Lam Research Corporation (LRCX) Earnings Call Transcript & Summary

Hey, everybody. I'm Joe Moore. Glad to see you guys again. I'm going to be on stage literally all day today, so you'll get tired to me, I'm sure, by the end of the day. But very happy to have with us today the CFO of Lam Research, Doug Bettinger. Doug, I think you have a safe harbor you want to read and then we'll get going.

Yes. I always like to keep my lawyers happy and Ram too, by the way. So I'm going to read the safe harbor real quick. Today, discussions may include forward-looking statements that are subject to risks and uncertainties and actual results may differ materially. Additional information concerning factors that could cause actual results to differ materially from these forward-looking statements can be found in the risk factors disclosed in our public filings with the SEC, including our most recent 10-K and 10-Q. And by the way, I don't intend to say anything new today, so I didn't really need to read that, but I did anyway. So we can get started, Joe.

Great. Well, maybe you could just give us a little bit of a summary of where we are. You guys last year saw your biggest market contract a lot with the NAND...

A lot. A lot.

In NAND and yet, you had a pretty good year when all said and done. Can you just kind of talk to the priorities that you guys have? And we'll get into some of the details.

Yes. No, Joe, maybe it will help. I'll start at a very high level, paint a picture of last year, this year, what we see. But you're absolutely right. I mean last year was a very challenging year, certainly in memory in terms of investment. Frankly, I haven't seen things down as much as they were last year in my entire working career. So it was pretty dark, pretty bad. Little bit better this year. I mean, you're starting to see some green shoots, and we can -- I know we'll talk about that as we go through. But I think when we look at WFE last year, our assessment was it was kind of low $80 billion. As we look into this year, I think we're mid-high 80s. Within that, obviously, we'll talk about this, but leading-edge, foundry and logic is quite strong. And by the way, we are making really, really good traction, gaining new positions in that segment of the business, which is a large reason, Joe, why you saw how well we did even with memory down so much last year. Now having said that, we do see memory up a little bit this year, DRAM up a little bit, NAND up a little bit. But by no means would I suggest to you, we see a recovery. I wouldn't call it that. But I do think we're beginning to see some green shoots. So that is -- the foundry and logic positions that we've gained were a big part of why things were so strong last year. In addition to what I call the customer support business group at Lam, which is comprised of spare parts, upgrades, service and then the Reliant product line, which services more of the specialty market. That actually was reasonably strong as well last year. And we were pretty pleased, Joe, when we came to the end of last year. We always give the number of chambers in the field. It eclipsed 90,000 chambers exiting the year. And we give that number every year because it defines the opportunity for us to continue to grow that installed base business. So we can talk more about that as we chat back and forth here.

Okay. Great. One of the debates really has been around the geography of spending. You've seen a very robust China at -- in the -- above 40% of your revenues in the second half. And I know you've talked about that moderating a little bit. Can you just talk to that dynamic? And how strong are the geopolitical forces there? And how do you balance out the desire for self-sufficiency, which could imply that we spend for quite a while, versus the fact that there's still an incomplete product portfolio to put into those factories?

Yes, maybe I'll paint a picture of what's going on in China. That's always a topical question for everybody. Yes. China in the second half of last year, our business geographically in China was quite strong. There were some dynamics that caused the second half to be stronger than the first half there. But you're right, it was 40% plus in the second half. And I continue to see a relative strength in that region of the world in the beginning of this year, although I do believe it will moderate as we go through 2024, Joe. You've got a broad set of customers spending in areas where they're allowed to spend, right? Obviously, there's technology restrictions that you don't ship to, you can't ship to. And we're Uber careful about making sure we're compliant with all of that. You can't run and follow those things. But anyway, it's a broad set of customers, and I would describe it to you as some in memory, DRAM, in particular, was strong in the second half of last year, continues to look strong in the first half of this year, will moderate as we go through the year in China specifically. And then a large tail of what I would describe as customers investing at trailing edge or specialty node technologies, it's not one or two. It is a broad set of customers aspiring to win business in areas where there's demand in China. I think largely, you see a lot of the investment from our China-based customers striving to supply-demand in China with some level of success. They're going to get better and better and better as time unfolds as they ramp these technologies and learn how to run volume. I would suggest to you, I mean, a lot of people will tell me, "Hey, these are older technologies. This is easy stuff to do." That's not true if you haven't done it before. And so I see a lot of this investment targeted towards more mature nodes for sure, but with customers that are learning how to manufacture and ramp these technologies. So that's kind of what I see going on, Joe.

