Lam Research Corporation (LRCX) Earnings Call Transcript & Summary

Good morning and good afternoon. Thank you. I'm Tim Arcuri, semiconductor and semi equipment analyst here at UBS. Very pleased in the session we have Lam Research, and we have Doug Bettinger, of course. Most people know Doug. He's VP and CFO of Lam. Before Doug and I get into it, we're going to have a safe harbor.

Yes. Tim, just give me 10 seconds, real quick. Everyone, please take a look at the safe harbor statement. It's consistent with our filings on sec.gov. I may make forward-looking statements, in which case the safe harbor is intended to provide guidance around how you should be thinking about that. So with that, Tim, why don't we jump into it?

Thanks, Doug. So obviously, last call, you talked about WFE this year being in the mid-80s. I think that everyone is sort of in the same range. And since you spoke, AMAT's been quite confident on WFE next year. Bob was talking about it growing 10%. And at the end of the call, Bob, he's saying that he thinks that the bias on 23 is up mostly because of the supply issues, and that's just going to push shipments out longer as you kind of work on the backlog. So I'm not asking you to commit to 2023 at all, obviously, but I'm just kind of wondering your current perspectives on the market and some of the supply issues and sort of when you think you can catch up to demand.

Yes, Tim, thanks for the question. Yes, as we walked through '21, we kind of incremented WFE up kind of bit by bit as everybody got more confident around the industry's ability to supply, frankly, was kind of how things looked up. And you've got it right, mid-80s is generally what we're talking about for this year. Now our normal practice is when we release earnings for the December quarter, we give you the color about next year and what's up and so forth. We haven't done that yet. But what we have suggested as we look into next year, looks like a strong year next year. I haven't gone as far as to talk about '23. That's still a little beyond what I'm comfortable talking about. But as we look into next year, every segment of the industry looks decent. Like I said, growth year next year. Maybe part of it is as we're -- and maybe this is implied in your question, we exit -- we're going to exit this year with unmet demand. I think everybody in the industry is because of the supply constraints that we're all dealing with. And I know you're going to ask me about that later on in the conversation, so I'll save that for a moment. But yes, when we look across all aspects of the industry, look pretty good. And so a step back, so what's going on? One, demand for semiconductors is very strong, right? I think that's not going to change going into next year. When you couple that with the process node transitions, capital intensity going up kind of node by node, looks pretty good next year, frankly. So when we're going to get caught up with demand? I'm not ready to say quite yet. It's a bit of an incremental activity is [indiscernible] one thing. You got to work on the next thing, and we're progressing through that reasonably. I know you're going to ask me about that later on. So Tim, maybe I'll just -- I'll pause and leave it there, and you can redirect if you want to redirect on anything.

Sure. So one question that people ask. I think I know how you're going to answer it. But one question that people ask is they say, "Well, because lead times are significantly longer than they typically are, that it's not double ordering, but it's causing customers to place orders earlier than they otherwise would have, something to get "in the queue." Now I guess my question is, it's difficult to tell, but what portion of your backlog and kind of your orders do you think is because -- or is customers getting in the queue earlier than they otherwise would? Is there any type of...

I think what I would tell you, Tim, is, yes, lead times are stretched out. And so as a result of that, we have better visibility for a longer period of time, frankly, for the reason that you're alluding to. Now I'm highly confident that there's not double ordering because, frankly, I don't know how many tools will go into every premium space. We've got people -- that's their job, right? They call in every account, people at every fab. And so we kind of know what that looks like over time. But I do believe because of the elongated lead times, we probably have better visibility for a longer period of time and going into next year. And maybe that's what you're implying.

Got it. Got it. Can we just talk about China? So domestic China, obviously, a big growth driver. Some of your peers are suggesting that domestic China could be down a bit next year. I'm just wondering sort of what you can say about domestic China.

