Lam Research Corporation (LRCX) Earnings Call Transcript & Summary

Good afternoon, everyone. Thank you for coming today. I'm Stacy Rasgon. I cover the U.S. semiconductor and semiconductor capital equipment sector here at Bernstein. It's my great honor to introduce our guest today, the President and CEO of Lam Research, Mr. Tim Archer. Before I start, I want to -- if you have questions that you want to ask, there will be, at one point in time, QR codes on the screen. You can scan them, and that will bring up a link to -- it's called Pigeonhole. You can submit questions. And we should have some time for Q&A at the end for those. Let's talk about semi cap. So the semiconductor capital industry has been enjoying a renaissance, I think, over quite a few years now. More recently, it really seems to have come into its own, as both the industry growth and I think industry capital intensity have been inflecting higher. The much anticipated sort of end of Moore's Law, far reducing the need for customers to spend on equipment has made -- actually made the contributions, I think, from companies like Lam more important than ever as materials-driven innovation move to the forefront of process technology development really keeps things going. On a personal level, my background is in semi cap. I used to build plasma etchers prior. I love this space. So to tell us all about what he thinks of and hopefully, he loves it too, but it's my great pleasure to welcome Tim to our session. So thank you so much for coming today. I really appreciate it.

Great. Well, thank you, Stacy. And I do love this industry. So I -- and I think we only have 50 minutes today, but we probably could go on for hours now.

We could, yes, if you get me started. I guess there is a safe harbor statement.

That's right. Please refer to the safe harbor statement. As I often joke, if I didn't give you some forward-looking statements, you probably wouldn't have shown up. So perfect.

so let's talk about the industry, right? And in this session, I don't like short term -- in this environment, especially it's hard to avoid it. So I will try to get over it. But it's a weird environment out there right now. And demand is apparently off the charts, and it's been off the charts, and we can talk about why. But you and as well as most of your peers are having an awful hard time serving. And it's -- it'd be funny if it wasn't so sad, right, which is I think they can't get semiconductors. So they're not really -- can you just talk to us in general, like what is going on out there? Supply disruptions, like demand, like COVID lockdowns in China. So what are you guys seeing, like at the highest level, like what's going on?

Sure. Well, thanks, Stacy. And I do appreciate you having us here to speak. If I think back just the last couple of years to think about what's going on, the last time I was here in New York City with investors was March 2020. So it's been a while. It's nice to be back. But if you think about what's happened since March 2020, the world is just completely transformed. And I think more than just about anything, people have recognized the absolute criticality of semiconductors to everything in the global economy. We saw a huge pull-forward of the digital transformation. For the most part, our company -- a good portion of our company has been working remote, just like people across the world. And that's really created this tremendous demand that you're talking about. And as you said, it's unfortunate that we can't actually fulfill all that demand today because many of those same challenges that were created by the last 2 years and the COVID environment disrupted parts of the supply chain. And I think we're just, in the last couple of quarters, realizing the full extent of some of those supply chain challenges. And those have ranged from everything, as you mentioned, COVID lockdowns, which I think continue to still surprise us; chip shortages, which I think is a little bit hard for people to understand that the makers of the equipment that make the chips can't get the chips to make the equipment. And that's -- I think we're making a lot of progress on that end, but it's probably one of the things that's created a little bit of this vicious cycle of chip shortages in a prolonged, longer than we thought. And then you've got labor shortages. And a lot of these are for very high-skilled positions, things that aren't real quick to ramp up. And so as demand grew so rapidly over the last 2 years and you needed to add capabilities across the entire ecosystem for highly skilled positions, many of the companies within the supply chain had challenges ramping at the same rate as demand was scaling. And so I think that -- again, you've heard us talk about what we're doing to try to mitigate some of those shortages, but it's really about getting in, digging in with each supplier, figuring out what their constraints are, whether it's raw materials or labor or chips, and helping them solve those challenges. And that's really where most of our attention has been focused over the last 6, 9, 12 months.

Got it. Is it just semis that are short, though? It doesn't sound like it's just semis.

No. It's a broad array of commodities, but they have similar characteristics. I mean -- again, one of them is chips. We've also talked about some of the highly specialized fabricated parts for our equipment, thinking about things where you're building high-purity, say, gas assemblies or something that delivers the critical materials into our system. Again, those often require specialized materials or specialized skills to build them. And while it's certainly something that we and our suppliers know how to ramp, there's just a time element to how fast you can ramp capacity in those spaces. And I think that now we've said we're making progressive improvement and expect that to continue, but it's taken some time to get that ramp up at full speed.

Sure. And I guess the goalposts aren't stationary either, right? Like every day, it seems like we're getting new stuff going on in the supply chain.

That's right.

