Okay, we're gonna get started with our next session. I'm Tim Arcuri. I'm the semiconductor analyst here at UBS, and I'm very happy to have Lam Research with us, and very happy to have both Tim Archer and Doug Bettinger with us. Rare to get both Tim and Doug on the stage together, so thank you. This is great.
Thanks for having us, Tim.
I think this tells you it's a big conference.
A big con-
Took two of us.
Yes. Great. Well, maybe let's start, Tim. So, your competitor reported the week before last, one of your competitors. Did you hear anything that surprised you in what they said that is different than what you're saying, and sort of the, you know, outlook on the market? You know, DRAM pricing certainly is getting a bit stronger. You know, any changes in the market that you might have seen, you know, over the past a month or two?
Yeah. Well, first, thanks again for having us here, and I think you all caught the safe harbor statement that was up there, since you dived right in with the, "What are you seeing in the market ahead of us?" No, I think that, you know, to kind of go back about six weeks ago when we, you know, had our earnings call. We talked about, not yet our 2024 guidance. We'll do that at our, our January earnings call. But, you know, we talked about how the year was playing out, where, you know, we were seeing NAND being a little bit weaker, DRAM being stronger now, especially driven by HBM. We see Foundry, Logic, continuing to be pretty strong at the leading edge, maybe a little bit weaker at the mature node.
I think that's a pretty, pretty much the view we're hearing from all of the equipment companies today. We meet the same customers, and so, that is, not a lot's changed, and I think that with that view, you know, we see a really great opportunity for Lam ahead, just given in the short term, the markets where we're strongest happen to be the ones that we believe have fallen the furthest relative to the investment, and therefore should be, coming back stronger in the year.
Just in terms of WFE, I mean, I asked this question, I think, to every company so far.
It's become the Tim Arcuri question, by the way.
Yeah, I guess.
You know what I'm talking about. You're laughing.
So I look at the WFE this year and, you know, your share already. Even if I assume $80 billion, your WFE share is going to be as low as it's been in seven years because the mix is very unfavorable.
It's all mix.
Mix. Now, if WFE really, if I include litho, it appears that WFE probably is going to end up to be closer to 90 this year, which would then mean that your share is even lower. So, so can you just talk about sort of as you deconstruct what's gone on in your share, is it truly all mix?
It's all mix. There isn't any meaningful change in share at an application level. And frankly, litho, if you're buying a litho tool, because of how long the lead time is, we come afterwards. And so there's a little bit of that going on, and there's a little bit of, like, the sales into China being more than we realized, 'cause we're not participating into that part of the restricted marketplace. That's all that's going on, Tim.
Yeah, no, I think that, you know, when we look at it, when we look at the application basis, you know, where we're winning and, and the increased use of etch and dep, I mean, I think we feel really, really good, actually. You can't do much about mix, and in fact, as I said, from a short-term setup perspective, the fact that NAND WFE is down more than 75% a year clearly hurts us quite a lot. But we also know that, you know, many of those tools that were shipped in 2021 and 2022, when the, when the industry starts to come back, not only are you going to be adding, you know, sort of starting those up, but you're going to be doing tech conversions. Those tech conversions will come at higher layer counts.
For us in NAND, higher layer counts not only means more tools to build and etch the higher stack, but it also means additional tools to help with the added complexity of the stress on the wafer that's created by more layer counts, the additional steps needed to do multi-tier stacking. So I think we're very positive that, you know, actually our share will continue to grow with etch and dep intensity increases, given the complexity of 3D.
Yeah, and a couple of other kind of idiosyncratic drivers off the bottom. Whatever happens to WFE, would you say it's fair that there's been all this litho capacity that's been bought and has been installed, and so the greater capture rate, that you should have a greater capture rate of WFE moving forward because it's been so litho-heavy, and now you have to fill in around all the, you know, films tools around the litho. Is that fair?
