Lam Research Corporation (LRCX) Earnings Call Transcript
June 9, 2021
Earnings Call Speaker Segments
Well, good morning, everyone. This is C.J. Muse, semiconductors, semiconductor equipment analyst with Evercore ISI. Welcome to day number 3 of our Inaugural TMT Conference. Very pleased to host Lam Research. And this morning, we have Doug Bettinger, Executive Vice President and Chief Financial Officer. Doug, welcome.
C.J., great to be with you today. Great to -- are we kicking things off here? Am I your first fireside chat this morning?
This is the first one this morning. Yes.
Awesome. So hopefully, we get the day going great. Listen, I know I've got the safe harbor slide up.
Great. Thanks, Doug. So I figured start off on the near-term environment. And I guess, as a starting point, where are your lead times today versus normal? And how are you thinking about capacity coming online from Malaysia in the back half of the year?
Sure. Yes, awesome place to start. First, I'd say, C.J., they are extended. We haven't quantified how far out they're extended, but they are beyond what a normal lead time for us would be. And we're working and doing our best to mitigate that. And you hit on the right thing. Basically, I would describe, every manufacturing location we have in the world, we are working to do more to extend our output capability. And you hit on an important one. We were fortunate in that we're building a new factory in Malaysia. That ultimately will be the biggest factory in our network. And one of the things we're trying to do to increase our output capability is to pull that schedule forward. And we're having some level of success, although I would tell you, COVID is making it challenging. Even though COVID is getting a whole lot better in the United States, it continues to be challenging in the rest of the world, especially in the Asia regions. But be that as it may, we're doing our best to pull that schedule forward and having some level of success. C.J., I talked on the earnings call about a couple of other things as well. We're extending our output capability in Korea. We have manufacturing capability in Korea. We took full ownership of a joint venture in Taiwan, where we manufacture our refurbished or increasingly new equipment, but our Reliant product line, which is targeted at trailing-edge process node. So we took ownership of that. So that -- the complete ownership of that so that we could more directly control the ramp capability there. We're doing similar type things in our factory in Livermore as well as our factory in Tualatin, Oregon. And so when you look at that, we're increasing our output capability pretty much everywhere that we build tools and working with our supply chain partners to do the same thing so that they can support what we're trying to do.
Excellent. Another question on the near term. For the first time, we're hearing multiyear elevated CapEx projections from leading chip makers. And just curious, how was that changing the environment as a supplier to those players? Is your soft backlog visibility improving? Are you seeing customers looking to get in line sooner for projects that may not take equipment out 1 or 2 years? How are you seeing the current environment perhaps changing structurally the business, if at all?
Yes. C.J., maybe it's changing in a little bit. But to be perfectly honest about it, our customers have always had multiyear roadmaps and multiyear plans and multiyear locations where they're building new fabs in different geographies. And as part of that communication to us, they're always telling us, this is where I'm going. This is what I'm doing here. I'm building this new fab in Arizona or in China or in Korea or I'm extending. That's not unique to the environment we're in right now necessarily. We always have those conversations. Now the fact that supply is a little bit tight, perhaps those conversations are happening with a little bit more urgency, if you will. But I don't know that it's fundamentally changed the nature of the long-term conversations that we're having with customers. You know how this business works. To the extent that the industry is working on a new process node or a future architecture, there always has to be a robust conversation with companies like Lam Research to enable new capability and road maps that don't exist yet. That's just the nature of how this industry and how this business works. And even though business is very strong right now, I don't know what I would tell you. It's fundamentally changed the nature of those conversations. They always are happening.
That's helpful. And I guess maybe from an end demand perspective. How are you thinking about NAND, DRAM, foundry/logic trending into the second half relative to what you're seeing in the first half and anything kind of noteworthy change-wise there?