Got it. I guess in terms of what you actually have been shipping. So you've given a little more specificity on like deferred revenue levels than you're used to. The number was enormous. What is it $2 billion, right? These are tools as far as -- this is stuff that's actually been delivered. It's sitting there. You can't recognize revenue just waiting for components. Is that true? And is that part of what gives you some conviction in like what that forward trajectory looks like from here?

That's right. So let me explain a little bit about deferred revenue. So deferred revenue, as you described, in general, for the most part, our tools that we have delivered to the customer that are now waiting for final parts that make those tools fully functional and allow us to recognize revenue. It's -- what it gives me conviction around is, certainly, our financials looking forward, but also the fact that we are ramping output from our factories. I mean this is whatever percentage you want to assign to, but it's 98% -- 90% of the -- 99% of the work has already been done within our factories. But yet none of that revenue was yet recognized. And so many people, I think, as we reported financials in the last quarter and gave guide, were wondering about our run rate from a revenue perspective and how we could be confident about a second half that looked better. And the reality is, is because our factories themselves have already been shipping tools at a higher rate, we're just not recognizing the revenue. And as the supply chain progressively improves through the remainder of the year and we get those parts we need and that deferred revenue comes down, then you'll start to see kind of our shipments and revenue start to match up a little more closely, and that's how we get caught up.

Got it. Got it. I know you've talked about like unconstrained WFE versus constrained WFE this year. And like where -- remind us again what your thinking is on those numbers.

Yes. So we've talked -- we haven't really given a number for unconstrained WFE other than to say, we think it's quite a bit higher than our constrained view of about $100 billion of WFE. And I think that -- again, that is partly due -- and primarily due to the supply chain shortages. In some ways, though, it's not necessarily -- it might be a bad thing right now, the underrunning the unconstrained demand.

Next question.

And the reason -- and maybe I'll let you ask the question, but it's -- fundamentally, I guess, I'd give you my view and then you can follow on.

I mean the question was like, look, is it actually like sort of -- like everybody that I talk to is obviously worried that we're going to go into a downturn in 2022. And we can talk about recession and -- but fundamentally, people are worried that we're going to be in downturn. And look, I think at some point, we will. Like it's 100%, it's a cyclical industry still. It doesn't mean we can't grow long term. But does this sort of make you perversely feel better about '23 if we're actually taking strong runs in '22 and actually pushing -- maybe people would start worrying about '24. I don't know, right? But [ what do you think ] about '23?

The industry is always characterized by people worrying about when the next cycle is coming.

I joke -- by the way, we've covered this space for about a year, and it was just so much fun to launch, and it's a little less fun to cover it sometimes because of this.

But I think that if you think about constrained versus unconstrained WFE, clearly, there's been a -- and even as I look at how we've reacted to supply chain shortages, you're really trying to acquire everything that you can acquire. And so I believe that if there's a silver lining to the inability to supply everything on the date it's needed is it's giving the industry a little bit of a chance to look at what does -- how is the end demand really evolving. You've seen with inflationary pressures and perhaps some softening in some of the end consumer-oriented markets, whether it's smartphones or PCs or other markets, it gives customers -- our customers a chance to rethink investment plans and make adjustments and really try to manage this industry for its long-term health. And so maybe that's me looking for this silver lining in the supply chain shortages, but if there was one, it's perhaps stretching out a little bit of this cycle.

Is it fair to say that like even in the midst of like some well-known shortage or demand issues in consumer stuff, PCs and smartphones and TV. We haven't seen any broad changes as far as I know in terms of like investment plans. There's other things that are picking up that slack, data center...

I think that's why we are gaining conviction around $100 billion constrained WFE is that when you've seen some adjustment as a result of a change perhaps in an end market, whether it's consumer, you then -- and you about strength in enterprise and any opportunity that opens up within sort of the ability to deliver equipment, quickly taken up by another part of the market that's actually seeing tremendous strength. So I think it comes back to what I said at the very beginning, which is semiconductors are now so pervasive across the economy that when one part is a little bit weaker or maybe showing a little bit of weakness, another part is likely strong because not all parts of the end markets tend to line up in terms...

Got it. What are your like long-term views on capital intensity in terms of -- I think it's going up. Like I don't know what -- I don't know if Lam has ever sort of put a stake in the ground on where they thought it could go.

We haven't really -- I don't know that we've actually put out a number. But what I would ask you is the kind of a question of -- so do you think complexity is rising. And I think that everybody would agree that devices are becoming incredibly more complex to manufacture. And specifically for Lam, if we think about the types of technology changes that are on the horizon, the world going more 3D, I think that the capital intensity or the complexity and the need for etching and deposition tools is actually rising faster than overall WFE.