Yes, I'll remind you what we have said, and again, I haven't given all the color. I'll be sure to do that when we release earnings this quarter about, hey, what's going on in China. But China is pretty strong, frankly. Remind you, it's been 35%, 36% of our revenue in the last couple of quarters. And so understand -- I always remind people of this, that when you look at my geographic revenue, that's both multinational customers with fabs in the region as well as domestic customer base. Split fairly balanced when you look at the last couple of quarters between domestic China as well as the multinational fabs in China. And so that's important to remind people of. And so we said last year, domestic China WFE was about $10 billion. And then what we've said this year is it's growing this year. Now I haven't made any statements about next year, but I have no reason to believe that the strength that we're seeing with the local Chinese customers doesn't continue. I don't know why it wouldn't, frankly. It's pretty broad-based set of customers between big NAND customer, big DRAM customer, and then a whole bunch of, I guess, I call it lighting edge foundry, logic investment occurring broadly across that. And so I have no reason to believe strategic intention of all of our customers in China, the Chinese customers doesn't continue, right? In fact, I'm feeling pretty good about where it's going. Check back with us, Tim, after we release earnings. I'll give you a little more color on China when we get there.

Perfect, Doug. And just going back to the constraints, and you did say that I would ask you about it. And so I just wanted to see if there's any characterization that you could provide on sort of any magnitude between what customers are asking you for and versus what you can do. I think there were customers sitting in your parking lot pounding on you to get tools early, and I'm sure they...

They're pounding on Tim. They don't pound on me quite as much but...

Right, right. Well, I'm sure that the pounding hasn't ceased, but maybe now they know that pounding on you is not going to help. So like maybe there's been some reduction of the pressure to get tools earlier. But I'm just kind of wondering if you can characterize the unmet demand that if you had the supply, you could do X hundreds of millions of dollars more in revenue in December and even in the third calendar quarter?

Yes. I didn't quantify a number, although it's not a small number, Tim, in terms of "unmet demand." Lead times are stretched. We're working on this, and we're being pretty transparent with customers about what we can do, what we're doing, when they can expect certain things. And that's progressing. I feel pretty good. And frankly, if my supply chain team is listening, thank you. I mean it's -- there's some heavy lifting in there, right? These are complex tools with many, many subsystems, many individual components. And frankly, when I go through and think about how we've managed this, I don't know, through the last 6, 9 months, we've done, I think, actually, as good a job as anybody in the industry managing this, although nothing is perfect by any means. I think we've done a nice job managing our own internal capability. We were fortunate in that we had a new factory coming online in Malaysia. And in fact, we were able to accelerate that ramp. You've seen us announce a new facility in Oregon, actually where a lot of our deposition tools are. And frankly, in all of our facilities to the extent that we're able to, we're expanding our outlook capability everywhere, be it in Ohio, in California, Oregon, I talked about. Our factories in Korea, Taiwan and Malaysia, we've been pretty transparent about what's going on there. So I feel good about what we're doing with our own facilities. But as I alluded to, we've got a very complex supply chain with a lot of supply chain partners. And frankly, that's where the constraints are being worked on, right? It's a layer or sometimes a layer or 2 deep in the supply chain. And one of the things I've been talking about actually are a little bit circular is -- frankly, one of the things that's constraining us is semiconductors right now, right? I mean the industry needs more equipment to generate more semiconductor output. And in fact, that's one of the things that we're working on and mitigating the constraints there with our supply chain partners, obviously. We don't directly procure semiconductors, but we're helping that because that has emerged as one of the things that is a couple layers deep in the supply chain. I would tell you we're making progress. I feel good about kind of what we're working on. It's incrementally improving, and we'll keep going at it. But I haven't quantified kind of the unmet demand, Tim. I know you asked directly. I'm not ready to tell you that right now. But I will tell you, it's getting incrementally better.

Got it. Yes. And it seems like the situation bottomed maybe about a month ago, and we're sort of moving maybe slowly in the right direction, but we're at least moving in the right direction.