Yes. I mean there's a lot of areas you can innovate at the trailing edge. We were talking about silicon carbide coming in and [indiscernible]...

Yes, that's an example. I mean there is a tail in China of silicon carbide investment, right, and we're...

We debated about what those [indiscernible].

Yes.

But the -- yes. No, that makes a lot of sense. I guess when you talk about the moderation this year, is that simply that I know there was some inflation of the second half versus first half last year because of export control earlier in the year, there were things you could ship in the second half, so you had some catch-up in error. Is that the moderation? Or is it more than that?

I think what I would describe, Joe, in China specifically is year-over-year, it's pretty steady, consistent. Last year, because what you were mentioning was somewhat second half weighted. We were unclear as an industry, hey, can you shift to that technology node? I think we all sorted that out and realized, yes, we can ship. So that was second half weighted for that reason. This year looks a little bit first half weighted, specifically just timing of investments of the customer base. Year-over-year, though, I'd describe it as steady, consistent is the way to think about it. It's not precisely the same number, but it's an investment that continues. And I often will get people saying, "Hey, when is this going to stop spending?" I don't see that happening, right? I believe there's sustainability of this. It doesn't mean it grows every year. Nothing grows every year in this industry, but it's not going away from everything I see today.

And just the last question on this. I know you -- it's very difficult to predict any future government export controls, things like that. And I won't ask you to. But can...

Good. Because I don't have a crystal ball, Joe.

But like what is -- what's the -- is any of that speculation impacting near-term activity? In other words, do you think some of that spending is because there's a fear of future export controls? Is any of your behavior changed by the fact that there could be? Or is this kind of we're sort of assuming this is more the status quo until told otherwise?

It's that as far as I can tell. No, I wouldn't know is somebody doing something more quickly or not because they're worried about something, hard to handicap something like that. But we're just executing consistent with the rules that are in place today. Could something change? Of course, it could. Do people speculate about that all the time? I don't have any new updates from anything that I know about, Joe, to talk about today.

Yes. But clearly, geopolitics has been good for the business in terms of the reshoring and s*** like that.

It feels like every region of the world, there's government money beginning to come into this industry. Listen, it, to me, is an acknowledgment from all of these governments that, hey, this is a critical industry. Right? I mean semiconductors, I think the world woke up to the fact that we're critically enabling to almost everything going on in the world today.

And that's the case from subsidization and it's also the case from the export controls, like that was the effective export control mechanism just to use U.S. equipment.

I think that's right, Joe. Yes.

Maybe we can go into some of the markets. DRAM, you talked about getting a little stronger. I mean it seems pretty strong now. I know there's a China component to that and an advanced packaging component to that. But it seems like the underlying spending has already kind of come back quite a bit. Can you just talk to how you see that environment?

I guess when I look at it, we're still in a cycle, right? Don't run to the conclusion that because spending is ticking up a little bit, that the cycle is over. It is not, right? There's still inventory out there from everything I can tell. But there's a product cycle in here, right? You've got DDR4 going to DDR5 because all of these new server CPUs are out there that require DDR5. Additionally, right, everybody is all bulled up about AI and I am too, low-latency DRAM is critically enabling for all of that compute architecture. And it is packaged to go. I think of it as packaging, but it's really die-to-die interconnect using a through-silicon via process that we have extremely strong positions in. I've begun to call it drill and fill, creating the space to have the die connect together and then depositing the conductive material with an electroplating process. We own those applications for the most part. So extremely strong positions. It's very enabling to what's going on there. And it's an area where there is a real cycle -- product cycle going on that's trying to get pulled forward. So I'm excited about that. But don't run to the conclusion that because there is some of that cycle in there that the broad inventory cycle's over. It's not. On top of that, Joe, there's one specific customer in China that is -- like we talked about a little bit earlier, was the second half weighted spending last year and looks to me right now to be a little bit first half weighted this year. So when you put those 2 things together, you've seen somewhat of an uptick in DRAM. You haven't seen that in NAND quite yet.