Yes. I'm -- I think it's true. I mean we'll see how it plays out. Well, I mean, I look at some of the drivers of capital intensity historically -- I mean obviously you peaked in the, I don't know, late '90s, early 2000s, and it came down. It probably bottomed in 2010 or something. And that was -- to my mind, that was more the 300-millimeter transition that was sort of playing, and it was deflationary. For those of you who don't know, you were getting more efficiency. It was not great for the WFE industry. That's done, like we're through that. I don't think we're going to see 450 anytime soon. You do have complexity increases, whether it was moved to 3D structures and more layers and everything else. We obviously now have another piece, which is the sort of at least talk of regionalization of investment of semi, although if they can ever pass the regulations, like I don't know, we haven't seen it. But I guess like what are your thoughts on -- even if you don't want to put a number out there, what are your thoughts on some of those other drivers, whether it's just the lack of 300-millimeter, if its dye sizes are no longer shrinking because shrink is slowing if it's regionalization. Like are there any of those that you think are like larger drivers? Or does it all sort of like to speed into it?

Well, I think that -- I wasn't trying to dodge the question to the extent that I do believe capital intensity rises. There's always -- whenever you say that, though, there's always the question of what about sustainability of this in the business model. But there's lots of ways in which I think that companies are going to invest to deal with that. First of all, it starts with creating tremendous value at the end market, so people are willing to pay for that. But even within our space, we're continuously focused on productivity. It's one of the key elements. You compete on technology and productivity. And you really compete on who can deliver the technology in the most productive way and high-volume manufacturing. And that's where you need to the capability, you need that etcher that has the technical capability. And it's going to have a cost. So you try to figure out how else do you make that fab very productive so that the customer can still spend lots of CapEx, but have their total running cost be such that it matches their business model. And that's where some of our newer efforts in things like Equipment Intelligence, advanced services, where you're using more data that's coming off of the tools to keep the tools up longer and -- so that the customer is getting more productive time out of those tools. You're looking at things like the Equipment Intelligence driving advanced services that eliminate how much labor you need within the fab, again, removing costs. It lets them put more of the expense pie over into the equipment bucket and a little less into some of the other parts of running a fab. And I think that if we continue to transition the industry in that way, some of this question about, yes, you have to spend more on capital intensity and more on the tools, but I'm spending less elsewhere, and therefore, my business model still works. Our success depends on our customers' success. And so we're partnering very closely with them on how they can be financially successful and we can as well.

I'm going to ask -- I'm going to -- I think it's a provocative question, but let's see. So you're adding all this value to your customers, you're enabling things that they need. There's not that many players. Like all of you guys play in your own niche mostly, right? Why do you -- and even if it is, why do your gross margin start with a 4?

It's a good question. It's -- maybe it's provocative. Let me start with a comment of we're always trying to get paid for the value we deliver. And perhaps at times, we have to do a better job demonstrating and convincing people of that value. But the last couple of years have definitely shown, you can't build advanced chips without Lam equipment, and I think everybody recognizes that. Otherwise, we wouldn't be getting so much pressure within this period of shortage. But I guess it comes back. And maybe I -- we start every thought process around the gross margin and pricing. And I just talked about the business model of our customers and their long-term success and ours. We're very focused on continuously building customer trust. There's not many big customers out there, and there's not many big equipment suppliers. And so I think that over time, what we're trying to figure out and work with them on is how to appropriately share in the long-term benefits of the industry. And so we've talked about, in this period, clearly inflationary pressures and other things have hit our gross margins. We're operating probably about 300 basis points below where our target model is.

They're not still bad relative to where they've been running.

Not bad. And we focused on trying to keep it -- operating margin at a very high level. But I think over time, inflationary costs will be priced into our tools. And I think those are discussions that are being had, and I think you'll see this move up. Our strategy, as you say, is about how to not start with the 4, ultimately get higher, is continue to deliver critical value in the most difficult applications to customers for their long-term road maps. And that's where our R&D is squarely focused on our new products.

Do you guys see a lot of price competition like across? I always feel like it's hard certainly to swap out once your process of record. But I mean, is there a lot of price competition like when it comes time to the bake-off or...

It's -- by that time, it's generally well set. I think that for the most part, most of those discussions around costs of ownership of an application are occurring much further upstream as development tool decisions are being made. We've got to think about if I'm going to commit the development time required to qualify this tool, what does its cost structure look like and what does the running costs look like in the long run? So I'd say those -- they happen much earlier in the process than they historically have.

Got it. Got it. And I guess, when we're talking about that sort of value, it's a good segue into the services business. So this is something that's been -- I mean it's been an increasing piece of the story, certainly much more recurring and sustainable. And I think even in the last downturn was like 2019, and I think your equipment revenues were down 20% in a market that was probably down 20%. The services revenue still grew. Not a lot, but they grew that year, right? Maybe you can talk a little bit more about what -- like maybe just start at a high level. What is in the services business. There's a number of different components to it. They've got different dynamics. And then I'd love to hear a little more about some of the value-add services you're doing, whether it's yield improvement or endpoint detection, whatever is that's actually driving real value to customers and how you're getting paid for it.

