broad -- and like I said, the demand drivers are very broad. It's a lot of different applications with broad geographic footprint, actually. If you look at some of the investments here, it's in Europe, Japan, China. It's where some of the older pattern in the United States, actually a lot of stuff on the United States, but it's a broad set of customers playing with a fairly broad set of demand drivers.

And then when you look at the leading edge, I mean, you've seen announcements and rumors of all 3 leading-edge foundries building some sort of facilities in the U.S. I'm sure Europe will get involved at some point as well. We've been waiting probably longer than everybody would expect to get the actual funding here. But -- how do you think about that? I mean, Intel has said, "Hey, we're going to have a base level of CapEx, but if we get more, it's a net CapEx and we'll spend. Do you think there is even room if the government gets involved for that leading edge to even be a little bit better?

Yes, potentially, right? And a couple of things I'd like to point out to people as government helping out with some of the stuff doesn't necessarily drive new demand. You've got to think about end demand at the end of the day, right? Now if it gets a little bit more affordable to make these investments, there's an elasticity component that actually might enable incremental demand, but don't lose sight of the fact that really what matters most of all is in demand. But -- so yes, it's -- you're probably going to see -- not probably, you are going to see a broadening out of the geographic footprint of leading-edge foundry and logic capability. Clearly, you've seen a lot of announcements, yes, in the United States for sure, Europe as well. Maybe even a little bit in Japan, right? All those that may or may not be leading edge stuff, there might be more specialty targeted. But yes, you're starting to see it broadening out. And so it'll be less disciplined when you think about that. And the other thing, there is incremental equipment needed if you're running multiple fab locations just from a redundancy standpoint. So there's a little bit of upside from that aspect of it as well, Blayne.

And how do you think of Lam's ability to capture greater percentage of that wallet. Obviously, you have a very dominant franchise and memory that we'll get to, I'm not going to penalize you for being so doing so well there. But clearly, I think there's more that you can do on the foundry logic side. Can you talk about that either from a product or a technology or a customer perspective?

Yes. I mean the good thing about all the businesses we're in etch and deposition is at every single subsequent process node, our market gets larger, right? And in foundry, as you walk from 14 to 10 to 7 to 5 to 3 and beyond, our market opportunity gets bigger. So you're seeing aspects of that right now. In fact, last quarter was an all-time record level of business for us in the logic space. I think 2 or 3 quarters -- 3 quarters ago, I said the same thing about foundry. You're seeing a continued nice trajectory there. There's that aspect of the addressable market grows every node as well as -- we're doing well in market share. But you're right, we have an amazing franchise on the memory side. But don't we set the fact that actually we've got a great trajectory in foundry and logic right now as well. And frankly, you need all aspects of foundry, logic, memory, storage, system architecture. So we're focused on all of it.

Maybe moving to NAND. I wanted to ask you, I mean, the market has been waiting for an oversupply and some sort of negative pricing for who knows how long, almost a year. It doesn't seem like we're getting that big correction. Spending came down a little but really not that much. There's no huge fall off. In fact, people are getting a little bit more incrementally positive about next year. You said memory would grow. How has NAND found the bottom, is it insight? And is this a proof point that it's maybe a little bit more rational versus these big boom bust everybody is waiting for?

Yes, Blayne, I'll talk a little bit about our positioning in NAND because I love it. It's our strongest position. But yes, I see an industry actually -- again, I alluded to this a little bit that it won't grow every year. But I see rational behavior by industry participants, right, everybody managing profitability, everybody managing their business, everybody focused on medium and long term. Where are things -- I mean, frankly, those are the nature of the conversations we're having with our customers right now is what does next year look like, not necessarily what is pricing doing this quarter in the spot market. People are focused more longer term, what does the growth look like into next year and the year after because frankly, the investments that are going to be made next year, likely don't generate much output next year. It gets with an eye toward where is the year after head and that's the nature of how this stuff works. And I know everybody knows that, but I'll just remind you of it. And those are the conversations we're having right now, right? What's the road map look like for next year? Where is it going? And I will just talk a little bit about NAND. We fundamentally enabled the NAND process well in terms of depositing those alternating some layers, some of the most complicated etches down through that stack. We own that critical process step, the metallization and tungsten metallization. We pretty much own all of that. So we're in an amazingly strong position in terms of NAND enablement. And frankly, you heard Tim, our CEO, talked in the earnings call about the fact that cumulatively, when you look at the 3 most critical steps through the NAND process flows, we have processed cumulatively 45 million more wafers than the next nearest competitor. That just enables such a differentiation and such learning that enables us to continue to evolve forward in metallization, high asset ratio ex capability. The incumbency factor is so important, and we're fortunate that we have that incumbency in NAND. And as that stack gets bigger, the industry just needs more of the things that we supply to it because we enable that growth in [indiscernible].