Yes. Okay. And the HBM opportunity seems pretty exciting. And you talked about owning some of the key applications. Applied has put some numbers around that for this year. I'm not going to ask you do that, if you don't want to, but can you just talk to your opportunity versus supplies in there?

Yes. We have put numbers maybe a little bit roundabout way related to this. We've grouped together all of the advanced packaging stuff in terms of some of the things we've said. So I'll go back and tell you what we said there. Before HBM started taking off, I had been describing the advanced packaging "opportunity" is hundreds of millions of dollars of business for Lam. That was before HBM took off. As that began to show up, I started describing it as business that I can easily see as $1 billion of business in a several-year time frame. But more recently on the earnings call, my CEO, Tim Archer, described it as business that we think is going to nominally triple this year. So when you put all of that together, that kind of helps you in your mind size how big this is for us. It's a $1 billion opportunity is how I think about it, driven by HBM, but also driven by some of the advanced packaging, things like [ co-loss ], as an example, or [ Fubross ] or other things. But we put that all together and largely sell a lot of the same stuff.

Would you say you're equally weighted to both of them, to [ callus ] versus HBM or is there -- I think of you as more memory, but...

Very similar stuff, right? It's electroplating and the deep silicon etch. It's a tool we call SABRE 3D in Syndion. Do you want to go to our website and have a look at what these things are?

Okay. Okay. Great. You mentioned NAND stays low. Everybody that we talked to yesterday said no real signs of any kind of pickup there other than maybe a little bit off the bottom.

Certainly pricing's get a little bit better but by no means are -- is...

Look, not to make it about me, but I'm very bullish on NAND when you see this level of spending. But like I feel like the investor sentiment is just like this is never coming back. I mean can you just talk to -- maybe, of course, it will come back some, but the levels that you saw a couple of years ago, do you think that's the next peak? Do you think that was too high? Like how do you size that?

Listen, I get asked this question all the time, which is, "Hey, is NAND ever going to get back to the investment levels it was several years ago?" And my answer is I absolutely believe it will, and it will eclipse it, if you're a believer in a semiconductor industry that approaches $1 trillion at some point. And I do believe that. I'm not going to put a time frame for you. I think everybody -- you can read all the reports that are out there. If the industry grows to that level, and I believe it will, NAND investment is going to grow beyond the previous level that it was at. I don't know what year that's going to happen. But I am a firm believer in the need to store data. At the end of the day, that's what NAND flash does, it stores data. And data is exploding at exponential growth rates and obviously, is going to be required as we go forward.

And 2 of your 5 customers are struggling with balance sheet issues and the other 3...

You can't invest if you don't have cash. So that's an important thing to understand as well.

Yes. And you end up where you're diverting spending to DRAM. So I mean, because of AI and HBM and some of the other opportunities. So seems to me a very bullish backdrop, understanding that utilization needs to come up and there's some headwinds we may have to do.

Everybody gets all excited about AI and certainly, I do too. If you look at, I don't know, an average AI server, maybe just allow me to describe it a little bit, there's 8x the DRAM content in an AI server versus an enterprise-class server. There's 3x the NAND content. And I don't know, I think the square millimeters of logic silicon is nominally 4x. So that's exciting, right? You need a lot more equipment to generate that level of output.

And the hyperscalers are very focused on having low-latency storage at the edge, and there is a lot of activity around that. So it's not the same as HBM. Like I get 5x the economics or something, but it definitely seems pretty compelling to me. Anyway, enough about me, but it's important for Lam as well. Can you talk about the competitive environment? Tokyo Electron has been uncharacteristically kind of noisy about the cryo etch.

Actually, for a long time, they've been uncharacteristically noisy.

Yes. But normally, they don't sort of talk about that. So just like obviously, you're very dominant in those key process steps. But can you talk about might you see more competition going forward even if it's just at the seams?

I guess the thing that I've been describing is, at the end of the day, when you've got a very strong position, will a customer look at somebody else's equipment to keep you on your toes? Of course, they will. Is that happening here? Maybe it is. Have any production decisions been made at this point. No, they have not, right? And what I would describe to you is listen, this particular application that they talk about, we have owned for over a decade. Right? Since the advent of 3D NAND, we have 100% of the application share of this, and that equates to over 6,000 chambers in the installed base. As investment begins to come back, which will be dictated by the market environment, the first thing that's going to happen is the installed base will get upgraded, right? I mean it's sitting there. It's the most economical way to get incremental output. I know for sure that's what's going to happen. Is the majority of the spending going to go to that? Of course, it is. Does that mean that the customer may not once in a while look at somebody else's tools to keep us on our toes? Like I said, they'll do that. But understand the majority of the spending is already locked into our installed base.