It feels that way to me. And when I talk to my peers and our suppliers and just generally listen to people talk, yes, it feels like it's getting incrementally better, but there's still a lot of work left to go, I think.

Yes. I guess Malaysia is a point of potential concern because of a lot of your suppliers or several of them have key sites there. So does that concern you? Or speaking of you, I mean, yes, you're ramping your own site there. With the new variants, have you heard anything that worries you that maybe they're going to institute some more stringent lockdowns the way that they did not too long ago here?

I think we're all kind of in a wait-and-see mode here with the Omicron variant coming out. I haven't seen anything new. Obviously, we're all monitoring it very closely, not just in one country, but everywhere, right? But so far, still nothing to tell you. What I would say, though, Tim, we've been operating in this COVID environment, I don't know, coming up on 2 years now, right? At Lam, we've learned how to do this pretty well, right, with PPE and distancing and everybody's temperature gets taken as they come in every day and whatnot. And so I observed that happening across the totality of the supply chain at our customers here in our own company and the rest of the industry, our suppliers. It feels like we kind of know what the playbook is here to manage this. You don't know what you don't know. And I think we're monitoring things very closely. But to your direct question, we're still operating pretty much everywhere in the world, no change there.

Got it. Got it, Doug. Speaking of Malaysia, it seems like maybe for you, your ramp of that new location is costing you maybe 50 basis points or something like that right now in terms of gross margin. And then, of course, they are afraid and there's logistics constraints. But it sounds like if you excluded all of those factors and you sort of normalize for those, that you'd be more in line with your 48% model versus the 46% that you're guiding. Is that reasonable?

Yes. Tim, it's not unreasonable. I've been starting to tell people that, frankly, last quarter, 46% -- guided to 46%. If you go back to that March of 2020 financial model that we provided, that I provided, implied in that model was a gross margin kind of 47.5 to 48. And frankly, at these revenue levels, you might expect us to be at that, if not for the challenges that you're pointing out, right? Malaysia is a little bit of a headwind as we ramp that facility, and it will become beneficial to gross margin in the latter part of next year. And then the freight logistics and other supply chain challenges that we're dealing with, there's a cost associated with not just freight logistics, Tim, although freight logistics is the biggest piece, and maybe the one I've been talking most about, but there's other stuff too. And as we work through that, that will get better over time. But yes, what you're suggesting isn't an unreasonable way to think about it. At the revenue levels we're at, it's sort of where the financial model was predicated on. And so yes, we're a little shy of that.

Got it. And then I just wanted to ask about DRAM versus NAND. It seems to me like people just kind of lump memory together, but it seems to me like there's a little bit of a different story happening in DRAM and NAND. And it seems to me like DRAM, if you -- And I -- I mean, I have quarterly estimates about how big the DRAM WFE TAM is and how big the NAND WFE TAM is. And you guys do very, very good work on this stuff. And so I'm kind of wondering whether you would agree with the idea that DRAM WFE is sort of like peaking right now. And if anything, it's going to come off a bit through next year. So if there is -- I don't want to say downside to something, but if something was going to be down next year, it's most likely DRAM. Versus if you look at NAND, NAND is almost like -- I don't want to say troughing right now, but it sort of has, if anything, maybe an upward bias as you kind of move through next year. And then, of course, obviously, we know Foundry/Logic is strong. So I'm wondering, is there any sort of finer point you can put on DRAM versus NAND as you kind of look to next year?

