I think we're gonna get started. I'm Tim Arcuri. I'm the semiconductor analyst here at UBS. Very pleased to have Applied Materials with us for the next session. Pleased to have Brice Hill, who is the CFO of Applied. And thank you for the time, Brice.
Thanks for hosting us, Tim. This is a fantastic location and great time of year. Appreciate it.
Perfect. So obviously we had some hundreds of pages of export controls that came out.
Mm-hmm.
200+ is my understanding, and you said that it doesn't change anything for you. Your guidance is your guidance.
I think the nuance there, Tim, is we said we're not changing our range. So there will be some impact. So the nuance there is, probably the midpoint will change, but no change to the range.
To the range. Okay.
Right.
Okay. And then, you didn't really provide much commentary on next year. How do you sorta characterize, in light of the export restrictions and just in light of everything, how do you sorta think about next year? What are the puts and takes? Obviously you believe that next year's gonna be an up year, growth year. But what are some of the puts and takes as you think about next year?
Yeah. I think, first, we don't guide the year, so it's, it's not a matter of visibility. We just, by practice, don't guide the year. What we try to do instead is give people as good a sense of what we're seeing in our reported quarter, Q4, and our outlook quarter. And so it's kind of like the speed of the business as best we can tell, and then we share that. And so, you know, as you look at our Q4 and you look at our Q1, we tied it to the fastest moving part of the market is the AI-related elements of the market. So if you think about High Bandwidth Memory, if you think about DRAM, and you think about leading-edge logic, leading-edge logic, we said we see accelerating all through the course of 2024.
So it was at kind of a low point where companies were waiting to ramp the Gate-All-Around technologies, but now it's beginning to pick up. DRAM very strong and High Bandwidth M emory, very strong and advanced packaging very strong. That's where there's a lot of energy in the market, and we see that both in our Q4 and Q1. ICAPS, you know, we had a lot of discussion about ICAPS because ICAPS is our legacy nodes, the nodes that are larger than nanometer, selling to end markets, IoT communications, auto, power, sensors. That's the ICAPS. Those markets, well, the end markets have been weak. The equipment market has been very strong for three years. A lot of that's been the buildup in China of capacity.
We noted that it's a little bit slower than it was in the prior year, but it's still very strong on the ICAPS side. You've kinda got an accelerating leading-edge, which we expected with the oncoming Gate-All-Around node. You've still got a strong ICAPS, a strong DRAM, strong advanced packaging. Then, you know, where it's not as fast as on NAND, but we did say in Q1 we saw it pick up in our NAND orders. That's kind of a sense of, you know, what we view as the market. I think this year will be a tug-of-war between does ICAPS have any slower rate of investment and how fast versus how fast does leading-edge accelerate.
Can you talk about DRAM? You've done a great job in DRAM gaining share. At least according to my numbers, over the past 10 years, you've gained almost 1,000 basis points worth of share. It's been amazing. How have you gained so much share in DRAM, and how do you characterize your position? There's sorta two different pieces, right? There's front-end wafer, and there's back-end advanced packaging. And you're very good at both. So how do you think about DRAM, and how have you gained so much share?
I think the architecture of DRAM is moving closer to the architectures that we deploy in leading logic. And so some of the solutions that we're offering are adopted in the DRAM architecture. And as we look forward in the roadmap, it becomes more and more 3D. And so same thing, you get heavier applications of the types of materials engineering solutions that Applied does. On the packaging side, you know, the innovation that you allude to is the stacked High Bandwidth Memory. And so, I think we've described that in the past that there's about 19 extra steps in the DRAM process to do the stacking. And Applied has a, you know, advantaged position. We sell more than half the equipment for those steps.
And so as HBM, HBM High Bandwidth Memory is growing at a 30% rate, goes back to this AI data center, these high-performance systems. As that continues to grow quickly, you know, customers are building out the capacity of the related packaging equipment. So that's been very strong in the market. Yeah.
I wanted to ask about advanced node foundry. You said there's a $2.5 billion base this year. I think you said fiscal 2024, $2 billion-$2.5 billion. You said that that's gonna double next year. I wanted to dig in on how much of that is incremental. On one hand, advanced nodes are more WFE intense per wafer, but on the other hand, the new advanced node spending is ramping well. Spending on the current FinFET nodes is falling off. So it's not all incremental per se. How do you think about how much is incremental? I get these questions all the time, you know, how levered is Applied really to Gate-A ll-A round?