Maybe, the second question is just on the calculus that people use to try to figure out, all right, land spend this much, it must equate to this amount of bids or wafers. You laid it out a while back, you haven't updated it. But without the exact math, are people getting us a bit wrong? Clearly, I think you just said that you're not going to get any more wafers, it's more technology transitions. But have some of these factors changed since you last laid this out and is it -- the strong spending that we saw this year, is it more just a function of the de novo migrations and the increasing capital intensity?

Yes, it is, Blayne. And yes, we gave a number 5 years ago, I'm not even going to repeat it because it's a sale number. And frankly, it's higher today than it was 5 years ago. And every process node as you go from a 9x layer to 128 to 176 or 1.9x or what have you, it gets more expensive. And I think it's intuitive when you think about it you're depositing more films, but the etches take longer, the metallization to the positive take somewhat longer. And so it is more expensive is the point. And I don't know if people are getting it right or wrong. But I have recently been pointing out more -- forget about that number we gave you 5, 6 years ago, it's a sale number. it's getting more expensive. And I think everybody in the industry generally understands that.

What type of visibility do you have? You talked about spending next year for maybe kind of output the following year, but your lead times are quite long at this point, too. So what kind of visibility you have on the NAND side?

We have pretty good visibility in every segment of the market, frankly, because lead times across all of the customer base aren't really all that dissimilar one to the next, whether it's a NAND customer or a foundry customer, a logic customer or DRAM customer. Lead times are stretched for all aspects the customer base, frankly. And so I would tell you, visibility is as good as it's ever been for that reason simply because the industry is constrained -- if we don't have really good visibility, it's hard to plan the supply chain, manufacturing and so forth. So visibility is as good as I've ever seen it.

I want to ask on the DRAM side because all the fears about NAND, DRAM was actually the one that was a bit more volatile. There wasn't spending, and there is spending. And now some people are talking about next year being down. What's your perspective on just supply/demand in DRAM? And do you agree with this thought that maybe there needs to be a little bit slower spending next year?

Like I said, we haven't quantified next year, but all segments of the customer base looks strong, I'll describe it that way. And frankly, in DRAM, like elsewhere, I see fairly methodical, rational planning of everyone's business. Obviously, in DRAM, it's a very consolidated set of customers, even more so than NAND, and I think everybody knows that. And so I see a pretty rational behavior. I see people talking with us about road map plans into next year, again, with an eye towards what does output need to be the year after or late next year. And yes, there's always an ebb and flow around where is pricing and near-term profitability, but the nature of the conversations, truthfully are much longer term oriented when the customers are thinking about the road maps, right, and especially perhaps because like we've been talking about, lead times are fairly stretched out.

I wanted to ask you on China. That was a big part of the increase in WFE. I think there's some view that maybe that would be a little bit down. What -- I guess everybody is concerned about some big domestic blood of spending that would somehow reversed. I think you consistently said it's more broad than that. But why would it be down next year?

I haven't said that it will be. In fact, I'm -- again, I haven't provided color into next year. I think the China customer base is going to be strong next year, too, just like the rest of the market. If you allow me, I'll spend a couple of minutes talking about our business in China because sometimes people confuse or don't remember everything. 35%, 36% of our revenue over the last couple of quarters has been to China. It's a big region for us. What I'd like to remind people all the time is understand though that that's fairly evenly split between multinational customers with fabs in China and local Chinese customers. So it's not all just indigenous Chinese customers. So that's one thing. There's a fairly broad balance. And there's a lot of fabs, a lot of multinational customers with fabs in China. So I don't really say that. Then within China itself, you've got maybe one big NAND customer, one big DRAM customer. And then a very broad set, I would describe, of foundry specialty technology investments occurring. And I think sometimes people don't see the broad-based nature of that because a lot of -- some of these customers are new emerging customers, although some have been around for a long time. But I feel really good about what's going on in China. I haven't said next year, allow us to get through the end of the next quarter and do all the normal analytics, and we'll let you know what it looks like. But -- we've said in 2020, local China WP was about $10 billion. And then I've said in '21, it's grown. We haven't given you a hard number, and I haven't said anything about '22, but I have no reason to believe that you're not going to see continued strength in China. In fact, I'd be surprised if it was anything but that.