Tim, honestly, like I alluded to in your first question, everything looks pretty strong next year, right? And I haven't given segment level, this is up, that's flat or what have you. But frankly, the nature of the conversations we have with customers, it's fairly long term in nature and maybe because of what you suggested. Like I said in the first question, we tend to -- are somewhat elongated, and so you get better visibility. Everything looks pretty good. And the investments I observed being contemplated in the industry are made with a lot longer-term orientation not just where spot pricing today [indiscernible]. Frankly, spending next year is going to generate output late in the year or the year after we have is really without orientation, honestly. And I'll remind you, and I know you know this, but I'll remind the audience that as we look at the transition from 1Y to 1Z, the 1 alpha to 1 beta and so forth, our addressable market gets bigger. Similarly, in NAND, as you go from 96 to 128 to 17x or 19x depending on the customer, again, our addressable market there continues to get larger as well. So I'm not ready to say -- to tell you, hey, up this much or what have you for next year. But just to remind everyone, our market gets bigger every subsequent process node. And that's why I'm really excited about the position Lam sits in. It's somewhat unique in that regard.

Got it. And just shifting to WFE share, your WFE share based on what you're guiding systems in the fourth quarter and based upon the mid-80s WFE number, which I totally agree with, is pretty flat year-over-year. And so I guess the question is as you look into next year, there's a couple of different factors. One, you have Sense.i coming on. Two, you've got more ALD. You've got more the sort of ex SAM is growing because of selected etch. And so I'm just wondering if we're here this time next year, do you think that you will have gained WFE share next year? Or is the story more just that these NAND-specific things are more in the out years versus in 2022? Because I ask because obviously, you're several hundred basis points below what the model implies for your WFE share as a portion of total WFE spending. And so I'm just kind of asking for a bridge that you can give us there.

Yes. I mean I'll remind everyone on the webcast that, again, every node, our addressable market gets larger. And I think everybody understands that in different segments of our customer base, we have a higher wallet share. In some, it's a little bit lower wallet share. And frankly, when you look at those process transitions in each segment, our addressable market gets larger. And so there was a time component to the things you're asking about, Tim. In that model was a '23-'24 model, and we're sitting at the end of '21. And so there's no progressions in foundry. As an example, from 14 to 10 to 7 to 5 to 3, our market gets larger. Our share aspirations because of the investments we're making, right, you referred to Sense.i, that's one aspect of growing the foundry logic and NAND and DRAM. And I'll remind everyone, also in NAND and DRAM, we have a similar progression from 1Y to 1Z, alpha-1 beta as well as the growth in NAND layer count. So the time piece of it is important consideration, Tim. Sense.i is another important consideration. The enhanced ALD that you alluded to is another aspect of as we come into some of the applications we're targeting in foundry and logic, our intention is to displace older technologies, frankly, that we're not in, like a PVD, [ resist ] and on dielectric, there's a handful of those things. In addition to -- and this is something we continue to be very excited about the dry resist tool that we first talked about in March of last year, seeing nice pull on that, although it's still relatively early days in that. So when you put all of those things together, that's the time component that it just needs time for these node transitions to occur.

Yes. I guess from that point of view, Doug, so the model is 2023-'24, and you didn't really give a specific breakdown of service versus product. So we don't really know what the product mix was in there. But if you make pretty pedestrian assumptions on the service side, you get to a 17% number for WFE share that's kind of inherent in that model, and you're sitting at just above 13% now. So there's 350 basis points, maybe 400 basis points worth of share to get from where you are in '21 to where you want to be in '23, '24. And those -- and all those things that you had talked about are pieces of that. But is mix also a piece of that? In other words, if the mix stays this skewed toward Foundry/Logic, is that 17% off the table?

I don't know that it's off the table, Tim. What you're alluding to is right, though, right? Our highest share of wallet is in NAND followed by DRAM, followed by foundry/logic. And like I said, we're targeting incremental SAM in every one of those segments, incremental share in every one of those segments. But mix is a piece of it, certainly, Tim. And I don't know, I probably need to update the model at some point. We'll probably do that not all that far down the road and go rescrub all of the assumptions. Mix is a key assumption, but there's lots of other assumptions that went into that as well. And yes, CSBG, frankly, is reasonably well ahead of where we suggested. So yes, I need to give you a model at some point. We'll get at that, hopefully, not too far down the road.