Yeah. I think the way that we've described that mathematically is our addressable TAM for Gate-A ll-A round specifically grew $1 billion for 100,000 wafer starts of capacity off a $6 billion base. So call it a 15% increase in the addressable market. It actually happens to be the same number for backside power delivery. So if that happens to be in the node that's shipping, then you've got also an increase in addressable market from that perspective. But to answer your question, I would break it into two pieces. One piece is the intensity, which is the 15% increase. There's more steps. There's more materials engineering, and that's the 15% increase. The second question will be how big will that node be? Will it be a really good landing point for designs, or will it not? Like from my perspective, 10 nanometer wasn't a very large node.
seven was a very large node. five maybe not as large. Three probably larger. 2 nanometer right now looks like a large node with the early point of moving from pilot production to high volume manufacturing, production. So we'll have to see what the buildout is. But those are the two components. More intense, 15%. And so, the, I guess, just circling back on your other question, that would be the 15% I would call out as incremental. Most of what we're shipping today is that leading edge. We're not shipping very much into, you know, existing nodes.
Got it. One thing that, and I think this is my 26th year covering Applied. So, one thing that I've really seen lately is that you are pushing pricing a little more than I've ever seen in the past. And it's really interesting because we don't hear about that from a lot of other companies. And you can do it because you're so broad and you're sorta equally exposed to all these different end markets. But at the same time, we have, you know, Intel now is talking about selling some equipment that they never deployed at 7 nanometer, and we, you know, Microchip might be selling tools. I don't know. But there is a little more supply potentially coming onto the used market.
So, does that change the sorta like when you think of how much you can push this pricing theme? Does that change that at all?
I think both are factors, but, you know, the blunt answer would be no. I think you have to have a practice in pricing. And what we're talking about in Applied is we probably in past years, there was enough cost re-engineering to offset inflation such that your tools could roughly have similar pricing, and, you know, you'd demonstrate similar margins over time. With COVID and the supply chain challenges that we all faced, that changed. And all of a sudden, the cost of your equipment and the cost of your provision to your customers is much higher. And there wasn't a process in the company, you know, if you think about like the auto industry. The auto industry has new cars every year.
They get the opportunity to price the new model years, look at the market, all those things that you're kind of thinking about. Those processes weren't as mature, at least in the Applied Materials business. So now we're building that process since we're putting more cost into the tools, and we're working on the leading edge and innovation. We're co-innovating with our customers. We're working, you know, at the leading edge in terms of the things that are required to solve difficult engineering problems. We truly wanna understand what's the value that we're providing, and do we get a fair price for the work that's been done. So I think that process, no matter what happens with any used equipment on the market, that process will continue to mature and continue to focus on.
You know, to the question about secondary market equipment, we're also, you know, sometimes get to look at that equipment. We'll think about refurbishing some of that equipment. I don't think it's been significant to that point, but that is always a factor in the market.
Got it. Can you just talk about gross margin? It there was a slight change, last quarter, well, the quarter prior to when you just reported, was, you know, 47.4% was sorta the baseline of gross margin, and that's ticked up. And now you're saying, well, you know, 48% is the baseline. So in the vein of the same question, you continue to push margins higher. How high do you think you can push gross margin? Can Applied be a 50% gross margin company?
Well, we certainly think we can. In the short term, what we said is we think 48% is representative of where we are today. And I said that specifically because the Q1 guide is 48.4%. So it's a little bit higher, and I'm just signaling that there's a rich mix in Q1, so we're benefiting from specifically the equipment and segment mix in Q1 for you know, display, for example, is lower in our Q1 outlook. And so that benefits the gross margins. But I think the underlying rate has improved over the last two years since the supply chain shocks and since COVID, we've worked you know, significantly on inventory management, excess and obsolete materials, logistics, cost reductions. All those things have helped move us in the right direction.
And then the pricing element that we just described, I think it's been a more modest help in, you know, fiscal 2024, but I think that will, you know, gain speed as we go forward.
I wanted to ask you about WFE share, and I ask you about this a lot, and everyone has their own WFE number, and Gartner's different than, I think, a lot of us are, and you don't really give us a number as much as you used to, and I get why. Your share's been ticking higher, and yet there's a headwind from China because a lot of the incremental spending has been in China, and the domestic Chinese toolmakers are gaining a lot of share within China. So that's a persistent headwind for you. Now, a couple of them just got added to the Entity List. So when you think about your share, does that lift some of the headwinds you've had, even though you have gained share?