Yes. No real change from the earnings call. Maybe I'll just give you a little, like, chronology of how the year so far has unfolded for us. We came into the year expecting a pretty robust investment environment, but a front half-weighted year was how we saw things as the year began. And that's how we described it. And we had expected WFE to be high 60s approaching 70 at that point, right? Fast forward 90 days, basically, what happens, C.J., is the second half strengthened. It strengthened in every end market, in NAND as the industry begins or continues investment on a 9x layer and increasingly a 1, 2x, 128-layer device in the second half of the year. In DRAM, similarly, moving down the process nodes, 1z moving to maybe even 1-alpha. And then in foundry/logic, you have -- actually, spending both at the lagging edge as well as at the leading edge. The majority of the spending occurring at the leading edge, but a lot of investment occurring at the lagging edge. So when you think about all of how things unfolded, everything got a little bit stronger in the second half, such that we now see WFE trending above 75. And it pretty much filled in, in the second half of the year is what we saw as things have unfolded thus far.
Great. So I wanted to, I guess, drill down on your NAND business in the NAND market. You guys have talked historically about kind of $14 billion, $15 billion normalized WFE update and it sounds like this year is tracking maybe around $18 billion or so. And I guess, just curious, does that $14 billion, $15 billion still hold? Or are we seeing an uptick in cap intensity as we kind of see an end to single stack? And then secondly, as part of that spend this year, there's a good chunk from YMTC. And so I think if you actually took that out, we probably would be closer to that normalized level. So do you agree with kind of all of the above? And how should we think about the sustainability of kind of current spending on this bucket?
Yes. C.J., I agree with everything you said. I guess what I would say is we initially gave that number, I don't know, it's probably 3 or 4 years ago, I forgot exactly when it was, call it, 3 years ago, to try to help people kind of understand what they should expect. But you're right. When you have an emerging customer that may not yet be generating output, that might create some level of incremental investment as they ramp up. And then I think, as you know, as the stack is taller, the level of investment increases over the years. So we kind of gave round numbers but maybe a little bit stale at this point. And some of the things you're pointing out are absolutely right. And maybe just a little bit of color. What I see going on this year is a combination of some new wafers coming in as well as, and you see this every year, a lot of node conversions, right, moving from, I don't know, 48 to 64 to 9x structure. And that's kind of what is required right now. The level of investment we're seeing right now is a combination of new wafers as well as node conversions occurring, just to give a little color on what's going on. But the stuff you point to, C.J., it's right. There's probably some level of incremental investment that's not generating output that bumps that number up a little bit.
Excellent. I think one aspect to the NAND business that I don't think is fully appreciated on the market is that, at least, maybe I'm wrong, but I think of Lam as being agnostic to whether it's new wafers or increased layer stacks. And in fact, I would expect your share of wallet to continue to increase given your kind of leadership in the high aspect ratio etch world. So if I look back to 2020, your share of wallet for NAND based on my estimates for WFE, around 23%. And so how are you thinking about your share of wallet going forward in the NAND market?
It will continue to grow as the stack gets bigger. I think we have very strong positions in the alternating film stack deposition. We have very, very strong positions in the channel hole or memory hole etch. We have very, very strong positions in the metallization, the tungsten deposition in the structure. And so when you think about as that stack grows, the need to grow those things that I just referred to is really what happens, especially if the customer base is making conversions, right, where you're just having to grow the things that are depositing our structure. That's everything we do. And so when you think about that, it grows, right? Our share of wallet will incrementally grow as that stack gets taller because we're doing all of the critical applications in there. So that's the way to think about it. In a year where new wafers are coming in, the customer base is buying everything. In a year where a lot of conversions are occurring, they're really just buying those critical applications or a lot higher percentage of those critical applications, which is pretty much owned by Lam Research.
Got you. And so when you think about '22, and obviously not perfect visibility today, but is that a year of layer counts? Or do you also anticipate new wafers?
Yes, everybody wants to know what next year looks like. And I'm not quite ready to tell you yet because we haven't done all the work on it. But like I said, what I see this year is a blend of both. What you see every year is a good amount of conversions occurring because that's an economically advantageous way for the customer base to get bit supply growth at a lower cost. I don't know what next year is going to look like yet, C.J. You got to give us a little bit more time. I do think it's going to be a good year for the industry. To what magnitude, too soon for me to tell you.
