Very good morning. Welcome to this session. I'm Vivek Arya from BofA's Semiconductor team, and absolutely delighted to have the team from Applied Materials join us this morning. Brice Hill, Senior Vice President and CFO of Applied Materials. What I thought we would do is go through some of my questions, but please feel free to raise your hand if you would-
Yeah
Like to bring up anything. Thank you so much, Brice. Really appreciate you joining us.
Thank you for hosting, Vivek. It's great to be here.
Excellent. So Brice, I know we'll get into the nitty-gritty of WFE and all the near-term trends. I just wanted to, you know, get your perspective. Applied is the largest semi cap equipment company, and I think a key enabler of the semi-industry's march towards that, you know, $1 trillion right target that has been set for hopefully the end of the decade or soon thereafter. What are your areas of kind of strategic and operational focus as the industry is marching towards that?
Thanks, Vivek. First of all, very happy to be here. Nice to see everyone, a lot of familiar faces. You know, it's an exciting time for the semiconductor industry and for the equipment industry. Many of the governments in the world have incentive programs to increase their local ability to produce semiconductors. That's driving, you know, a sort of new demand function in terms of locations for semiconductors. Then, of course, we have this tailwind that is the generative AI and AI in general right now, that's driving a lot of excitement on leading-edge process technology, on advanced packaging, on DRAM memory, the high bandwidth memory, parts of the business that are all coming together to drive performance at the leading edge, to drive new capabilities.
That's creating a lot of excitement in the equipment industry, especially at the leading edge.
Mm-hmm.
I think for Applied, Vivek, you know, we maintain a fairly large portfolio of tools, and the way that we would like investors to think of that is, we're curating that equipment so that our customers can solve, you know, the most difficult problems on the leading edge. Our CEO likes to talk about. We try to estimate the technology inflections that are 3, 5, 7 years away, and then we work collaboratively with our customers to solve the new technology issues. If you think of Gate-All-Around transistor as an example of that, you're really starting to operate at the atomic level, at the nano level, and so that requires new materials, it requires new capabilities and more steps to build those devices.
I think the people here know that when you build these transistors, you build these devices, you sort of build them from the bottom up, layer by layer by layer. That's, you know, different types of depositing, treatments, cleaning, all sorts of different equipment. And one of the things that Applied can do is put its, different tool types together so that they're connected, and we can create solutions that no one else in the industry, can create, and that's what we focus on working on. So we talk about inflections that are coming up in the process technology side. Gate-All-Around is an example, Backside Power is an example, new architectures in memory will be an example, new bonding techniques, in Advanced Packaging.
Applied Materials works with our customers to develop those technologies that are several, you know, years away, and then develop the materials for, or the equipment for high-volume manufacturing, then put them in place and service them in the future. So really bright prospects right now. I don't think we've changed our view. Our view is semiconductors grow at about three times the rate of the economy. So if GDP is 2.5%, semiconductors have been growing, like, 7.5%, approximately, and we expect the equipment industry to grow as fast or faster than that, and we expect Applied to outperform that market. And then we expect our services business, which services that fleet of tools that's installed, we expect our services business to even outgrow our own equipment business.
So those are kind of the dynamics, and, you know, it, it's a good time to be in the industry.
Thank you, Brice. So let me pick up on the two points that you mentioned. So first is, you know, reshoring demand. How much of that has Applied seen already, and how much is still to come? Because, it's been in the news since the last two years, but I don't know how many actual dollars have been allocated, right? No, no, outside of the US, perhaps, but maybe not as much. So how much has Applied seen so far, and what does it mean for the next two, three years?
I would say very little so far in terms of direct orders. We've seen some because there are companies that are starting to claim the tax credits related to some of those incentives. But you can imagine, when a company decides to put a factory in a different place than they normally build, it takes a year or longer to build the factory itself. They'll order the equipment towards the end of that cycle to begin installing it once the factory is clean, and then they'll prove their process technology and begin to ramp. So it's probably on the cusp of starting to see that business, but I, I'll just wanna be clear that we don't view that as incremental business for Applied.