Well, I think that's certainly true, and I think that was kind of Doug's comment about, you know, many of the-- the litho capacity gets put in place, and I wouldn't say that necessarily that they're done with that, but there is a, there is a need to fill in with all the other steps that are needed to, to fully utilize those tools. In addition to that, you know, we grew our installed base from 2019 until now by about 40%. And so there's also a very much, a very, high capture rate that we get from the conversions that are going to be coming on the memory side. And so, you know, I think there's an element of etch and deposition have kind of suffered through this, this big build-out of, of the litho capacity and the infrastructure.
Even as you looked at CapEx overall, I mean, a lot of the building and fab infrastructure, and I think our day is coming now as you look at Tech Conversions that are very Etch and deposition intensive, you look at new buys in memory, the mix starting to come back a bit in our favor. And I think that's where we have an opportunity to capture a much higher percentage of WFE going forward. In addition to that, you know, in the last few earnings calls, we've used the opportunity to highlight a few tech inflections that you're really not seeing right now because of where we are in this cycle. But whether it's the impact of Advanced Packaging or and which includes HBM, Gate-All-Around, the Dry EUV Resist product we have-
Backside power delivery. In the last four earnings calls, we've talked about those, and that's because they're all billion-dollar-plus SAM expansion/market share growth opportunities for Lam. That's a really unique place to be in. That's really partly because of the under exposure we've had in some of these marketplaces for a while, and we've made a conscious effort to invest in R&D to grow our portfolio into those, into those inflections.
I want to ask you about that. So, out of all those, is there any one of those that you're the most excited about? Or, and in aggregate, how should I think about how that expands your SAM?
Yeah, well, I mean, actually, we're excited about all of them because they're great opportunities for Lam. But I think that each one is a bit of a different stage of sort of maturity/adoption. I mean, we have a very strong position in advanced packaging at this point, based on products that we already have in the marketplace and are very well proven. Copper electroplating, for instance, is a market in which we've had a very strong position for a couple of decades now. As it continues to play an ever more important role in advanced packaging schemes, in HBM, you know, that's market share that's right there that we should that we fully expect to win as that, those applications grow. So I'd say those are coming.
Their products are in place, products are being ordered, and our product position is really strong. Gate-All-Around is a bit of a more unique application, and it pulls in new tools from Lam, like selective etch, ALD. And it's an opportunity for us to participate more fully in the Foundry, Logic build-out. You look at backside power, again, leveraging a lot of Lam's strength in interconnect, one of the kind of the core of the company for a long time, big opportunity for us. And, and-
And it's like, backside power is something where you can, if the incremental WFE is X, I mean, is the capture rate for backside power, for example, going to be like 50% for you or some huge-
I think in each of these, when we look at the products that are needed to enable these inflections, we actually believe that our capture rate should be at least as high as our current market share within Foundry, Logic, which, you know, in itself is good. Our challenge in Foundry, Logic hasn't been market share of our served market. It's that we need a broader product portfolio to serve more of the market.
Mm-hmm. Yeah.
As you know, I think we've been often identified as the 3D NAND company.
Yeah.
But actually, the reality is we've used the learning from that 3D NAND experience over the last 5, 6, 7 years, to really build out the capabilities to enable 3D inflections across all device types. And that's exactly all of those device types I just talked about. Gate-All-Around is a 3D inflection in the transistor itself. Backside Power is kind of a 3D build-out on the backside of the wafer. You need etch and deposition to do that. Advanced Packaging, it's 3D. It's things like chiplets. It's HBM, which is clearly 3D. And so I think that, you know, we're kind of coming into our own relative to the use of 3D as both a performance scaler and, frankly, a cost enabler.
I mean, often these, these new technologies are designed to help address the costs that would come if you try to just continue to scale in 2D itself. And so I think you're seeing just Lam's expertise in 3D playing out across a number of opportunities. And I think that's gonna, gonna yield market share and serve market expansion for us in, in, in the years to come.