I feel like every dominant franchise in semiconductor equipment, there's always an alternative that people get noisy about it from time to time.

It's a dielectric etch application. The other competitor is a strong dielectric etch provider. But yet, we have the biggest application in the industry. So that's probably why they're talking about it.

And can you talk to -- when NAND spending does come back, I tend to think of Lam as capturing a higher proportion of sort of layer count increased technology spending versus adding capacity. Does that come back first?

Yes. So what we've described and what I believe will happen is, obviously, right now, utilizations are pretty low. The first thing that will happen is utilizations will come back up. And when utilizations come back up, that will show up in our business in spare parts, in service and then closely followed or maybe concurrently, the equipment that's there will get upgraded right? That's the way to get output most affordably, most quickly, if you will. And so obviously, that's going to benefit us because we're extraordinarily strong in the installed base. I'm not seeing that yet in a meaningful way, but that's what's going to happen first before any new wafer capacity gets put in place. It's just much more economical to upgrade what's there and to turn utilization back up from the low level set it's in.

Okay. Great. So there's one Chinese NAND vendor, YMTC, that's on the entity list. And I know you're not doing business with them, none of the U.S. companies are. But they have been pretty vocal that they're adding to wafer starts this year through some combination of buying from international tool vendors and China local vendors. Can you speak to that trend generally? Do you think -- does that enable competition in the Chinese market? And do you think that -- I mean, I don't think anybody saw China's DRAM coming back the way it has, is there a possibility that either from them coming off the entity list or other people in China that start to make these investments that there might be a Chinese opportunity for you?

Listen, we're not doing any business with that specific customer. I don't expect that we will unless something changes. Might something change? Nothing I'm aware of, but it's always possible. And yes, somehow, they're continuing to generate output, right? It might be unintended consequences. I think probably not as effectively as they would have been able to had we still been able to do business with them. But understand also that, Joe, and you know this, but it's still a fairly modest amount of the total global supply coming from that one specific customer. But somehow they are yes, getting output. I don't exactly know how, because we're not there any longer.

I'm less concerned about that supply and more just the notion that Chinese equipment companies are starting to ramp, that they're -- you're sort of forcing these companies to use Chinese equipment and maybe enabling some of the competition over time.

Over a period of time, potentially. Yes. Clearly, the global leaders in the industry, we're very far ahead of kind of the emerging set of customers. But over time, will they get better? Of course, they will.

Okay. Shifting to the logic market. Maybe starting with the trailing edge, which has been super strong and is now moderating.

It has been. Except in China. China is still quite strong. We already talked about that though.

Yes. But let me talk to that dynamic a little bit of trailing edge. And to me, there are sort of countervailing forces. There's a secular, we have to spend more at the trailing edge because we're not waterfalling the cutting edge nodes down, we don't have nodes anymore, but there's also an oversupply that we're wrestling with in areas like analog microcontrollers. Can you can talk about those dynamics?

Yes, sure. I mean when I look last year, investment at those specialty nodes was very strong. It's moderating to a certain extent. I think everybody understands what's going on there. The industrial and automotive stuff got in the semi cycle. It just -- it happens. And so in the mature economies or perhaps outside of the Asia regions, that's pulling back a little bit this year. They're still -- when I look at it and think about it, in fact, it's some of the stuff we were talking about here before we opened the mic, Joe, there's secular growth there. Over the through-cycle secular growth in industrial, automotive, image sensors, pick your favorite specialty node, silicon carbide other things. That does require incremental investment over time, but it's not immune from the cycle. As much as I wish this wasn't a cyclical industry, it is. It's a growth cyclical industry. And so that's what I see going on in that segment of the business right now.

And how much of that business -- or I want to ask about services in a bit, but how much of that business gets categorized as Reliant and part of the services business versus what you're counting as foundry?