Yes. I think that business adds tremendous value to the customer because anything that allows them to continue to get value from the tools they've already purchased is always like their first and top priority. So that business is comprised of 4 different product lines, the spares business, services, upgrades and what we call the Reliant business. It services the specialty technology part of the market. So basically, what someone would call trailing edge, we call specialty technologies, and it services that part of the business. And so each of those, like you say, has maybe a different driver. But in almost all cases, it's focused on thinking about spares, it's about delivering critical components that help the customer continue to operate the installed base at its highest level of performance. And as our installed base has now grown to -- at the end of last year, it was about 75,000 chambers, that business just continues to grow. And even in, as you mentioned, the 2019 down cycle, customers still run their tools, they still try to extract value from the equipment that's already on the floor before they purchase...

Revenues are down, they're not 0, right?

Exactly. And so that -- so the spares business is very resilient and kind of -- we've seen kind of continuously growing. The upgrades business also sometimes tends to be a little bit countercyclical as well because when customers are busy running their tools flat out, they don't have time to upgrade. When they think about now actually being able to have a little bit of time on their hands to make improvements in the installed base, so they come out of the cycle stronger, they focus on upgrades. How do I make the installed base better? And so that as business returns, I'll be more productive.

What's an example of an upgrade, by the way?

There are 2 types of upgrades. You have productivity upgrades and you also have technology upgrades. So technology upgrades help the customer generally move forward to the next node. Productivity upgrades tend to help them get more throughput, extend maybe parts life, improve uptime, these types of improvements. So and then the final part of the business, the -- well, services. Again, I talked about Equipment Intelligence, kind of the use of advanced services. That's really our focus. And that's about helping customers optimize using a tremendous number of sensors that are on the tool using the data that comes off of that to help maintain uptime, improve uniformity and help match chambers faster and better to each other within fabs and also across fabs as customers try to ramp maybe the same technology now in different parts of the world. When you talk about regionalization, you're going to have fabs that try to run and be qualified for the same products on one continent and another. Things like chamber matching using our Equipment Intelligence services will actually become very important. So that's a nicely growing part of our business. And then finally, the reliant business, I mean, you've seen -- you can't get a microwave oven or something because you can't get the controller chip needed for the touch panel, or you can't get a car because -- or can't get, the one I read was, seat heaters in your car because you can't get the chip that's needed. So that part of the business has just been really expanding as chips find their way into just everything throughout the economy, both the tech world, but also really in the more industrial- and consumer-oriented end of things.

Got it. Have you guys ever said how big Reliant is like sort of relative to the total company?

We have not. We have not.

Okay. Feel like saying it today?

I don't feel like saying it today. I'd probably get it wrong if I try to quote the number right now. But it's an important part of our company and one in which -- we -- why we call it specialty technologies, I'd actually say, is because it's transitioned from just the older stuff to now actually very specialized technologies. They just happened to use larger dimension structures, right? But they have needs for new materials, certain new types of deposition or etch equipment. And in those cases, it's not easy. And that's why it's the specialty technology. These are, for what they do, actually leading-edge chips. They do what they do best in the world. And so we're, in many cases, now developing new applications to expand our portfolio of products that address the specialty technology space. And so we see it not just as an afterthought, which is to sell into this market some of the same old tools we've been developed in the past, but actually now is an area of potential growth for us.

How does the economics on like a Reliant or a specialty tool look relative to like the more traditional mainstream businesses?

Overall, quite good. It's -- again, it depends on the tool and the value it's adding to that specific application. But in general, what you can think of is the margins are in the range of the corporate average. And if they're a little bit lower, it's because that tool is reusing R&D that was invested long ago, and therefore, the expense profile is lower. And so overall, it's a nicely contributing part of our business.

Got it. On the services piece itself, how much of that now is like subscription? Is that something that you've been...

We haven't really talked about it too much. But -- I mean I would say that when you start talking about the newer, what we call, advanced services, more of that type of business is on a subscription basis, meaning paying for access to either data or services or these analytical services that we're talking about. And -- but I'd also say that in many of those cases, it's a little bit easier to convince a customer quite often with a more results-based contract discussion, which you can...

Where you kind of -- we're going to save X and we can shave...