You're always so greater time around the business, but I do want to give you a couple of CFO questions as well. Gross margins have been 46%, I think, for 4 or 5 quarters in a row, while revenues up 30%, 40%. I think we talked about some of the supply issues, but is there a way to kind of quantify it a little bit better in terms of what are you facing still in terms of headwinds? And then how do you think about the mix. I guess you don't want to say what it will be next year, but just conceptually foundry logic memory, trailing is leading edge, or what's the mix component to think about as well?

When you think about the profitability of the company and the business model, there's not really a differential based on the segment of the customers, meaning foundry logic versus memory, that doesn't drive differential profitability necessarily. But you're asking the right question, Blayne. There's some level of headwinds related to COVID in the supply chain constraints that we've been talking about. And yes, we've been at 46%. What I've recently been pointing out to people is, okay, everybody wants to know how big is, and I haven't said. But what I have said is, if you go back to that financial model that we gave actually right before the world locked down in COVID with people remember visiting us at our Analyst Day, it was right before the world shut down. I provided a financial model targeted at 23, 24 different WFE levels. And obviously, WFE right now is decently higher than those scenarios. But be that as it may, what was implied in the financial models there to get to kind of that 33-ish percent operating income was 47.5%, maybe approaching 48% gross margin, and we're at 46% today. And at these revenue levels, you might have expected us to be closer to the financial model. And obviously, part of why we're not part of these headwinds that we're talking about. The biggest first lane and continues to be a straight logistics spending. The customer base for us, by and large, is in Asia. Our biggest factories are in the United States in California and Oregon. We have inbound freight bringing some of the supply chain output to the factories and then going back. So there's transpacific shipment in and out, that's occurring -- that is the biggest headwind we're dealing with right now. We're working on it. It will get better over time, I think, I believe. And so that's what's going on, frankly. And on top of that, ramping a new factory location, while this isn't a big fixed cost business, there is some level of fixed cost. And so Malaysia eventually becomes a nice competitive advantage and a tailwind to profitability. But right now, it's a little bit of a headwind, Blayne. So those 2 things kind of combine what we're working our way through.

And then maybe just finally, I wanted to ask you on capital returns. I mean, the semi cap industry has great profile versus kind of broad-based semis, maybe potentially superior growth, capital returns are also a similar range. When you look at your dividend, it does kind of trail some of those broader-based semi names a bit. I know you've been committed to the buyback. I think M&A may or may not be off the table. Can you talk about how you rank these 3? And is there room to kind of bring that dividend more up to kind of where the broad-based semi group is?

Yes. Let me talk a little bit about it. The one thing -- and we're out of time, so I can't really talk about the installed base business or what we call the customer core business group is a fantastic part of the business that is extremely cash generative. And as that grows, our confidence in cash generation continues to get stronger and stronger and stronger. We're out of time, so I can't really talk about that, Blayne, but don't forget about that. That is a huge contributor to our commitment to return 75% to 100% of free cash flow. Since we initiated the dividend, we've grown it every year. And I know people like to see the dividend growing every year, and that's our plan. We intend as the business grows to grow the dividend every year and then supplement it with the share buyback. I think large-scale M&A is largely in the past for this industry. It doesn't mean there might not be small tuck-ins here or there, but anything of any stainless scope, I think, is largely behind the industry. And so I think you're looking at all of us returning more and more cash. And certainly from land, that's our intention. Our commitment is at least 75%. And the last couple of years, actually, we've returned more than 100%. I've actually pulled cash down on the balance sheet over the last several years. So yes, business continues to be very cash generative.

Perfect. With that, we'll leave it there. Thanks for your time, Doug.

Yes. Blayne, good to see you. Take care everyone.