It's a little bit of both. I would say a lot of it is in the Reliant product line, which, again, is 1 of the 4 components of what we call the Customer Support Business Group or CSBG. But there is some new equipment there, too. I mean some of the things we do there as well does take advantage of the fact that we're making incremental investments in the leading edge and benefits from that at times, but a lot of it is Reliant, Joe, you're right.

I mean the economics of Reliant used to be good because you had a lot of refurbished tools in there. And then there were no more tools to refurbish. So maybe now that will come back and you'll get some of that.

That is a true statement. Yes. Historically, Reliant was "We buy used cars back. We refurbish them and resell them." That's still an opportunity over time, but you're absolutely right; there really isn't any used equipment available any longer. At least nothing in a meaningful way. So we are selling what I describe as old new equipment if that makes sense, old models, but new equipment.

And then at the cutting edge, we're having debates now that we haven't had for 10 years about which of your customers is in the lead on gate-all-around.

I'm not -- You're not going to get me to talk about that.

No, I'm not going to take you there. But I mean it seems like the fact that there's an Arm's race there, it's a good thing for spending levels and there's a lot of enthusiasm to move to those nodes as quickly as possible. So can you talk about that? And then can you also talk about Lam's ability? You mentioned doing very well in some of the newer technologies. Lam's ability to monetize that?

Yes. So yes, we've begun talking about these big new opportunities, specifically in -- a lot of it in foundry and logic. So the things we talk about is gate-all-around. So we've described gate-all-around numerically as for every 100,000 wafer starts of capacity put in place, it's an incremental $1 billion opportunity, not all of which will we win, but it's -- that's how big our incremental addressable market grows by. And it's areas taking advantage of our ALD deposition process as well as selective edge. So that's one, gate-all-around, of which different customers will invest at different time frames. Also backside power, right? The same thing. It's an incremental roughly $1 billion opportunity for every 100,000 wafer starts there. I almost think of these things as 3D structures. And when things go 3D, like NAND did a decade or so ago, we do well because you're depositing material, then you're moving material through an etch process. Typically, that's how these structures are put together. Backside power is sort of like that, at least when I think about it. You're moving the power delivery to the backside of the wafer of the die, if you will. That's our SABRE product, other things too, but specifically SABRE, which we've been extremely strong, I don't know, for a decade or longer. So that's an incremental opportunity, largely foundry and logic driven. Then the other one that is perhaps a little bit further away, but we've been talking about it and investing in it for a while is dry photoresist, right, to help enable the photon absorption from EUV. I'm very excited about this because this is a brand-new opportunity for us, something we didn't participate in before. Our excitement of this, I had my CTO last summer, I think, at SEMICON West with a group of investors, and he described it as it's a question of when, not if. Every customer that uses EUV has our hardware in their lab, evaluating how this works, how they want to have us change certain things. You don't see that kind of pull for something unless there's a pretty compelling value proposition, right? Lab space at the customer facility is highly valuable. They don't allocate space to you unless they see something of benefit. This one is perhaps a little further down the road, but I'll remind you, we describe this as a $1.5 billion opportunity cumulatively over a 5-year time frame, in that 5-year and beyond approaching that $1 billion opportunity. I like talking about these $1 billion opportunities to make things easy. This will be a $1 billion opportunity a little bit further down the road. So we're extremely excited about that one as well. The other one, obviously, we've already talked about, the $1 billion advanced packaging opportunity. So there's 4 things that I kind of put these same numbers around to make it easy and hopefully resonate in your mind of our $1 billion opportunities as we look forward.

And what are you seeing at these cutting-edge nodes? Because you've had -- we actually have seen some factory pushouts, Intel's Ohio, TSMC talked about a second fab in Arizona, moving a little more slowly, Samsung, Taylor. At the same time, a lot of that is just geographic like TSMC seems to be emphasizing Japan more than Arizona, things like that. There's also this question of who's the pipe cleaner for these initial nodes. It was Huawei and Apple, then it was just Apple. Now it seems like on 2-nanometer, Apple doesn't necessarily won't commit that kind of volume. At the same time, there's a lot more people that want to be...

They've got the AI group putting these [indiscernible].

Yes, all the AI guys are suddenly like in the first year or second year. So can you just talk to those dynamics? Does that -- it seems like that creates a pretty good environment for you guys.