Okay. Great. Just on that front, I had a question coming in from investor as to whether you see your incremental benefit from 3D DRAM, whenever that does happen, if you could compare and contrast your opportunity for 3D DRAM versus your opportunity in 3D NAND. It seems to me that it's maybe not as juicy as 3D NAND was, at least for you, but can you just sort of give us some relative compare and contrast?

Yes, sure. Frankly, Tim, 3D NAND for Lam was bases loaded walk-off grand slam home run. We're both -- we got baseball kids, right? So I figured I'd throw that one in. It's still early days as the industry tries to figure out what 3D DRAM looks like. I'm super excited about 3D DRAM. And I don't know exactly what the process flows are going to look like. It's going to be very positive for Lam because it's a 3D structure. There's some complex etching in there, new materials and things like that. Whether it's as strong as the transition from planar to 3D NAND was, maybe not a grand slam walk-off home run, but it's going to be solid. It's going to be a solid transition for us. We're super excited about it. There's lots of debate around what is the timing. Our view of it is second half of the decade, you'll start to see it. It's being worked on actively as we speak. R&D dollars are being deployed, jointly collaborating with every one of the customers that's going to make the transition. It looks like a very good transition for us. I'm not sure it will be quite as strong as NAND was, but it's going to be very, very good.

Got it. Maybe we can come over to service and you said -- CSBG, I know this is one of the things that you like to talk about the most and rightfully so. The really unique thing about your business is that NAND is probably the stickiest end market because a customer is not going to go -- if they wanted to switch vendors, they would need to go rip out their entire installed base because you're just buying more of what you already have.

That's right. Incumbency is so important in this industry, and I think people really underappreciate that. I mean when you're the incumbent, you're the incumbent. It's -- yes, you don't rip off the installed base. That's a very expensive proposition.

Especially in NAND. And I think that, that sort of fuels your service business, too. And so I guess the question is, can you talk about service and maybe break it down into what's recurring and what's more transactional? And because, obviously, as you said, that particular piece of your model is running way above what you had thought.

Yes. Maybe let me -- let me step back and just for those that might be somewhat new to the story describe what we call it, CSBG, the Customer Support Business Group, or the installed base business. Let me describe what it is and I'll come back to your direct question. The first thing to understand, and I think oftentimes, people don't realize this is our tools will be productive for a really, really long time, decades, frankly. And so once the tools -- or I refer to chambers, right, our -- a lot of our tools have multiple chambers. But they'll run for decades. And so what is this installed base business that I talked about? There's 4 components that go into it: spare parts; service; equipment upgrade; and then what we refer to as the Reliant product line, which is refurbished equipment or increasingly, just think of it as the specialty technology node because it's not just refurbished equipment anymore. The demand there is so strong. We're selling new tools into that. So the totality of that is what is CSBG at Lam. And so your direct question, how much of it is recurring? A lot of it is. And it's not necessarily because it's all under contract, although a lot of it is. The biggest individual component of those 4 things I described are spare parts. And when utilization in the industry is high, the intensity -- the spares intensity is high as well, right? So as service for that matter because to keep the tool in a pristine condition to maximize yield and output, you have to do the maintenance, along with the volume. And on the utilization side that occurs more frequently. That stuff is completely recurring. If the fabs are running, if the tools are running, you must replace spare parts and you must do service on the equipment, right, to keep it in a pristine condition and maximize output. So I haven't given a specific number, but that's a lot of the business, those 2 components, right? I would describe upgrades as more dependent on where is the installed base at, where is each fab at. Sometimes there's a little bit of lumpiness to upgrades. And then the Reliant product line, right, the specialty nodes or think of trailing edge lithographic nodes, 28-nanometer and above, if you think about it that way, 11 consecutive record quarters of revenue there. And we have suggested in the past, and I continue to say the same thing, I think those legacy -- WFE in the legacy nodes or the specialty nodes grows faster than overall WFE. So when you put it all together, WFE has been very strong. So it's shipped a lot of chambers this year, frankly, and that's going to provide opportunity for the business to continue to grow. I described this as a business that grows every year, again, because chamber comp -- grows every year. Now it doesn't necessarily grow every quarter because there can be timing aspects to certain things and lumpy upgrades sometimes. But because chambers, really, they never go away, they just get repurposed, this is a business that should grow every single year. And we have strategic intent -- and we've talked about advanced service. Think of it as guaranteeing some kind of a result where the dollar per chamber in the field grows over time because of our strategy around how we deliver some of these advanced service offerings. And the fact that things at the leading edge, frankly, are somewhat more spares-intensive, again, because it's just -- the chamber needs to be that much more pristine at the leading-edge nodes. So when you put all of that together, that's -- those are a lot of the reasons why this business has just been so strong this year.