Sure. How about your new Sense.i platform? You've indicated you've shipped to 3 NAND players, huge benefit from the smaller footprint. I believe you ramp this plan for the back half of the year. Can you kind of walk through the moving parts there and how we should think about the implications to your share and to your margins?
Yes. Let me talk a little bit about Sense.i. And for the people that know the company, you've heard the story before, but we probably have some that you haven't heard yet. So we are super excited about our new etch platform. It's the first bottoms-up redesign of the platform since the conversion to 300-millimeter, which was over 20 years ago. So we're very excited about kind of blank sheet of paper -- the leader in etch redesigning RF platform, right? Customers are excited about this. The customer pull is very strong. And so we're working really hard on ramping things, and it's ramping quite well. We have tools or soon will have tools at all the large memory customers. We announced Vantex, which is the dielectric etch chamber. So that was an incremental announcement from us, not all that long ago. We first announced since I had our Investor Day in March of last year, just to kind of frame it for everybody. And so what's unique about Sense.i? Why are we excited about Sense.i? I really think about 3 things with Sense.i coming on. And like I said, it's important to understand, first, blank sheet of paper redesign from us of the platform. So there's a lot of things I think our architects have wanted to do over the years, but we're constrained by the platform. We're not constrained. So we lifted those constraints. First, technology. This is a redesign of the chamber. And it will enable the road map for the company for the next decade or longer from a technical capability standpoint. When you think about all of these high aspect ratio etches that are in process flows today and increasingly will continue to grow over time, this takes us forward for the next decade from capabilities that are going to be needed there. So that's pretty exciting, just the technical road map piece of it. Second, it's the most intelligent tool we've ever brought to market in terms of its ability to -- I almost think of it as be self-aware about what's occurring in the tool itself. And to -- almost to a certain aspects of self maintenance, manage things like edge -- wafer edge yield, which is a challenge for the industry. All of the intelligence -- a lot of cameras, a lot of sensors, a lot of ability to self maintain, so that's pretty exciting. The equipment intelligence aspect of it. And then third is the footprint efficiency. It is the most efficient etch tool or it will be as it ramps, the most efficient output per square meter in the fab. And so that enables a cost road map for the customer base. And so when customers look at this, they're pretty excited and they wanted as soon as we can get it to them. So we're excited about it. We're executing the road map. In fact, we're trying to accelerate the road map this year to a certain extent. And the customer, you have -- I always know in this business, you have something unique when you see the customer pulling and wanting it sooner sometimes than you're able to deliver it, and that's what we're seeing with Sense.i. So we're pretty excited about it. It's a big part, and you're probably going to ask me about this, as we talk to you about how are you going to gain share of wallet market share. Sense.i is a key aspect of how we're going to gain continued share from where we are today to where we aspire to be because of those 3 aspects of capability, productivity, equipment intelligence and technical differentiation.
Excellent. So I guess 2 follow-on questions there. Is there potentially a replacement cycle here with this new tool given the productivity? Or no? And then secondly, how should we think of adoption? Is it going to be primarily NAND first? Or is it going to be broad-based?
We have chosen to go to NAND first, but it will be broad-based over time. It eventually will replace the 2300 platform that we have out there over time. In the 2300 tools, we'll get repurposed to other applications in the fab. But it's got a road map. It's got a phase cut in over time. We've chosen to ramp it with the NAND customer base first, but it will eventually go to DRAM, foundry/logic.
Excellent. All right. So I guess maybe moving on to DRAM. I think your share of wallet, maybe 2/3 of what it is for NAND. Curious if you see changes there into the future, particularly as we think about 3D and DRAM at some point in the road map. Would love to hear your thoughts there.