If you're in and you're saying, if I use this data or I provide these services that I can increase the output or the uptime of a tool set, customer very easily can recognize what that value is and therefore know exactly what they should pay you. And so that's part of our business as well, which feels a little less guaranteed except for the fact that we have strong conviction because we quite often have already replicated that same improvement on the same types of tools in other locations. So there's many different ways we engage customers. But the primary idea being we're always trying to add value to their end operation. And so whether it's a results-based or it's a subscription, we -- probably we've talked about -- I don't know if we've talked about it or not. But within a subscription, we also have something we call cost per wafer pass, which means a customer will actually pay you for the wafers run through the tool and will provide some of the services and spares required. And so again, it's a bit of a results-oriented commitment. This is -- we're the ones who know how to extend life of parts and how to minimize the amount of service that the tool requires. And you're kind of really aligning mutual interest in that case. They get a fixed cost per wafer, and we take on both the opportunity and the risk of providing and improving the parts and service. And so many ways that we partner with our customers, too, for mutual benefit.

Did your large customers used to mostly do the service on their own? And is that changing?

It varies customer-by-customer. And so we have different models. And again, in the interest of helping the customer, we will work with them on making a model work that fits.

Okay. Got it. I want to ask. I wanted to ask a little bit -- we talked a little bit about regionalization. You sort of like dodged the question on the actual regulations themselves. And like did you -- dodge is the wrong word, but it got buried. What are your thoughts on like what's going on with the chipset, both here as well as in Europe? And like it's just taking forever. It's been well over a year, and you guys are obviously like well involved in those discussions. And what is going...

Well, we're deeply involved in the discussions and have supported many of the discussions around the CHIPS Act and the importance of it to, I think, the overall ecosystem. I mean clearly, part of our message has been outside of this idea of regionalizing manufacturing has been -- part of my comments have been driven a little bit by what we've seen in the recent shortages, which is you need to invest in the entire ecosystem. And that ecosystem runs from what we would call more fundamental R&D into the types of materials and processes, plasma physics, for instance. We're engaged with the national lab on this point. We're really trying to establish the groundwork and the fundamental technologies that are going to make for a very robust semiconductor industry in the long term in the U.S. That's what we think would be a great use of government money as well as a lot of the training and labor development, workforce development. As we talked about, some of the shortages are caused by workforce shortages. And so just, again, a government focused on how to ensure that the semiconductor industry, as it grows in the U.S., has sufficient talent to actually deliver on its promise. And I think Europe, similar kinds of conversations. So I think, again, investment to strengthen the capabilities and resiliency of the semiconductor industry as a whole, I think we look at it as a very positive thing for our company.

Got it. I want to ask about end markets a little bit. Lam is often thought of as like the NAND company. I always feel like you've got to ask, like as an investor, do you like NAND or do you love NAND? That's still true in some sense. You get a lot of NAND revenues, et cetera. But it's been getting less. You've got other pieces like logic, in particular, has been growing a lot. And you've talked a lot about some of the opportunities at some of your large logic customers have moved from 14 to kind of 7 and whatever they're calling it now. Your logic [ customers have like ] doubled or more than doubled. But maybe could you just talk a little bit about -- I think we all know Lam's position in NAND. But could you maybe talk about some of your position in some of the other areas where historically investors weren't really looking at Lam?

Yes. The only thing I'd correct there is I would characterize Lam as the 3D NAND company.

3D NAND company.

And the big distinction there being the 3D piece is really important. And it's because...

But we thought 3D today almost, isn't it?

It is. But the fact that Lam has gained tremendous experience in how to architect 3D devices is going to be very valuable in the inflections that are going to come in the remainder part of this decade. And that's where a lot of our focus and growth really is going to be in foundry, logic and in DRAM as well and also in advanced packaging, multichip packaging. So that 3D element. Recently, we just introduced a new suite of products in targeted foundry, logic gate all around selective etch, also some of our new ALD products targeted towards helping to build and fill highly complex 3-dimensional structures, and we're making great traction there. You'll see that inflection come. You see the same thing in 3D packaging as you're having to do plating of these multichip packages. And you're seeing -- going to see it in the latter half of the decade as DRAM goes through. So the 3D element and our experience and position that we've built up there is really important. But beyond that, we're also focused on how to accomplish 2-dimensional scaling. And so you've seen our attention shift in some ways towards how to make EUV more productive and more capable. And one, if you're going to print an excellent feature, you need to etch that feature. And so we brought to bear high aspect ratio, really fine pitch etching. And we really are a world leader in that capability. There's a number of capabilities in our tool that allow us to compensate for incoming nonuniformities using one of the innovations we have called the Hydra chuck, which allows us to adjust temperature, about 100 different zones on the wafer. And so you can basically look at incoming wafer nonuniformity and balance that out through the etch process to create a great outgoing wafer. That's -- those are important for things where you're trying to open really small features. We're also working on our new EUV dry resist.

Dry resist, yes.

And that's an area where, again, when you think about the company, we're an etch and dep company, but we're also interested in expanding our SAM into new areas and so -- where we can still use etch and deposition. And so there, we're trying to disrupt the market that's existed for deposition of wet photoresist by spin-on, track market.

So like when you talk about a $2 billion opportunity for dry resist, the track market is about $2 billion. Is that the opportunity?