So some people, there's the tailwinds, which we're talking about. There's the idiosyncratic pieces that you're now playing in, you know, pieces of the market that, you know, you're going to get a very big capture rate of. But then people say, "Well, NAND WFE is never going to go back to, you know, $17 billion-$18 billion, you know, plus, plus like it was." You know, NAND WFE, I mean, if you believe AMAT, they want you to believe that, you know, NAND will never go back to even thirteen, you know, $12 billion or, you know, $13 billion. Granted, it was, it's like $5 billion this year, but still.
How do you, how do you handicap when people look peak to peak, and they say, "Well, okay, your WFE share peaked out at, I think, 13.7, you know, 14.1%, when, you know, NAND was going crazy." Maybe what you lose in NAND, because the market never gets back to what it was, you're only barely even if making up for what you lose in, in the fact that the NAND market never gets back to where it used to be. How do you handicap the, the potential headwind of the NAND WFE not getting back to where it used to be, or do you not think that's the case?
Well, I think, I think we don't believe that it. I mean, first of all, never is a long time. I mean, we're, we're fundamental believers in the, the continued growth of the semiconductor industry itself. And in that trillion-dollar market, NAND will be higher, in our view, than, than much higher than the last peak. But for Lam, I think the most important thing is not just what is NAND WFE, but what is our capture of that WFE? And as time progresses, the installed base plays a larger and larger role in enabling bits going forward in NAND.
Mm-hmm.
And remember, Lam's capture rate in NAND tech conversions is a significantly higher percentage of WFE. So it can be true that WFE can be lower, but Lam's revenue actually continue to be higher, and that's simply because at every technology node, you need more etch and deposition equipment to help enable, you know, 200 layer stacks, 300 layer stacks, 400 layer stacks, 500 layer stacks. And so, but you're doing it more efficiently. Tech conversions, which are highly efficient for the customer, but also very high in capture rate for Lam.
Well, we've made it, 12 minutes into the presentation, and I haven't asked about China at all. So maybe I'll start asking about China.
By all means.
So, China this year, I you haven't given a China WFE number within your $80, but it seems to me like it's in the low $20s, probably for the year, you know, maybe it's $23. But in the back half of the year, and you can look at the export numbers, the, you know, Chinese government puts out these numbers, and you can see them, it's easily in excess of $30, or the, you know, run rate's in excess of $30. And if you believe the export control numbers there, it's even closer to $40 run rate. So the back half of the year has just gone crazy. AMAT was sort of suggesting that, well, that kind of comes off in the first half of next year. How do you see the trajectory of China?
Is it something that you're like: I don't know when it ends, I'm just going to ship what the customer wants. How do you kind of handicap that in terms of your model? Do you think that the market can sustain at $30+ billion run rate per year?
Yeah, I mean, I guess two things. One, the way we try to run the company is we have to respond to what the customer wants or doesn't want, and so we try to be highly variable in the way we run the company. But to your point of the second half of this year, China being quite strong, I mean, some of that was a factor of the DRAM component, where we weren't sure and needed clarification on whether we could ship tools. Those tended to pile up in the second half, and so I think there's some explanation of why there was so much in the second half of this year. We have talked about a digestion period that's kind of there. It's natural for any fabs.
Take a big project, you take a lot of equipment, your next job is actually to install and qualify those tools rather than, you know, take the next big slug of tools. So I think there will be a period of time here, maybe it's the first half of the year. We'll give more color on that in January. We talk about first half, second half of 2024. But I think, you know, in the end, we see some element of sustainability of spend in every region of the world, given this drive that came out of COVID for a level of regional self-sufficiency, and that's true in China. You see it in places like even Japan, that's investing. You see it, of course, in the US and Europe, and I think much of that is still to come.
Those fab, you know, many of those fabs are just being constructed, joint ventures just being set up. And so whether it's China or elsewhere, I think you'll see this sustained spend for a period of time until people can then assess, do we have the right amount to satisfy our supply chain resiliency goals or not? Once those are met, then we'll have to see where it goes from there. But I think this is a multi-year event of this build-out.