Yes. No. Yes, DRAM, strong. I mean, this year, maybe the near-term comment, again, ramping spending in the second half to generate output. Really, when I step back and look at where are things trending, 3D DRAM [indiscernible] customers kind of on their own schedule, all their 3D [ mark ] far off from one another, ramps in the latter part of the decade, maybe beginning in 2025. And we're excited about where this is going. Lam does 3D well, right? If you think about 3D device architectures, you saw it in NAND. We've already talked about that. You're going to see it in DRAM, where the structure gets built up architecturally. And then that creates different etch challenges. In the case of DRAM, there's conductor etches. There's emerging selective etch becoming more and more important. And that's what we do well. So when you look at where that goes, we're excited about 3D DRAM coming in, but it's still fairly far away. There's a couple more planar nodes in front of us. And our footprint there is very strong. That's about patterning, enabling a road map. EUV is coming in, in a couple of layers here. We're working to help enable EUV in DRAM and foundry and logic with our dry resist capability. I know you'll ask me about that a little bit later on. But yes, I like where DRAM is going. I feel good about our share gain objectives there. We're the leader in etch in DRAM like we are elsewhere. So yes, it's doing quite good.
Excellent. I guess moving to foundry/logic. I think getting our arms around precisely what's driving kind of a share gain opportunity [indiscernible] is a little bit more challenging, but would love to hear your thoughts there as the industry is migrating kind of 5-, 3- and 2-nanometer. And then more specifically, would love to hear from you around gate-all-around. I imagine that's going to be a fantastic opportunity for Lam. Again, as we move more 3D transistor structures, would love to hear about your positioning there and how you're thinking about perhaps timing of ramp for you there.
Yes. It's funny, C.J. Everybody thinks of Lam as the memory-enabling equipment supplier. But quite frankly, over the last several quarters, we have had all-time record levels of shipments into the foundry/logic space as well, right? So there's nice momentum there. And it's complexity of the road map. It's -- patterning is in there. It's 3D device architecture beginning to show itself. Actually, gate-all-around is going to be good in terms of what SAM per wafer, if you think about etch intensity and deposition intensity. That's again with FinFET, though, to be honest with you, right? I view this as an evolution of architectures that are more etch and deposition intensive. Gate-all-around is going to be good. It's going to have incremental etch needs, again, in a selective etch type capability. But there's different complexities in there that increase the intensity of our share of spend. So yes, we have consistently said every node in foundry/logic as you walk down from 14 to 10 to 7 to 5 to 3 and beyond, our share -- or our SAM gets bigger per wafer and gate-all-around is part of what drives that at 3 and beyond. But there's incremental things, and it's complex. It isn't one thing. It's multiple things in there that drive that trend. And yes, you're seeing it in our numbers already, and I'm excited about where it's going.
Excellent. So I guess you spoke to earlier the new dry resist tool. Was hoping to get an update there in terms of your collaboration with ASML and imec and your engagement with your top 4 customers.
Yes. So I talked earlier about the pull for Sense.i. We're seeing the same thing for our dry resist tool. And this is a new space for us, as you know, right? This is a brand-new share of SAM. It's an area that we don't play in today that we have aspirations to grow over time. So we have or very soon will have an evaluation tool with every customer that's using EUV. So everybody's looking at it. And again, you don't get that space, you don't get R&D space if a customer doesn't see something that they think is attractive relative to where they want to go. We're solving challenges that they see in the future. And that's very much the case with our dry resist tool. And yes, we're closely collaborating with both ASML and imec to help fine-tune this. And now as we put these eval units with customers, collaborating with customers, which is even more important to fine-tune things, to work kinks out and to deal with technical challenges that we see. But I would tell you, C.J., I feel better about where we are today than I did a year ago. I see customers pulling on this. I see us figuring things out that we knew we needed to figure out, and momentum is very positive. And I still am very convicted in $1.5 billion over a 5- year time frame. I'm hopeful that it could be higher, but too soon for us to change any numbers at this point. But the customers are liking what they see. The collaboration is very strong. The partnership with ASML is very strong and imec. And yes, I'm hopeful a year from now, I'm going to be even more convicted when I talk to you about it. But momentum is good.