What we view that as, the industry is growing, it's growing at that, you know, mid- to high-single-digit rate that I just described, let's call it 7.5 or something like that. More and more equipment will be needed. These incentives will determine where that equipment lands. We don't think companies will order equipment that they don't need for real production, okay? We do think that some of that production will be a little bit less efficient than if the owning company were to put it in their normal, you know, economy of scale location, and that'll drive probably 3%-5% incremental equipment, just because of not as good economies of scale. But other than that, we just think it'll be location, and it'll feed the growth of the industry going forward.
Got it. So the sense is that whatever wafer fab equipment intensity, right, so WFE to sales ratios have been, that they are not going to be impacted to a great extent by just this one factor?
Correct. We don't think we don't think there's a one-time stimulus associated with this. We think that this will just be companies were going to make those investments. Now, they may choose to make it in a different country and benefit from some of the incentives, but and it'll have a small benefit to us, but not something that is a spike or that will go away.
Got it.
Yeah.
So part of that geopolitical landscape, China was a very big driver of demand for the industry last year. How are you looking at demand from your China customers this year, right, and, and over the next one or two years?
China's been extremely strong, over a many-year period. China has grown a significant amount of wafer start capacity, mostly for us in the ICAPS space, although we've been shipping, DRAM, too. We can, we can come back to that. We think it's a, you know, it's a country strategic priority to build. Our view is they're building, enough capacity to be self-sufficient, especially in those trailing nodes that we're able, to supply, so they're on a journey, to do that, and that's really part of the story why our ICAPS business, the more mature nodes, you know, what we call IoT, communications, auto, power, sensors, building all those edge-type devices. China's put a lot of capacity in, but again, we don't think that that capacity... It's been fast.
They probably got a little bit ahead of the demand function, but if you view those end markets as growing at 7.5% or whatever per year, then there's probably a little bit to absorb, but we think that will be a stable and growing market going forward.
I see. And, if we try to unpack that China demand between the DRAM project, right, which you mentioned could start to taper off, excluding that, can China grow or be flat or be down this year?
Yeah, if the last three quarters, our company had significant shipments related to China DRAM technology that the whole industry was allowed to ship to was you know met the requirements from a technology and trade rule perspective. That's going to be falling off in our second half, and what's replacing that because our revenue guidance was relatively flat, what's replacing that is strength in the ICAPS and then strength in leading edge. And if we pull out China and look at China, other than that short bubble that we saw in the DRAM, China's been very strong, and we continue to believe that that market will continue to invest. There's 30 factories that we're tracking right now, more than 30 factories that are in the process of ramping their equipment, installing equipment, and growing.
It's a much larger customer base than even that, and so we look at this as a long-term trend. We don't make a call on whether, you know, this year will end up being higher than last year or next year will be higher than this year, but it's, we don't see a cliff here. We don't see a change in the investment pattern.
I see. Brice, how do you, kind of, sanity check the ICAPS demand? Because when we look at a lot of, the U.S. automotive industrial companies, right, they have seen 30%-40% type of peak-to-drop drawdowns, right, in their demand. When we look at, some of, you know, let's say, somebody like a GlobalFoundries, right? The CapEx went from $2 billion to, like, $700 million. So a lot of those—So when you see, U.S., where there's a lot more credible data available, taking down their CapEx so much, aren't you concerned that if China is building so much, then the Chinese economy is not doing so well, that that could be pulling forward a lot of this, pure fear of restrictions and, and whatnot?
I think it is interesting, Vivek, because to your point, some of the end markets we all know have been weaker. Auto industrial have been weaker. For a while, PCs and smartphones were a little weaker. That's starting to pick up right now. The timing ends up being a little bit different, typically. I mean, we can use DRAM as an example. DRAM's been continually strong, even though the memory market has been down. So when you're building factories and you're installing equipment, people here in the room might know that it's a four-year horizon that you're planning for. I have to put floor space in place. I have to add equipment. I have to qualify that equipment. It's for the next technology node. I have to be ready to ramp that for my customers.