Does that make things better for you that, you know, you can gain some more share in China because some of these domestic players have been added to the Entity List?
Possibly. So we'll have to see how that works out. That's, you know, having those companies on the Entity List, they're going to have impacts on their supply chain. We'll have to see how that plays out, whether they can mitigate those impacts similar to, you know, we were able to mitigate some of the impacts that we had on rules. I, I'm sure they'll try. But to your point, it's been a persistent part of the business. So when you think about Applied Materials gaining share over the last five years, during that five years, the local Chinese vendors gained share. And so how could Applied have gained share if the local Chinese vendors are gaining share? Well, our strategy in the company is to continue the R&D roadmap and grow the addressable TAM faster than any share we might lose in the non-differentiated equipment.
When you think about the number of steps on a Gate-All-Around process and how fast that number of steps is increasing and how much more engineering intense those steps are, we think that there's plenty of innovation in the market, even in the ICAPS market, for Applied Materials to keep growing its addressable market, even if some of the, you know, lesser differentiated equipment is duplicated by those vendors.
I wanted to ask about ICAPS. You sorta created a niche and a you know nomenclature for it. But you're not the only one shipping into these markets. When you isolate ICAPS and you look at your position and you look at your share, how would you and it's hard for us from the outside to sorta isolate your share in ICAPS versus in the advanced nodes. But how do you think your share differs within ICAPS versus in the advanced nodes?
Yeah. I don't think it's significantly different. I think we've said in the past it's slightly accretive because the way I view it is if you go back to some of those nodes, 28 nanometer, 40 nanometer, 45 nanometer, 65 nanometer, if you had a perspective on what Applied's share was on those nodes, there's a good chance we have a similar share today. So I think there's continuation as those build out. In fact, I think some of the customers look at that and say, "What is a successful footprint in a node?" And if I'm gonna build out a node, I'm gonna use that successful footprint and use those vendors. And, you know, ICAPS, I think what I would say is when we think about that market, it's the largest semiconductor equipment market, right?
When you think about wafer starts, the ICAPS nodes, 10 nanometers and above, there's more wafer starts per month capacity on those nodes than all of DRAM, NAND, leading logic put together, plus a factor on top of that, and so now you look at the end markets and you think about automotive, industrial, power, image sensors, all these end markets. Every third-party forecast has those growing at mid- to high-single digits. There's more capacity put in place every single year in those markets, so you're right. You know, Gary and Applied started a ICAPS-focused business five years ago. They've introduced new tools in the space, and we continue to look at that roadmap for innovations, a lot of them on the power, efficiency component. We continue to look at innovations to drive that business, so I think it's a critical business for Applied.
I expect it to grow significantly over time. You know, notwithstanding if we have a little bit slower rate of investment because of the China buildup, we'll see how that plays out.
Got it. Can we talk about service for a minute? You've done a really, really great job, and you've done a great job signing more tools to long-term contracts, and particularly in China. I think tools under contract grew 10% year- over- year, and it seems like China was a big piece of that. And it seems like China's a bigger piece of your service revenue than what I hear from some of the other companies. So can you talk about what's going on in service, how you've been able to monetize that, and are the Chinese customers and also talk about ICAPS within that, are those customers willing and they want to sign up to long-term deals?
Yeah, I think so. I mean, from my perspective, when you have a customer who's building a new facility, even if it's a large existing customer building a new facility in a different place, a different location than they're used to, they're very likely to look at the Applied Services business because you get, you know, a successful labor force. You get people who understand how to ramp the tools. And a lot of these companies are looking for qualified labor. And so if Applied Materials can bring in people that can help ramp the equipment after we install it, ramp it to high yield, and help them maintain it at a high yield, that's a big advantage. And actually, you know, when you think about selling in China, that's one of the critical capabilities that Applied Materials brings into the space.
Not only are we offering the hardware that can do a job, but we're offering a successful supply chain behind it that has a lot of capabilities. We're offering a services business that can install and then maintain the tool over time and keep it at high yields, plus information on how to optimize the performance of that equipment over time, which changes over time because we continue to learn about the fleet and how to best operate the fleet. So I think all those factors help. And the fact that some of these customers are newer and smaller. They definitely look at the services capability as an advantage for them to get started.
The gross margin in services is a little lower, but the op margin is there. Is that correct?