Excellent. Do you have a sense of the productivity benefit of going dry versus wet?
I don't have numbers for you, C.J., but it is a decent amount or you wouldn't see the customers as excited about it. One of the things that maybe is somewhat underappreciated is like the benefit from a chemical consumption standpoint versus a wet solution, right? If you think about ESG and whatnot, it's a whole lot environmentally -- it's better, right? You're consuming a lot fewer chemicals in the process here, which is another aspect that increasingly people are asking me about.
Yes. I would think ASML will be giving you dollars to get the shop because it's going to improve the throughput of their tools, right?
It absolutely is. And that's why we're collaborating closely with them. It helps them and it helps us and it helps customers more importantly.
Yes. I guess we'll move on to CSBG.
My favorite part of the business model in a lot of ways.
I know. I know. So I wanted to hit on Reliant. And so I've got to use VLSI for this. So if I take your reported service fares category, and then I kind of take out what VLSI has estimated, it suggests Reliant is about a $1.7 billion business. That seems a little high to me, but I was hoping you could give us a sense of the relative size of the business and how we should think about sustainability of lagging edge spend beyond 2021.
Yes. Maybe C.J., if you like, give me a couple of minutes. Let me paint a picture of what that business is for people that may not be as familiar with the business as you are. CSBG is the customer support business unit. And we -- it's really the installed base, right, how we make money with the units -- the tools that are sitting in the field. And there's 4 things in there. One, you've asked about our Reliant product line. This is our legacy node equipment sales, right, which historically has been almost entirely refurbished equipment, right, where we will buy a tool back, refurbish it and resell it. We're also selling new tools into this space now, right? And that's part of the reason why we took ownership of the joint venture in Taiwan so that we could add more direct control over that road map. So Reliant is important aspect of that. And when we look at the lagging edge investment in foundry, we see it growing faster than overall WFE, maybe by 2 to 3x. So that's what drives that business. So whether your numbers are precisely correct, C.J., I just disclosed CSBG in total, not the component. So -- but Reliant is an important piece of it. The next important piece are parts. That aspect of the business is very strong right now. And you should think of that as kind of ebbing and flowing with industry utilization. When utilization is high, consumption of spare parts is a little bit higher. But what is more important are just chambers in the field. And we are shipping a lot of chamber -- incremental chambers in the field. So that's part of the tailwind there. So spare is second. Third is pure service, right? Our tools need periodic maintenance, and all tools in the industry do, but it's an important aspect of what we do. The service intensity is pretty high with our tools, especially etch tools. And so that's the third component of CSBG. And the fourth is upgrades, right, productivity, technology upgrades, where the customer gets incremental capability from an older tool to a new tool by us changing some component to allow or upgrading it essentially. So when you put those 4 things together, I don't know, it's about 1/3 of the business right now. Yes. I've always said, I can't envision a year where this business doesn't grow every year because chambers grow every year. That's an important thing for people to understand is our tools will run for decades. They almost never, like, go away. And so if you think about a strong WFE year like we're seeing this year, the tools in the field will grow a good amount this year. We'll -- we give you a chamber count every year, so you can see it. We usually do that at the end of the year in the December earnings. And when you have WFE as strong this year as it is, chamber comp will grow a good amount this year. And so that is a tailwind for what CSBG will look like next year. So anyway, I've rambled on a little bit, C.J., but I wanted to make sure people had a comprehensive picture of what is the customer support business group at Lam? That's what it is. And it's a recurring business. As long as the fabs are running, a lot of this just keeps going, right? It's an annuity in many ways.
Yes. For sure. So you have far exceeded your target model in terms of the growth for this business. And I guess, maybe it would be helpful to understand what has led to that and whether, I guess, we can think it's sustainable or not. And so I'm curious -- I imagine Reliant has surprised some, but also curious to hear how you're seeing kind of revenue per tool per chamber.