Yes, it does. Everybody gets excited about all this government money moving around different places. What I remind people is, though, temper your excitement a little bit because what truly matters more than anything is the demand for leading-edge semiconductors, which is sort of what you were mentioning. Is it mobile? Is it enterprise compute? Is it AI compute? At the end of the day, that's what matters relative to the investment in wafer capacity. Now as this becomes more geographically dispersed, you will need more equipment, right? Things to broaden out, it won't be as efficient as if it were in one place. And the timing of this stuff, it will push in, it'll push out, it'll ebb and flow depending on just different things, be it facilitization timing, be it where demand is showing up, when government money is becoming available. A whole variety of things, Joe, will have it kind of move around. But at the end of the day, the investment will occur based on demand for leading-edge silicon. Full stop. At the end of the day, that's what matters. So don't lose sight of that. Oftentimes, when I sometimes get confused about, hey, that timing is moving up, down or sideways, if you step back and think about the demand for silicon and semiconductors in the end markets, it'll get clear. At least it does for me when I step back and think about that. You've got to remind yourself of that at the end of the day. The rest of the stuff is just timing.

Yes. Yes, okay. Makes sense. I know your favorite topic is services. Can you talk to that? You saw a contraction last year around some of the export control utilization things. Kind of unusual, but can you talk about the growth prospects for that business?

Yes, for sure. I guess, let me step back. We give a number that's important to listen to in every December call, which is the number of chambers in the installed base. The important thing to understand is that it grows every single year. I think sometimes that surprises people, "Well, why doesn't? Doesn't stuff go away?" No, it really doesn't ever go away. Our chambers will be productively deployed in the installed base of the industry for decades, literally. We have things that are there for decades. That's important to understand because over time, the opportunity for us to sell into that installed base on average is at least as high as it is when we initially sell the equipment. And that shows up in spare parts, in service, sometimes in what you were talking about earlier, Joe. We'll repurchase things, refurbish it and resell it. Although today, there isn't really anything available and upgrades. And so when you think about those things, there's a long tail of opportunity for us to do things for the customer that delivers a decent amount of value to keep the installed base productively deployed. So that's important to understand for CSBG. I used to describe it as a business that should grow every single year. On average, that is still true. But that Reliant product line last year got so big, that will be exposed to WFE, obviously, right? It's the specialty nodes, the trailing edge node, WFE will cause that to ebb and flow. And that has gotten to be a fairly large component of CSBG. That won't grow every single year. In fact, this year, it's probably pulling back a little bit, offset to a certain extent by improving utilization and things like that, that's going to pull spares, service and eventually upgrades through.

Yes. I mean I understand the business is segmented that way because the business has operated that way.

That's how we run it. One General Manager, this is how we run things internally. And so I just disclosed to the investment community, the same thing we look at, Tim and I look at internally.

Every quarter, I'm trying to extract a Reliant number out, so I can understand what the [indiscernible].

Have you figured it out?

No. If you just give it to us, that would make it easier.

Yes. I'll think about that over time.

Okay. That makes sense. And then you talked about the opportunity to grow the dollars per system in the installed base.

That's still the objective, right? So the installed base, and again, we give you the number every year, our objective is to grow dollars faster than units, if that makes sense. Units is kind of a strange concept here, but chamber count, I guess, is the way to think about it. And the way we do that is increasingly trying to understand how we can deliver incremental opportunity from a category of things I describe as advanced services, right? It's more predictive type service as opposed to engineer on site, show up and do a task. That's historically what services look like. Increasingly, though, what we're beginning to deploy and actually beginning to get real pull-through is using things like cobots. Tim began talking about cobots a couple of quarters ago, more consistent service and better performance is able to be delivered by using the cobot as an example. There's different predictive type things where instead of show up and do a task, you predictably do something for the customer. That's beginning to be a more meaningful opportunity and it's great for the customer because it delivers real value for them. It's good for us, too, because theoretically, you should be able to get paid in a reasonable fashion for that. And that's part of how we deliver growth in dollars over and above growth in units.

Great. So I have a couple of financial questions. Last one and then I'll open it up for the audience if anyone's got [ one ]. Can you talk to gross margin? You've had a lot of puts and takes there. It seems like China's probably good for that. You've also moved their manufacturing footprint to more centralized Malaysia. Can you just talk to those dynamics?