That's right. We've actually said -- eventually, as we get out there, we've said over about a 5-year period, probably about a $1.5 billion opportunity for us as it ramps. But the nice thing about capturing a market like -- where you're tied to sort of EUV as number of layers, we'll continue to expand in the future. And so it's an important opportunity for our company and a place where we can think we can add real value based on this transition to a series of -- it's actually a series of steps. This starts with the deposition of a dry underlayer that helps with the absorption of EUV photons that helps with this dose-to-size reduction, improves the EUV productivity. Also the deposition of the dry resist and then the development of that process using also [ an ally ]. Yes, essentially. So we've just taken this and transitioned it to a series of deposition and etch processes, which is -- which are technologies that we know very well and have equipment that's readily adaptable. And so we're in the early stages of that. We have that equipment installed at all of our key customers that use EUV, and I think in time, we'll see exactly...

It's got other benefits, too. Obviously, there's a lot less liquid waste. You don't have like pattern collapse. There are ESG benefits from an environmental standpoint.

That's right. 5 to 10x less chemical waste. And issues that come with wet processing like pattern collapse due to the stresses put on through the drying process, those are problems that get eliminated in dry processing. It's one of the reasons why, over time, we've seen a lot of wet processes get replaced by dry processes. And I think into the future, that's a trend we believe continues to happen.

Got it. And I think that makes sense. One of the controversies just not for Lam, but just broadly for Lam and -- always is EUV going to like taken away from me, right? And it's funny like -- so we launched on the space a year ago. We've been actually working on and off on that launch for like a long time. And I got a whole section written at one point of why I didn't think -- and by the time we actually had logic, it was sort of decided, right? I mean I suppose is the right way for me to characterize it, without EUV, I guess, there probably would be a lot more, in theory, etch and dep, but there's still a lot of etching and dep even with it. And without EUV, everything comes to a screeching halt anyways.

I think it's back to your sustainability and extendability argument, which says there is a world where you can't have -- you don't have EUV and you need a whole lot of etching and deposition equipment. But that world is too difficult to implement from a cost and a yield perspective. So it's a future that actually would never really exist or a lot fewer wafers will get run in that world. And so I think we've embraced the idea that EUV is good for moving the technology road map forward. And therefore, we've jumped in with both feet with ASML and imec to figure out how you just make it more productive, so it can be implemented in a way that doesn't disrupt the business model of the industry. And then I think that's, again, part of our motivation, and I think we're making good progress.

Got it. Got it. Can you talk a little bit about the Malaysia infrastructure that you're putting in place? Now you've got a new factory. It is hitting margins now to a degree. That should relax at some point once it comes online. But like what is that facility? Why are you building it in Malaysia? What are the benefits that it brings to you?

Well, every time, we look for a new site to build a facility -- first, I'd describe we have most of our manufacturing capability in the United States. And so again, everybody -- we're looking for places where we can find the right talent, places where we're close to our customers, close to our supply...

And here, you've got to ship.

Yes. And so as much as there may be a lot more manufacturing in the U.S. we're still heavily concentrated with our customer base in Asia. And so building a large facility in Malaysia had multiple benefits, great access to talent, a low-cost region, place where we can convince many of our suppliers to ultimately congregate around and also take some of the benefits of that region. It has a very well-established semiconductor ecosystem already. And so we've ramped it at quite fast speed within Malaysia. Now its output and the ramp speed is now somewhat governed by our ability to get enough parts into the factory. But we're continuing to ramp it quite aggressively. And as you said, it had been a headwind for margin. I think Doug was telling me last week, I think now we're probably feeling it's a little bit margin neutral and becoming a little bit margin accretive as we move through the remainder of the year. So we're getting to a scale now. And we're going to continue ramping. We're getting scale now. We're starting to...

How much of your production capacity will be in Malaysia by the time it's done?

We haven't said, but I think it depends on -- obviously, it will be a key growth site for us. But today, we've been -- and for the last 2 years, we've been ramping every manufacturing site we have worldwide. So assuming demand remains strong, we'll continue to ramp every site we have. And its percentage won't rise quite as fast, but I think in the longer run, we look at Malaysia as a very important part of our -- and a very large part of our manufacturing network. And -- but we've still got several years to go to get to that point.

Got it. And I guess, talking a little more about Asia and then let's talk about China. So one of the -- I get a lot of people look at the China revenue. I can't remember what it was last quarter, 30% or probably somewhere in that ballpark.

Roughly 30%.

Is that amount of revenue exposure a concern. People are obviously worried about potential regulatory threats. I know you can't ship everything you would like already just because of some of the players. What are your thoughts about, just in general, the revenue concentration in China and some of the potential like -- like if they don't build it, does somebody else have to? I mean...