My customers are qualifying those parts in their products that they're building. So there's a flywheel, and that's, you know, a little bit separate from the demand of the quarter for the end devices. So going back to the ICAPS, even though those end markets were weaker, we saw continued investment. I mean, even last year, you know, ICAPS was very strong growth across the board and across the world, companies making investments. So that's one thing that I would point out, is that the performance of the consumer market, if you will, is a little detached from the planning cycle of companies. And then going back to China, all of ICAPS has been lower in utilization than you would like. Last quarter, I would have said it was in the mid-seventies from a utilization perspective.
Generally, factories like to be, you know, 85% or higher from a loading perspective. So there is some digestion that has to happen or increased demand that has to happen to get right back on the beat rate of adding, you know, 7%-10% of capacity every single year. So we do think, we've said there could be some digestion, but again, in our Q3 outlook that we just gave, our fiscal Q3, ICAPS is growing and Leading Logic is growing, so we don't see that currently.
Understood. On AI, which is right, all the buzz today, why haven't we seen WFE benefit more from it so far, and when does it start to benefit from it?
I think, you know, if we break things down for our business, where you do see benefits is strength in DRAM or high bandwidth memory. During the quarter, during this last quarter, the allocation of DRAM wafers to high bandwidth memory went from 5%-15%. That's a significant change in one quarter. In our business, as you know, we highlighted that our related HBM equipment, high bandwidth memory equipment, was growing rapidly. We think it will grow 6x in the course of the year for the equipment that we sell for those DRAM steps. If you look into the leading nodes, utilization has been low, just like we said, for the rest of the industry, PCs, going back to PCs and smartphones primarily.
But if you look for factories that are very well loaded or even 100% loaded, they're the factories that are on the leading edge. So our view, at least my view, would be that you're definitely starting to see the benefits of that demand function, high-bandwidth memory, DRAM loadings, DRAM pricing, Leading Logic fab utilization, and we're on the cusp of ramping a new process technology with the Gate-All-Around transistor, so a new transistor era. So we think those are all positives for Leading Logic. And if you break down the devices, you know, companies right now are racing to build the highest performance computing systems. Those are taking networking chips, graphics chips, CPU chips, accelerators, pairing them with high-bandwidth memory, all on a single platform with advanced packaging capabilities.
Companies are racing to figure out, how do we run these workloads more efficiently at better power, et cetera? And that's gonna push demand on all those functions.
Got it. Does, quickly on that AI topic, does edge AI and the, just the adoption of AI in smartphones and PCs, does that mean anything for WFE, or is it too early to make that call? Does it mean, I don't know, larger die sizes? Because all these NPU and other functions that have to be added, I imagine they will consume a lot more transistor strips.
Certainly, certainly on the server side, we know that servers are much more intense. An AI server-focused server is much more intense than a different workload server, intense from a graphics, a networking, a memory perspective, et cetera. On the consumer devices, I'm not sure yet. I haven't looked at the detail. I would expect they'll have more memory-
All right.
than they would before, but we'll have to see.
Okay.
Yeah.
Gate-All-Around, can you maybe describe to us where all in, in your product portfolio Applied Materials is involved? And I think you suggested about $2.5 billion for this year, potentially doubling, right, next year. So how much of that is market growth, and how much of that is kind of, you know, Applied specific, growth?
So, yes. Gary highlighted, our CEO highlighted during the earnings call that, this year being the first year of shipments for, you know, high volume manufacturing Gate-All-Around nodes. We're seeing about $2.5 billion of business that's associated with, the transistor, building the transistor, and also the wiring, for that transistor. As you probably know, Applied has, significant share in the wiring for the devices, also, all the different interconnect layers. We do expect that to double. There's many different, you know, equipment functions that are involved with that, and it just goes along... If you ask, "Is this market growth, or is it technology?" If you ask that question, it's probably hard to sort that out. In the first stages, I guess I would put this mostly on technology.
Again, if you, if you are our customer, and you're planning a new process node, and you have multiple device customers that are designing on your node, this is something that you all work on for several years. And so the most important part is when the device customers are ready to launch their parts, our customers have the capacity in place at high volume and high yield, so they can deliver those. And so I think that's mostly technology, and then the question will be, how much is the capacity?