And Doug, what do you think -- is 10% to 15% the right long term -- I mean, obviously, 30% year-over-year. But what's the right long-term number?

It's a hard answer for me to give Tim because it depends on what happened to WFE in the year before it, frankly. Clearly, we're ahead of the model. And again, like some of the other things you've been asking me about, we probably need to update the financial model and give you a sharpened pencil on this as well. And again, we'll do that.

And I guess last thing on service, do you feel that the market is beginning to value your service business? I mean it used to be that nobody would ask about service. And now people do ask me about service, and I know that they ask you about service and you like talking about it. I mean some of these sort of toll-taking models trade at very, very high multiples. I mean very high, much, much higher than your stock. So do you get -- this is more of just sort of like a touchy-feely question in the sense that you get the feeling that investors based upon their questions are beginning to value that piece of the business. Are they willing to value the stock maybe on sum of the parts? I'm just wondering what your sense is from talking to investors and the tone of their questions.

I think, Tim, we first began to disclose this externally every quarter. I think it was March of last year. I'm talking about it a lot more. I get asked about it more than I used to. But even still today, frankly, when I have an investor meeting that's 45 minutes long, it's the last question that gets asked oftentimes. And sometimes I have to say, "Hey, you forgot to ask me about this." And so I'm talking about it more. I'm reinforcing it a lot more. I think the fact that we are now reporting revenue for it every single quarter, it's getting better. But I still think people underappreciate the quality of this business, which is why I talk about it all the time. In many ways, it's my favorite part of the business model, although don't tell my etch and deposition guys that because I don't want them to think I don't love their business because I love their business, too. But this is really -- it's a fantastic business model, grows every year, very sticky. It is very cash generative. It doesn't require anywhere near the level of investment of like selling the new equipment as it benefits from the design of the new equipment, obviously, because all the parts are designed where the tool itself is designed. But yes, I still continue to feel like it's not completely appreciated. The quality of business isn't completely appreciated. So I just keep talking about it and encourage you to keep writing about it. It helps people to understand. It's a very sustainable part of the business model.

Got it. So maybe we can pivot to capital return, which is another thing that you love talking about too. So for the past, I think it's 14 quarters now, you've been returning more than 100% of the free cash flow, well north of your model. This is, I guess, on the bane of all these questions I'm asking you about the model as well. But that has suck the cash down now. You used to have $7 billion. Now you have a little bit under $5 billion. And I think you, in the past, said you don't want to go below $4 billion. I don't want to put words in your mouth, but I seem to recall that being sort of the level that you feel comfortable with around $4 billion. So how should we sort of think about the return plan going forward? How much cash do you need? And I guess part of that is a question about capitalization, but I -- well, I'll ask that as a follow-up.