Revenue per tool trending as expected, maybe a little bit ahead. What has really driven the strength of this business is -- I mean fabs are running at a very high utilization, which drives spare parts and service consumption. That investment in lagging-edge nodes when you think about -- I mean, just look at all the headlines of constraint in industrial and automotive. I mean semis are constrained at the lagging edge that drives the Reliant business very strongly. It also drives the upgrade business. So when I look at our 4 cylinder engine here, we're hitting on all 4 cylinders right now. So it's been a great year for CSBG, will be a great year for CSBG. And I feel very convicted about where we're going from a growth standpoint because of how many chambers we're going to ship this year.
Excellent. On the capital return side of things, you've historically stated that the dividend will be supported by CSBG. And I guess, given how strong that has been, do you still view that as tightly connected? And what does that speak to over time as to what kind of growth in the dividend rate we should see?
When I think about capital returns, C.J., what we've said is our plans are to return 75% to 100% of free cash flow back to the equity holders, and that's very much still the plan. The constituent components of that are the dividend and the buyback. And I've told people, you should expect us to grow the dividend every year. And yes, I have said the sustainability of the growth in dividend is somewhat coupled with CSBG. But the cash that -- by the way, CSBG is a very cash-generative business. It doesn't require a huge investment. It's very recurring in a lot of ways. It's why I say it's my favorite part of the business model in a lot of ways because of that cash generation. As a CFO, I love cash, right? But yes, CSBG comfortably supports the level of dividend. We plan to grow the dividend every year. And then to supplement it in terms of that 75% to 100% of free cash flow with the buyback, which we've, I think, been pretty judicious with how we've gone at buying our shares back, and we intend to continue to do that.
Got you. We got a couple of minutes left. I guess, maybe we'll close on gross margins. Your target model, 48% to '23, '24 time frame. I was hoping you could kind of walk through the drivers of getting us there including, I guess, COVID-related costs today, but also the benefit from the new factory in Malaysia and anything else we should be thinking about.
Yes. C.J., the 2 things you referenced are really what I see is the things that will enable us to expand gross margin. That are fairly clear line of sight. So right now, when you look at what's going on, the biggest headwind related to COVID that we have -- it's freight logistics spending, right? We're flying inbound and outbound freight, big stuff, right, subsystems and completed tools. And just to kind of paint a picture, our biggest factories are in the United States, and most of our customers are in the Asia region. And so we're flying things, right? And freight lanes are pretty expensive right now because commercial aircraft aren't flying at the volume that they were before, right? And so there's an elevated freight logistics spending that will mitigate over time as freight lanes normalize. So I haven't said exactly how big that is, except that I have said it's a noticeable headwind to gross margin when I look at it. So that's one thing that will enable us to expand gross margin over time as when that normalizes and isn't a state of a headwind. Second, and you hit on it, C.J., is as we ramp the factory in Malaysia, and like I said, the Malaysia factory will eventually be the biggest factory in the network, it will have a lower cost structure. You'll have a little lower cost structure from the labor component. It will also have a better cost structure from a supply chain standpoint. We're bringing the supply chain along with us here. And also, if you think about my comment on freight logistics, Malaysia, obviously, is in Asia where the customers are. So we're going to be flying things at shorter distance. And so that -- when you look at all of that, from where we are today, C.J., to where we want to be in that financial model, those are the things that get us there, in addition to some things we're going to continue to do around pricing and new tools with better gross margin. We've talked about Sense.i and all of those things. When you put it all together, that's how we get to where we intend to be from a gross margin standpoint.
Excellent. Well, I think we've run out of time, but great to connect with you. Great to hear all these wonderful things. And hopefully, next time, we can do this in person.
Yes. I hope so. And thanks for having us, C.J., and I hope the conference goes swimmingly today.
Thank you. Appreciate it.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete Lam Research Corporation transcript - plus 251,000+ transcripts from 12,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.
Get an API key View API docs →For developers and AI pipelines
Programmatic access to Lam Research Corporation earnings transcripts and 251,000+ others is available through the
EarningsAPI REST API and the hosted MCP server.
Quarterly plans from $105 - full transcripts, speaker segments, full-text search,
and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.