Right.
I think this demand function that we're talking about, certainly AI is a great tailwind for the demand function, across the board on leading edge.
Got it. So advanced packaging, number of companies, right, selling GPUs and ASICs are talking about CoWoS being one of the leading supply constraints for them. What is the outlook for Applied's advanced packaging business this year? How do you look at it, you know, growing next year? Like, if we just look at the leading players' CoWoS capacity, is that what your growth is then ultimately correlated to? Because they seem to be more than doubling, right, over time.
We have said our advanced packaging business is approximately $1.7 billion this year. That includes, by the way, the $600 million I referenced for high-bandwidth memory packaging. We do think it will double over the next few years. We're not being more precise than that, so, you know, that implies a growth rate of 20%-30%, and we think that's consistent with what we're seeing on those edge AI workloads, that type of growth rate. Could be faster. We'll see.
Okay. On services, where are we right now? Like, what part of your services business is more cyclical, right? What part is subscription based?
Right. The, you know, the spares and the equipment that we offer to keep the machines running, if you will, that goes up and down with utilization. The utilization across the entire network has been lower over the past year. You know, it's probably... Memory was lower, probably in the low 70s. ICAPS was also in that region. Leading logic was probably better than that. And what we've seen in the last quarter, Q2, and in our Q3 outlook, is utilization is improving across the entire industry. It's improving, especially in memory. You know, DRAM is a good example of that. And so as the companies put all their equipment back to productive use, they start requiring more spares and services, and so that's the part of the business that's, you know, cyclical, if you will.
But 80, 80% approximately, a little bit more than 80% of the business is what we call recurring, and then about two-thirds of that is under subscription for us, and those are subscriptions that are two and a half years, typically, in duration, and we have over a 90% renewal rate on that. And with the incentives that we started with this morning, different incentives for companies to put factories in new places, they're more likely to ask us for help with services because we have the qualified workforce in those new places, and so that's a tailwind for our services business. Very exciting business, different economically from a, you know, modeling perspective than our equipment business. The equipment business, people call it growth cyclical, right?
The services business is based on what's your installed base, and every day that we ship a tool, our installed base grows, gives us the opportunity to sell spares and sell services to that, equipment base. So whether it's a great year in equipment or a not so great year in equipment, that installed base is growing, and then we're developing new services for our customers. AI is an area that's very helpful. We can see our entire installed base of tools. We can help the customers pinpoint, how do you keep those tools operating at the highest utilization, the highest yields? These are the knobs that you can turn, to make that operation, work as well as possible.
Those are services that we can offer to our customers, so you grow the value that you can provide for each tool, you grow the number of tools, and then you improve the utilizations. We've said that the services business will grow at low double-digit rate for the next few years. And the other thing, Vivek, that I like to point out to the investment community is, I draw a tie between the profits we make in our services business, which we view as very stable and growing because of the subscription-type model, tie that to the dividend. So when you look at the operating profit for our services business, it more than covers our dividend, and we've embarked on a path. We raised our dividend 23% in 2023.
The board just approved a 25% increase in March of this year. We said we'll over the course of the four years starting last year, our intent is to double the size of the dividend, so rates growth rates of similar size, and then we'll see where we are, and what's affording that is this stable services business. And then just to complete that equation, from a distribution of profits perspective, we'll distribute. You know, we can't do much M&A at this point. Globally, that'll be pretty small, so we'll distribute, at the end of the day, all of our excess profits or nearly all of our excess profits. What doesn't get distributed from a dividend perspective will go out in the buyback function.
So I think that's a good way for investors to think, you know, how confident are we in the dividend? How confident are we in the growth of the dividend? You can look at our services business to provide a lot of that confidence, and then the growth cyclical will be the buybacks.
All right. You answered four of my questions already.
Well, you know-
I anticipated that. Okay. Any recovery prospects in NAND over the next one or two quarters? Are you seeing any, let's call them, green shoots at all?