Yes. Okay. Yes. I think we've been very judicious with the balance sheet. Our priorities are always going to be, first, you've got to make the right investments for the business and make sure that is well funded and well financed. But even after doing that, we generated a lot of cash, obviously. And so the model, 75% to 100% of free cash flow is what people should rely on. And I know people like the dividend growing on an annual basis. We're committed to doing. In fact, not that long ago, we just raised dividend, about 15%. And you're right, over the last couple of years, we've done a little above 100% of free cash flow. We have pulled cash down. I'm pretty comfortable where the balance sheet is today. We've got a very well-termed-out debt structure when you look at it. We've been pretty thoughtful, have been very thoughtful about kind of stacking the maturities out in a way that makes sense. I think our average duration is north of 10 years. It might be closer to 15 years actually. Last year, we raised $2 billion of new debt for the first time in history of the company. We actually put a 40-year note out there, which is great. I think debt investors really have begun to appreciate the quality of the business. And so yes, I mean it's 75% to 100%, Tim. Every once in a while, we try to be opportunistic about how we do it. It's not the same dollar every single quarter. I think we've done a pretty good job since we brought Lam and Novellus together in terms of how we've executed this. I think the average repurchase price is like $140 or something over that time frame. So I feel pretty good about what we've done, and I'll stay focused on it. But I'm not here to give you a new commitment or a new number. I'm pretty comfortable with what we've described and how we're managing it.

Got it. But I guess the message is that you don't feel like you're sitting on a lot of excess cash to where you were maybe 2 or 3 years ago.

One of the big things I always remind people is, too, that helped us immensely was when tax reform occurred in the United States. It avail-ed us to all of that cash, and a lot of our cash gets generated outside the United States. It's where the customers are, right? Think about where the fabs are, where the cash gets generated. And it was challenging at times under the old tax regime prior to 2017 because you had to repatriate -- you pay taxes. Right now, I have to pay tax regardless. And so that has helped us do exactly what you're pointing out that we've done, which is pull the cash on the balance sheet to a lower level.

And I guess, Doug, just you did put out some new debt. But what do you think the right model is for the company? You're basically net cash neutral right now. But if you look back last cycle, I mean, even during a huge memory downturn of a few years ago, your revenue barely declined year-over-year. I mean I think it was down year-over-year for one quarter, maybe it was 2 quarters. It's a very, very stable company. And it seems like you could run with a fairly significant level of net debt, not that I'm suggesting levering up the company to a significant degree. But it seems like net cash neutral, is that the right capitalization for the company? Would you like to be more net debt?

Like I said, I'm pretty comfortable with the balance sheet as we sit here today. It's not too far off what you -- if you go look at the entire semi industry and the equipment space, it's kind of where you see a lot of companies at. Not to say that, that means that's the right answer. But I'm pretty comfortable with the balance sheet where it is. I'll be -- we're thoughtful about it. I -- my treasurer and I talk about it, debate it all the time. But I feel pretty good about where it's at. Like you've observed, we pulled cash down to a certain extent. We've termed the debt out, just like I said last year, raised a couple of million dollars from incremental debt. I feel pretty good about where we're at, frankly.

Got it. Time for a few more questions. One is coming in from an investor. And just, Doug, you and I have been around this industry for a long time, and there certainly seems to be an argument that the cyclicality of the industry, I think, in some respects, this is true; but in others, it isn't. But the question really is around the cyclicality of your. Business, and it harkens back to what I just said, when we went through a huge DRAM cycle and -- or memory cycle and you were fairly down year-over-year. So do you think that the industry is becoming less cyclical? Or are you concerned that when we come out of this cycle because of all the things that have happened, all the shortages and all the getting in line and that there's going to be such a digestion phase that we do worry that you could have a down 20, 30 -- I mean, some like big, big numbers you just talked about.

Because it does. But the natural cycle is just very different than it was. If you've been following the industry, as long as you and I have been in the industry, right, consolidation has mitigated that. Diversity of demand for semiconductor, frankly, has made it less volatile when it used to just be PCs. You had one thing up and down. You had PCs, phones and everything going on in the cloud and emergence IoT and different things driving demand has mitigated that a little bit. The consolidation has mitigated a little bit. Rising capital intensity, frankly, I think, has mitigated it a little bit because it's just too expensive to put too much capacity in place if the industry doesn't have an eye towards selling it, frankly, right? I mean, back in the day, you used to have everybody chasing the same market share, and it created part of the cycle. Those things are all fairly different. And then when you think about our business, you've been nice enough to ask about CSBG. But as CSBG has grown, that also dampens the up and down here, right, because CSBG just comes along every single year regardless of whether WFE might be up, down or sideways. So when you think about all of those things, I think the nature of the cycle, this is very different than it was 10, 15 years ago.