Is it also fair to say that you, or some of your peers that maybe have better DRAM WFE share than you. I mean, a significant piece of what's gone on here during the back half of the year in China, it's been DRAM. I mean, a big, big piece of it. And so you haven't captured as many of those dollars because your share in, you know, DRAM's not quite as high as, you know, some of your peers might be. So-
I mean, it's still pretty good in DRAM.
It's good still, but I mean, but, you know, DRAM in particular seems to be a big reason why you've seen this big increase here in the back half of the year.
Yeah, because of what Tim Archer just described, which is we couldn't ship in the first half or we were unclear, and then we could. Likely, that customer would have been spending through the year, more steady, if we knew that we could ship to them, and so that back piled in the second half of the year. I think that's the dynamic that's showing up in everybody's numbers right now.
Got it. And how concerned are you, there are these—there's, you know, the local Chinese film companies that, if you add up their revenue, you know, they report their revenue, and they're going to be probably 3% of WFE in 2024, maybe approaching 4% in 2025, if you believe what, you know, UBS is projecting anyway for these Chinese companies. And then you consider, well, okay, you know, China's, let's say, a third of WFE, and, you know, they don't serve the entire China market. They just serve Dep and Etch. So it, so they're, they're actually a pretty big percentage of their SAM in domestic China. So is that something that over time, do you see them more as competition?
Or, you know, we spent the first 10 minutes talking about technology, and in the areas where technology wins, you don't see them, or do you even see them pushing into those areas?
I think that, you know, we're very realistic that at the lower end of the market, I mean, they will make progress, and they have made progress, and more so given the more recent situation, uncertainty about what could or could not be shipped into China. But I think in the long run, we'll compete against them just as we do everywhere else, which is to advance the technology faster. And one of the things that's unique about companies like Lam is, and this is true when we compete against any... I'm not sure. I don't think that's me.
I don't know who that is.
You know, when we compete against companies in whether it's China or anywhere else, Lam has access to every leading company worldwide. Many of these local equipment suppliers access a much smaller customer base, and therefore, given how important those customer collaborations are to advancing the technology, and not just, you know, can you etch something or create a uniformity profile? It's often funny, the best uniformity profile that creates the best yield on a wafer definitely is never a flat profile. It has to be customized to the customer's needs based on all the other accumulated steps that they run.
And so you need a very tight feedback loop with leading-edge customers in order to make sure that your tools and your hardware and your processes are not only ready for the technology, but really for mass production over an extended period of time on tens, if not hundreds of thousands of wafers. And so I think that always will, the equipment suppliers that have more access to more customers and more engagement will innovate faster and therefore, in the long run, win. That's especially true, many of those items I talked about, whether it's Gate-All-Around or it's dry EUV going to High-NA EUV resists. I mean, these are. You need access to absolute leading-edge, best-in-world customers.
Maybe we can shift to, and I'll direct the conversation to Doug, just because I know, well, you're both so excited about CSBG, but you know, Doug, I mean, you know, both of you loved CSBG so much. But can you talk about your service business and maybe. It is a bit different than your peers, the components of it. So can you talk about the components of the service business and maybe why you're so excited about it? I think, you know, by our math, you know, service alone is $15 in earnings. I mean, so.
Anyway, yeah, that's working. Yeah, I love the CSBG business. In a lot of ways, it's my favorite part of the company's business model. And when you decompose what's in it. And by the way, Customer Support Business roups, four things: spare service, upgrades, and the Reliant product line, which goes into more of the specialty node investment area. It's a business that, before this year, I used to describe as grows every single year, and it should, in a normal environment, grow every single year. The challenge this year is utilization on the memory complex is so low that that's a headwind for spare service and upgrades, frankly. But it's part of the business model that just delivers cash generation and revenue over many, many, many years.