Well, a few questions there. So maybe just first on the regulatory. Breaking down the 30% of our revenue in real rough numbers, about half of that goes to multinational companies, about half of that goes to domestic China companies. And so from that perspective, while you can never predict exactly what a regulation might look like, half of it is not exactly what people think when they see shipments into the China.

And even the local stuff, like how much of that is like leading edge versus trailing edge?

And then again, already, as you said, there are restrictions about -- around shipping of leading-edge tools to customers in China. And so to a certain extent, in divisions, it's mostly targeting these trailing edge nodes. There is, of course, a memory side to that business as well, which is -- has been expanding. It's an important region for us. But I think that we're deeply engaged in those discussions with the government and really partnering to understand how can we help them accomplish what it is they're trying to do without weakening what we think is a very important industry for the U.S., which is the American equipment -- semiconductor equipment industry. And so I think that we'll always follow the regulations. But first, we want to engage in the conversations. And we've built up a government affairs team that's giving us [ insight ] of those conversations. And I think they're -- no decision -- no, I think this will be an ongoing process for quite some time.

And like I have some investors who worry that the Chinese customers have been stockpiling tools in front of potential action. How long could they run those tools without spare parts if that were the case?

Not acknowledging or not acknowledging whether that's the case. I don't believe it is, but the reality is Lam provides important spare parts.

That was my tool. Right. So yes. You're not going to be able to just -- if you're cut off, I mean like the tools become junk and -- junk is the wrong word. You're not going to run them that long, right, if you're cut off, is that the right way to think about that?

Yes. We believe that our ongoing support with spares and services is important to operation of the tools.

Yes. Got it. I've got another -- maybe it's a provocative question. But you and your competition all have a very positive view on the market. You all have given like long-term targets. All of those long-term targets imply share gain. So how do I think about reconciling this? And look, I mean this is like -- I look at how you guys as well as many of your competitors have executed against those long-term targets. By and large, you've executed very, very well. But usually, it's because the WFE number comes in a lot higher than you thought. And the share, it's always tough, right? And some of that was EUV getting like bigger as a percentage of WFE. How do I think about market share? I generally think about this market as -- market share as being relatively stable. At the same time, though, I do see things like your logic isn't that big for you, but again, it has doubled, like in a rapid fashion. To me, that feels like. I don't think CapEx hasn't quite doubled for them, not yet, right?

No.

How do I square the circle on market share?

Well, I think everything you said is kind of true. Within the established markets where applications are reasonably set, we think those positions don't really move much. In fact, it informs our strategy, which is -- it's rare for us to aggressively go after an incumbent -- a position where there's a strong incumbent. Instead, our strategy is to go after and be patient and wait for that next technology inflection, spend our time looking at what new tools are going to be able to really address that need and go after that. I mean -- so if you look at many of the new tools we've introduced and people say, when will I start seeing revenue. It's like, well, it might be when 2-nanometer comes in. But our business is such that you have to invest pretty far in advance of the inflection to ensure that you can understand the need, you can develop the right tool, you get the tool into the development fab, and you can get it qualified, and then you ramp. That's often a 3-, 4-, 5-year time horizon for that to happen. So when we talk about our confidence level, it's where we've now put a number of those new tools on the floor where there is no incumbent. We think we have the best tool and now we need the technology node to ramp. And in some of these cases, where we've talked about having been patient and placed tools and gotten them qualified, but we're waiting for the ramp. I spoke on the last earnings call, there's a particular customer where node-to-node, we will see our conductor etch share double from one node to the next. We've been waiting a while for that technology transition to occur. But it's -- but we're really running and investing for this longer horizon rather than trying to go compete for short wins that usually end up being either unsuccessful or highly unprofitable once you do win. So that's -- part of that, I guess, is just also to explain from our business perspective, it's because many of the companies in our industry, including Lam, we've built incredible moats around our positions. You talk about within 3D NAND, if you look at the number of wafers, and we've quoted a number that's probably now a little bit out of date, but I think it was last year, we said we had at that time run 45 million more wafers than the competition through the 3 most critical steps. The learning you get from that and the customer engagement and the installed base the customer has says, it's a very high switching cost. And it doesn't mean that we don't have to like meet their technology needs and meet their productivity asks, and we're partnered with them, but you definitely get first shot every time because that is the best choice for them is to figure out how to extend that equipment. And that makes many of these positions where they don't turn over for 10, 15, 20 years as long as that equipment is there and doing its job. And so you have to -- if you want to move share, you have to do it on new applications they come up and, as architectures inflect, gate all around 3D packaging, these types of inflections.

Got it. We've got about 5 minutes left. Should we go to the lightning round? Except I just lost it here. So hold on one second, Lam Research. Let's see what we've got. Okay. In the event of a downturn, what kind of flexibility do you have in the business model to trim costs? I guess the question is probably more around like resiliency in a downturn for your model. That's probably a better way to ask it.