That's right. High level. Yes. Yeah.
And is there any capability when you sign these deals and you go in and you upgrade the tools and you maintain them and you extend the performance of them and the customer sees and says, "Wow, I'm actually getting a you know a return on this," is that reflected in the, like, is the outcome of that they come to you and say, "Hey, I'm willing to re-up to more long-term contracts," or are you able to say, "Hey, listen, I wanna take some of that upside. So I just gave you money. I want some of that money back"?
Yeah, I think, you're right. The length of our contracts, I think the average contract in the course of the two years that I've been here has grown from two and a half years to I think it's 2.9 years at this point. So what we're seeing is customers are making longer-term solutions decisions, with those. I'm not sure if that's, you know, evenly, deployed across China, but, certainly some of the larger customers are making those longer-term, commitments. And that's better for both parties, Tim, because, you know, for Applied Materials also, you have to plan your labor force, and you have to plan your warehouses. You have to plan your supply chain. And so when we are able to sign a longer-term arrangement with a customer, those become, easier to plan and easier to deliver a successful level of service for the customer.
Brice, I think based on what you said, it seems to me like ICAPS is maybe 50%-60% of your foundry logic revenue. Is that fair?
Yeah. Before the, you know, a couple of years ago, I think we viewed it as more 50/50, but because ICAPS has grown so strongly, it's more than half of the foundry logic. Now, this year, we expect to see foundry leading logic accelerating. So that balance might, you know, turn back the other way a little bit, but we'll have to see how the year plays out.
Can you talk a little bit? You're not as levered to NAND as maybe some of the others are, but can you talk about what you're seeing in NAND? It sounds like you're a little more excited about DRAM than you are in NAND.
Yeah. I think just from a architecture perspective, I know Applied, you know, when we're thinking about new technologies in DRAM, I know they're very exciting to Applied because of the materials engineering requirements, the 3D nature of them. There's a process called 4F- squared that, you know, there's a lot of R&D going in that space. Further out in time, there's 3D DRAM. So I think it's fair to say that our integrated solutions and some of the solutions that we provide have potentially more applications there. That said, you know, NAND is a focus for us also, specifically some applications in that business. It's hard to, you know, the difference between the two, I think at a high level, DRAM wafer starts are still growing. NAND wafer starts haven't grown, so it's mostly an upgrade market.
Mm-hmm.
And that also changes the complexion if you're trying to, you know, win applications. It's harder probably in an upgrade market than it is if you have new capacity being put in place. So there's a little bit different dynamic there, but we do think the NAND market will improve. And, you know, we've seen that a little bit in our Q1 outlook.
Got it. Not a market that you have a lot of revenue in today, but can we talk about flat panel for a minute? Because it's been dead for a while, but a lot of companies are working on large glass for, you know, fan-out panel level packaging, and that's gonna replace, you know, wafer level stuff like CoWoS probably in the not-too-distant future. You know, TSMC's talking about, you know, going to fan-out, as soon as 2027. So it's not that far out there. So can you talk about what you're doing with respect to large substrates?
So we're definitely evaluating, you know, at least all the different technologies and substrate types that are being contemplated. Applied is doing R&D, and, you know, we expect to work with our customers on those solutions. So I think you're exactly right. What's happening? It goes back to, you know, this AI, high-performance systems. And what all the major players in the world are doing is trying to figure out how do you optimize the performance of these systems. It's the mix of accelerators, GPUs, CPUs, High Bandwidth Memory, other components, and then also how do you lay them out on a board and on a system on a chip such that the speeds and interconnects are optimized. And the substrate, changing the substrate and using different techniques of interconnects, those are things that everybody's experimenting with, and we're certainly involved in that.
Do you see, like, when this takes off, is it gonna be a, I mean, once TSMC starts to do it, you're, I mean, it seems like the kind of thing where your business should start to really go vertical?
Yeah, I think we have, you know, when we talk about packaging and specifically advanced packaging, the business is today it's a $1.7 billion business. We're including the packaging, you know, equipment that we sold into HBM, the High Bandwidth Memory in that $1.7 billion. And we've said that, yeah, we expect that to double over the next coming years, just like it has in the, in the recent past. And it'll be that energy that you're describing and sorta the solution, the search for more efficient solutions in a performance realm.
There's a question from the field about e Beam and sorta some of your activity there. And maybe you can talk about just process control generally, how you view your share and your position.