On average, after we sell an average tool, we generate more revenue after we sell it than when it first shows up in WFE. I think people don't realize that, and it's something we've been talking about for a few years. This business just keeps generating profit and free cash flow. It's a great part of the business model. Prior to the real strength we saw at those specialty node investments, spares, I used to describe as the biggest individual component. I now describe spares and Reliant as the two biggest components in CSBG, but it's a great part of the business model, Tim.
Great. And Doug, just to comment about you generating more revenue over the-
On average, not every tool.
That's over the-
Over the life
lifetime of the tool.
These tools, by the way, our tools run for decades. This isn't something that almost ever goes away. We have tools in the installed base that are 30+ years. They don't go away. They'll get refurbished sometimes, they'll get redeployed from one application to another, they'll get moved from one fab to another sometimes, but they don't go away. And so that's why, like I said before, this year, with the heavy utilization downtick, I would describe it as a business that grows every year, usually because chambers grow every year, and that's a number Tim always gives at the end of the year. We'll give it at the end of this calendar year, 'cause that's an important part of what our monetization opportunity is.
And I think, Tim, this is, you know, the, the number I gave earlier about since 2019, the installed base having grown about 40%. I mean, that's 40% more chambers in our, in our installed base generating spares and service opportunities. And it's why, you know, a lot of our focus is how to help the customers get maximum productivity from that installed base and at the same time create new revenue opportunities for Lam. And, you know, we talked on our last earnings call about our, our new collaborative robot, cobot.
I was gonna ask you about that.
You know, it's the. You know, we recognize that there are economics to this business for our customers where, you know, as tools, honestly, as you deliver technology, they're becoming more complex and more costly. And so we want to find ways to help make those tools more productive, such that it works for all of us. And, cobots is a way in which you can take kind of labor-intensive, costly service, that's also complex to do, meaning sometimes even a trained service engineer cannot get it right first time.
You replace that with a cobot that's specially programmed to do a certain maintenance task, and so far our experience is really great in terms of improving that first-time right percentage, which saves you, you know, parts that get lost when you have to go back and redo a certain maintenance, and also the lost productive time of that system from having it out of the manufacturing line.
So is that. I thought that was pretty cool, too. And is that something that's you see becoming ubiquitous within the, you know, service offering?
I think that what we would like to do is continue to work with customers to find those most labor-intensive, most challenging maintenance tasks, and then see if a cobot can help with that. Obviously, we've picked the first instance is our high aspect ratio etch process, which obviously has to be maintained at a really high level in order to ensure that we can do that 3D NAND etch. But I think over time, it will spread to more and more tool sets and applications.
If you can keep the tool running longer, you can, you can improve uptime. Could you share in some of those economics?
Well, I think that what we'd like to do is offer this as a service, as a replacement for the customer having to hire and train and deploy their own staff to do this. And so to a certain extent, in many ways, we've often been offering service contracts in what we call a results-based, on a results-based basis. And so obviously, if we have high confidence that the cobot can truly deliver higher uptime and therefore more wafers out per day, I believe we can capture some of that benefit back to Lam and let the customer keep some of it as well.
I had a question about the controls, the U.S. government controls for shipments into China. It's our understanding that in the U.S., there's what's called long-arm controls, so the U.S. government can reach all the way into the service, how the tool is being used in the field, and the burden is on you as the supplier to then have your service people, you know, have a report back and say: Well, the tool is being used for a process node that's an allowable process node. Whereas for the non-U.S. companies in Japan and in the,
Netherlands.
Netherlands, they don't have long-arm controls. So is that something where you, where you do have to have your service people, the, the burden's on you, where you have to report back to the government to basically prove that your tools are being used in a way that they're allowed to be used?
Yeah, we need a license for service and selling equipment, so it's one and the same in terms of what we're able to do, where we can service, what we provide. It's, it's no different between service and the equipment itself. Tim, you wanna add anything?