Yeah, we're definitely seeing green shoots in NAND. First of all, utilizations have improved, pricing has improved, inventory's improved. At least I would be in the camp of it's been hard to explain on the NAND demand side, why it's been as low as it has. It's been hard to explain that because when you think about the AI workloads, and you think about the data required to do the things that we're talking about with GenAI, it requires a lot of storage... So we expect that the NAND market will grow. We've said over the long term, we expect memory to be about a third of the WFE market, roughly equally balanced between DRAM and NAND. There's probably just not as much of a tailwind for NAND right now that DRAM has with the HBM.
Green shoots, utilization, pricing, inventory is all better, and we see several of our customers, they're publicly saying that they think the storage requirements will be increasing with these workloads.
Right. That's interesting you say a third. Historically, it's been kind of a 55-45 mix, and now you sense it's more of a 65-35. So memory, right, I guess what it means is that you expect, you know, those memory buyers to be disciplined for a longer time.
Well, what's happening in the memory space, both DRAM and NAND, is in a good way, especially NAND, Moore's Law, at least from my perspective, is very much alive. So the bit density on a wafer or a bit density on a die-
Mm-hmm.
is increasing at such a great rate that they're able to service the bit demand growth, which is very healthy with the same number of wafers. So on the memory side, it's almost all technology. We're not adding a lot of wafer starts in the industry to service what is a growing bit demand for memory. Does that make sense?
Yeah, I understand. On gross margins, so you guided to 47%. The long-term target is 48.5%. Can you help us bridge, you know, from where you are today?
Yeah, so we were at 49 or 47.9 in Q1, we were 47.5. Those were elevated because of the DRAM shipments we were having to China. Our China mix, you know, is falling through the year. We've signaled that that DRAM, it won't go to zero, but it'll be, you know, closer to zero. So as, China mix normalizes in the second half, closer to something that's normal for us, like 30%, our Q3 guide is consistent with that. The gross margin, about 47%, 47.0, I would say that's where we are. I would call Q3 kind of a normal mix quarter from that China perspective. And so how are we going to get to 48? How are we going to get to 48.5? It's cost control on one side, but it's really pricing.
We're introducing new equipment all the time. We're working with our customers on, you know, what we consider to be the highest value problems, but we're also working on making pricing adjustments with our customers because the cost and value of the equipment that we're making is higher, higher value than what it was three years ago. Part of that was the supply chain changes that we all faced, although some of that has receded, and part of it is just, you know, the cost profile, as everybody knows from inflation, the cost of labor, the cost of new equipment, the cost of utilities, et cetera, has gone up, and so we're working on the pricing equation to share that with our customers.
Any big CapEx projects that you anticipate, that investors should be aware of?
We do. We announced a rather large CapEx project in California. It's called EPIC. It's a platform where we will build a large lab that we can co-invent and collaborate with our customers. That was recently approved by Sunnyvale. And so, the only disappointment we had with that, Vivek, was we were also applying for CHIPS grant money related to that project. We recently were disappointed to hear that there's not grant monies available for that, although there may be other opportunities. So we're still going to make that investment.
The size might not be what we originally anticipated, but we're still going to make that investment, and I think the company views it as a new development platform, where we'll have multiple facilities around the world that are this EPIC platform, where we'll be able to co-innovate with our, with our customers.
Okay. Finally, SEMICON West is around the corner. I was hoping you could give us some kind of a sneak preview.
Absolutely not.
Okay, I tried. Let me ask this other thing instead. The Sculpta pattern, shaping tool, what success have you seen so far?
That's, that's an excellent example of a new innovation for Applied. It's a patterning shaping tool, as you mentioned. One customer has announced that they have multiple uses for this tool.
Okay.
You know, we've said that we think it'll be a $500 million run rate this year. We expect to reach revenues of $200 million on this tool. It's just beginning its journey, and we think that, you know, the ability to shape the designs that are already put in place by litho equipment, as an example, is something that customers will find more and more applications for, and we're excited about that. And in general, you know, it's a good example of this innovation at the leading edge.
Okay. Terrific. Thank you so much, Brice.
Thank you.
Really appreciate your time.
Thank you.
Thanks, everyone.
Thank you.