So we operate a highly variable model. I mean the best thing I can say is go back and look at 2019. WFE was down last down cycle. And we still delivered, I think it was, 26% operating margin and second best EPS in the company's history. So we can deal with a downturn reasonably well.

To be fair, even where you're running right now, though, like it's elevated versus like where we've seen before, though, right? I mean I wouldn't -- I don't know if I'd call this a downturn just yet even if -- the numbers were a little weaker than...

No. We're talking really about demand -- I think you're talking about the demand in WFE downturn, I think. No, 2019 was a good example where there was a small correction.

Yes.

And we actually took actions within our operating model and we're able to respond.

I guess your trough margins in '19 were probably higher than the peak margins were in the prior cycle.

I didn't check that. But I'd have to go back and look. I think we're -- what I'd just say is we're trying to get stronger through each cycle. And part of the model we have around how we build our manufacturing supply chain is to ensure that we continue to maintain that variability in our model.

Got it. There's a couple of questions here that all sort of kind of address the sustainability of capital intensity going up. But I think the big question is, obviously, if it's going to keep going, who is going to pay for it. And I guess, just like what are your views in general? Like does it mean the returns for your customers have to go down? Or do you customers -- that means your customers have been more successful at proving the value to their customers? I mean, obviously, somebody along the line has got to pay for it.

I think it's one -- I think one of those answers is likely part of the equation. And as I mentioned, we also play a part in that, which is to try to help continue to drive productivity. It's easy to -- it's actually quite easy to deliver technology. it's hard to deliver highly productive technology. And so that's where we come in. And if you look at our most recent etch tool, it's the highest footprint density tool that's available in the market. And that's so that customers can actually produce more wafers in -- within a same fab footprint. And those are the kinds of things that we can contribute to helping the customers' business model sustainability.

Got it. This one has a number of votes. I'm going to ask it. You may not like it, but I'm going to ask it. So when you look across the different nodes or businesses during the pandemic, is there any spot where the business is well above trend? Ideally meaning more potential for it to normalize? Is anything like more on leading, more on trailing? I think it's just like versus where you were running? More elevated than anything else?

Well, I mean the question is relatively easy to answer because really the single biggest part of the market that moved up was foundry, logic during the last 2 years, above historical spending. Now whether that's in line or not in line with long-term demand, it's probably for somebody else to answer. But clearly, that part of the market is what's really grown during the last couple of years.

Got it. Math on long-term services growth and margins. I know you've given some numbers on installed base growth, but like, I guess, think about like maybe services, content per tool and like how does it flow through the margins.

Yes. We gave at our last Analyst Day this -- a bit of a line that showed how we were trying to expand revenue capture or capture per chamber in the installed base. And I think we've done really well against that objective. You've seen our...

Was the growth 50%, or it had grown 50%?

It had grown 50%. I mean obviously, what we've said is it will moderate over time as the installed base continues to get bigger. But one good rule of thumb is to think about our services business growing faster than the rate of growth of the installed base.

And that's growing 5% or 10%, 10% maybe for you? Or the other way around?

Maybe in that range. And obviously, this last couple of years have seen significant installed base growth as WFE has been very high. But our objective is to grow that installed base business faster than the installed base itself is growing. It's the easiest way to think about it that way.

So we are almost out of time, and it's unfortunate because I could do this for another hour or more. But I want to give you your soapbox. Like we're at the end here, you've got a room full of folks. Why should investors buy your stock?

Okay. Well, it's -- I think that if you're thinking about buying semiconductor capital equipment stock, in particular Lam, first, you'd have to be bought into the idea that semiconductors are fundamental to the long-term picture. And I think that people are pretty much in agreement with that. And so our underlying driver of our business, I think, is really solid. And most people are coalescing around at least $1 trillion semiconductor industry by 2030. Actually, I've seen recently that number even moving up. And...

At 5% a year, it's $1 trillion in 10 years, plus or minus. So yes.

So that -- so the underlying driver of our business is good. And then back to your question, a second point, are devices becoming more complex or less? I'd argue they're becoming a lot more complex. Is the majority of the technology inflections moving in the direction of more etch and dep capital intensity? I think the road map show the answer to that is yes. And then you look at the track record of our performance. Pretty good at -- while you say it's hard to gain share, we've shown an ability to gain share over the last 6, 7, 8, 9 years that have been pretty substantial, and I think that we have an ability to continue that into the future. And then maybe finally, just from an investment perspective, generate a lot of cash. And we've made a commitment to return 75% to 100% of that cash to our shareholders. And I think that's -- that means you have a growing business, nice cash return.

You've been doing more than that, too.

We've been doing more than that. But our commitment is 75% to 100% into the future. So...

Got it. I think that's as good of a place to close it out. Thank you so much.

Yes. Thanks very much, Stacy.