No, I think that, you know, first, firstly, we spend a lot of time and effort to ensure that we're compliant with all rules. And, you know, one of those is to work closely with the customers to understand the applications. And there are many different ways in which the U.S. government has controls to ensure companies remain compliant. And I think that, this is something that, you know, over the last few years, we've built up a very strong team, in our government affairs and foreign trade, global trade, to ensure that we follow all those rules.
Doug, there was a time where we were all talking about deferred revenue.
We still are.
We still are, yeah. What, what do you think is sort of a normal level of deferred revenue?
Yeah, I keep getting that question. Frankly, I wish I never started talking about deferred revenue back when we were shipping those incomplete systems, because now everybody's asking: Where should it be? Blah, blah, blah. Tim, I think it's about $1 billion, and we finished last quarter at $1.004 billion. The nature of what shows up in deferred revenue today is different than when we were shipping those incomplete systems during the supply chain challenges we had. Back then, it was literally that. We were shipping incomplete systems, shipping sub-assemblies that had to get installed in the customer's fab. They wanted that. That was good for them, but you couldn't recognize revenue if the tool wasn't functional. That is completely caught up now. It's not that any longer. That's at a completely normalized level.
However, we have a set of customers that I, I describe as newer customers, some of whom are private, that we're not sure of the creditworthiness sometimes. Like, I can't see the balance sheet, they're private. And so in that environment, you do business with us one of two ways: you get a letter of credit from a bank, or you pay us money upfront. Today, the incremental component in that deferred revenue is that cash-up-front component, because some of these customers are newer. We aren't completely sure about the creditworthiness, so you've got to pay us in advance. That's what's in the number that's got it a little bit elevated today.
I think I ask you this, both of you, this, every conference that I see you, but when you look at your stock price-
Mm-hmm.
Do you think there are things that people are missing when you look at how the stock is valued, you look at the growth prospects of the company, and you hear the questions being asked in these meetings? Are there particular things that you think investors really miss about the story?
Well, I think that, you know, what I'd say is that it's difficult to talk about the stock price, when you guys can establish, sure, what's fair or not. But, in terms of the growth story of the company, I think there are some elements that are underappreciated. I spoke a little bit about that. Lam's in a very unique position, both in the short term and long term. Short term, obviously, we're coming out of the worst memory downturn in maybe ever, certainly recent memory. And so short term, where I think we see a lot of tailwinds for the strongest parts of our business. We've also used this downturn to do a couple of other things, and that is to ensure that we continued to invest in the R&D needed to expand our SAM in Foundry, Logic.
You know, there aren't many companies that are it sounds terrible, but as underexposed in Foundry, Logic as we are-
Mm-hmm
which we look at as a lot of upside.
Mm-hmm.
So I just ran through this, these inflections, dry EUV resist and gate-all-around and advanced packaging and backside power. I don't think that the long-term growth opportunity there is fully appreciated. I think when we look at 2030, if we're at a $1 trillion semi industry, we're at $150 billion of WFE, and Lam has executed on these SAM expansion opportunities in Foundry, Logic and held all of our positions in NAND, which we fully feel capable of doing, the company will be, will look very different from a revenue profile, a resiliency to cycles, a profitability. The installed base will be dramatically bigger, and I'm not sure that people have quite done that math.
Mm-hmm
To see what a little bit longer picture looks like. Today, we really get tied up in a lot of the questions around 2024. And I think if you look longer, you realize there's an underappreciated growth story out there that is pretty unique to Lam because, like I said, we serve mid-30s% SAM of WFE today, and we have competitors who are probably north of 60%. It means we have a lot of white space to develop tools, grow into, identify new opportunities, and really be pretty choosy about what we go after, and I think that's a great opportunity for us.
I agree, it's the best idiosyncratic growth story in the sector, so.
Keep telling people that.
I do. So anyway, thank you to you both. Great.
Thanks for having us.
Appreciate it.
Happy anniversary, by the way.