Got it, Doug. So I want to ask about logic. This is something that we used to talk about a lot and just the idea that you had done a lot better sort of from like a PTOR point of view, you have done a lot better in the subsequent logic nodes. But then, of course, those didn't transpire on the same time line that we thought that they would. But it certainly seems like that story is getting kind of back on track. And so can you just talk about logic generally in terms of -- people always talking about NAND and DRAM for you, more so in NAND, obviously. But can you talk about maybe logic and maybe some of the per wafer opportunities you have at these more advanced 5 -- foundry 5, logic 7 nodes versus the last time logic was a big piece of the mix where you didn't have much of a position? And then maybe you can touch on foundry as well because your share of foundry WFE seems to be inching up as well.

Yes. Both are incrementing because of the reason I described earlier in our conversation. Every process node, our addressable market gets bigger than the one before it. And I know you know this, Tim. I'll remind everybody on the webcast here, last quarter was an all-time record, another category. And I think 2 or 3 quarters ago, I said the same thing about foundry. So you have a very strong memory, but we are making very nice progress in foundry and logic, again, because the addressable market gets bigger. And then, yes, you remember the story about the North American logic customer that we've been describing for, I don't know, it's been a long time, but you're now seeing in the numbers -- showing up in the numbers, right? And the story there is as we went from 14 to 10, 10 to 7, it wasn't just the addressable market getting bigger, there's a market share story because of differentiated capability technology in our conductor etch capability. When I shared the story, it's showing up, right? So it's not just the market is getting bigger there, like it is in foundry, same kind of dynamics driving both foundry and logic, obviously. It's similar process flows, obviously, but the market share story, and that's for us, is very much happening and going to continue.

Got it. And then I guess last question, Doug, really about NAND and about the cost curve in NAND. You've become very levered to NAND. And as you said, it was a [indiscernible] plan for you. But as you look out a technology road map where it's -- that the industry can scale with layers for a long enough period of time that you don't have to worry about it from a risk point of view, that if the industry does move to an alternate technology or an alternate process architecture or whatever, that it's beyond the limit of you running the company that you're really worried about at this time.

As far as we can see, there's not an end to the scaling, right? It's a very long road map in front of us in terms of where demand customer base is going. So we can't see into it, frankly. I was going to say there's always engineering challenges as you go from one node to the next. And as we go from single stack to double stack, there's new challenges inherent in different process flows, but we see a very long road map in front of us, Tim.

And you're -- and just to be clear, so your WFE -- so your opportunity per wafer in a string stacking world where we can almost not infinitely stack these 64-layer strings, but I mean -- but it's become a modular game. Your opportunity per wafer is similar in this new modular world that it wasn't a single-shot world? Is that correct?

That's exactly what we've been saying. And yes, you remember how we described it. If you think about it, as layer comp grows, you're depositing more film on the wafer. That's what we do, right? Those are our applications. That very complex, difficult etch down through the channel hole or memory hole depending if -- I mean people call it different things, we own all that. Metallization, we pretty much own all of that. And so when you think about the stack getting bigger, everything I just talked about, our strong market share, it's just -- there's more of it, right? And I think one thing, Tim continues to talk about -- our Tim, Tim Archer, cumulatively, 45 million more wafers have run through our tools than the next nearest competitor in those 3 critical applications. So we just get better at doing this because we're getting all the volume going through.

Great, Doug. Well, thank you. We've run out of time, but super, super appreciate your time.

Good to see you, Tim, and thanks for having us.

Thank you.