Yes. I guess -- let me see. Could ramble on about this for the rest of the time, but I'll try to keep it succinct. People will always ask me, hey, how do I think about the long-term financial model? And admittedly, it's a little bit stale. The model we gave was literally about 4 years ago. So I need to update that model, but it's still the right way to think about things. And we gave it just before COVID#, March of 2020. Embedded in the financial model is still the right way to think about it was a gross margin in the high 40s, call it, 48% sustainably. And then if you think back to the time frame, we went through a period of several years of just crazy inflation. Inflation never really goes back. It may slow down, the rate of increase may slow down. So that changed things. Secondarily, as business began to downtick. Last year, we took the opportunity to maybe accelerate a pivot to the Asia region, both in terms of our lab footprint as well as just the factory footprint, right? We've built a fairly large factory in Malaysia that as business begins to grow again, whenever that happens, increasingly, the incremental volume will show up there. The benefit from a cost standpoint there is labor for sure, but more importantly, a more affordable supply chain, shorter flight routes, right? The entire industry continues and will continue for quite a long time to be an Asia-based industry, independent of the fact that, yes, some people are onshoring and whatnot. The industry is still an Asia-based industry. And so having manufacturing in a bigger way there will save us a reasonable amount of money in terms of freight, logistics, things like that. So offsetting some of those inflationary headwinds over time will be a more affordable cost footprint as, like I said, business grows again. So that's in the longer term, how to think about it, that those high 40s numbers are still the right way to think about it in the long term. Now let me step back and peel the onion back. In the near term, gross margin has been at 48% nearly, and that's where we guided the March quarter, largely because of very favorable customer mix. That is not sustainable I don't believe in the longer term. That will mitigate to a certain extent as we go through 2024. So in the near term, we're at those numbers, we won't be able to maintain it, at least not in the short term. I think in the long term, we will. That's the right way to think about it. Anyway, I rambled on a lot about that. There's a lot of puts and takes in there. That's the way to think about it.

If there's any questions from the audience.

I see a mic in the back if there is? It's too early. You got me up here [indiscernible] and everybody didn't have their caffeine yet.

Can you talk about -- I mean, you guys have a lot of free cash flow. Can you talk about how you think about uses of cash?

Yes, sure. In fact, maybe a little bit surprisingly, 2023 was an all-time record year for Lam in terms of free cash flow generation, in a year where business was meaningfully down. Now that happens when business contracts, working capital [indiscernible] cash. So that's kind of what you saw last year. More coming out of accounts receivable. Inventory is still sort of sticky in terms of the level that it's at, but that will take care of itself over time. But yes, I was super excited about the free cash flow generation of the company. So how do we think about uses of cash? I guess, first, the primary thing we're going to do is invest in the business, right? I think that goes without saying. We see opportunities to continue to improve the product line, to continue to invest for these technology inflections, that is always, always, always going to be the first thing that we do at the company. So even with that, though, we generate more cash than the business needs from an investment standpoint. And so I think we've got a fairly generous buyback program. The intention of the company is to return 75% to 100% of free cash flow on an annual basis. And we've done that. I think last year, I forget the exact number, 70%, 80%. Years prior to that, it was north of 100%. So I think we're pretty generous with how we allocate cash. Our plans are -- and listen, because I know a lot of people in this room and a lot of people on the webcast, like that dividend and frankly, I do, too. And so our intention is to grow that dividend on an annual basis. right? It's been made very clear to me that's what people want to see. And so that's our intention. And we've done that, I think, pretty generously over the last several years. I think we put the dividend in place for the first time in 2014, and we've grown it pretty much on an annual basis since then.

It's hard to keep up with the stock price. That's a good problem to have.

Listen, it's a good problem to have when your dividend yield goes down because the denominator is driving it down and that largely is what you're seeing here. And obviously, then we supplement that growing dividend with the share buyback and we've taken a meaningful reduction in the share count over, I don't know, the last 5, 6 years because we generate a lot of free cash. So...

Yes. I mean I think it's particularly notable how much cash you generated in a year where your memory customers depleted a lot of cash. So it's a testament to the business.

A lot of it came out of working capital.

Yes. That's good. So great. Well, we'll wrap it up there. Doug, thank you very much for your time.

Okay. Joe, thanks for having us.

Thank you.