So we have number one position in eBeam. We actually just had a sales conference in Athens for that business group. Very strong roadmap. The company is investing significantly all across the different areas of inspection and process controls. And so I think, you know, it's a critical business for Applied. And also, as you might imagine, as we think about these different innovations in packaging and other areas, we're getting to smaller and smaller dimensions. It offers new opportunities for inspection and metrology. And so I think there's a lot of opportunity in that business, and it's a, you know, it's definitely an internal focus area from a portfolio perspective.
Got it. This question's more about sorta what the normalized earnings power is for the company. And this is not just an, you know, question for Applied. This is more a question for the whole industry that I'm asking you. If China consumes between 20%-25% of all semiconductor revenue dollars.
Mm-hmm.
Would you say it's fair to believe that over the longer term, China probably will be 20%-25% of wafer fab equipment as well? So, so you as a CFO, you think any number above that is an over-earned number for a period of time. So over the longer term, and, you know, you're not down to 2025 yet. You're gonna you're down to 30. You may go below 30 next year. You're getting there. But is that a fair way to think about it, or would you say, "No, I push back on that because of reason X or reason Y"?
Are you asking, the percentage of equipment that they'll consume?
Yeah. Yeah.
Yeah.
So I'm saying that they're, you know, they've been at 42% of WFE. They're now down to, you know, next year.
Oh, I see.
But really, they're only 20%-25% of consumption dollars. So shouldn't their portion of WFE over the longer term also be 20%-25%?
Yeah. I don't see a reason why that logic doesn't hold true. We certainly think, first of all, I think my number would be closer to 30% on semiconductor consumption, but whatever. It's in that range. I see the logic of the, you know, I think we view it similarly. The country, we think, has a strategic goal to put enough capacity in place to be self-sufficient. So matching those two numbers makes sense, and then holding it for the long term makes sense. And then we've certainly seen a big ramp over the past three years in capacity in China. So when we see soft markets in ICAPS and we've seen a rapid ramp of capacity in China, that's why we said, "You know what? Our Q4 was a little bit slower than the prior Q4.
We think that there should be continued new capacity investments, but maybe at a slower rate. That normalizes. We haven't seen it yet, but that's kind of what, you know, we're watchful for.
Got it. This is a topic that, that we talk about a lot, but I wanna talk about WFE intensity. I would be remiss if I didn't ask you about this. And it used to be 11, 11.5, well, even below that, and it crept up, crept up and up. And then China went from a very low number to, you know, now it's $35 billion this year. And that corresponded with a very large increase in WFE intensity, culminating in 16.5% last year. It's been coming down a little bit. But the question is, how do you think about the, like, what the actual long-term WFE intensity is? Because certainly it's there's a, there's an upward trending underlying curve, but it's, but China's been inflating the, you know, number during the past couple of years. So is 13% the right number?
13 and a half, the right number?
First of all, it's. I think we agree with all the shape of that curve. So intensity being, you know, equipment sales over semiconductor sales. The dynamic that you described, I think it was rising. If you go back 10 years, I think it was rising because there was less used equipment available in the 200 mm space and definitely in the 300 mm space. So the intensity of the ICAPS nodes started rising, and that helped raise the intensity level overall. And then to your point, ICAPS in China was accelerating dramatically, plus memory, and that raised the WFE sales faster than semiconductor sales. So now we expect to be on the other side of that. Well, that would fall again. Where the landing point is, is hard to know because we've said mid-teens, Tim. So in the range, could it be 13? I suppose it could be 13.
Could it be 16? It could be 16. In the mid-teens would be our perspective. The dynamic on the leading edge is you've got a lot of pricing power in some of the leading graphics components. So that's raising the price on semiconductors without raising the WFE. On the flip side, you have more intensity at each new node on the WFE. And so those are kind of, you know, battling. So anyway, we do expect it to decline because that ramp won't repeat itself, but somewhere in that range for the a medium-term normalized level.
Got it. Can you talk about capital return? You've always been a very good, you know, repurchaser of your stock. You know, every quarter you're in the market to a fairly high degree buying back your stock. Is it more of a systematic program, or is there an opportunistic aspect to it? And I ask you that because your stock has come down a lot during the past, you know, three months. Does that, as the CFO, does that kick in an opportunistic thought in your mind, "Hey, you know, maybe I do wanna be more aggressive down here"?
Yeah. I think, first of all, you know, if we back up to, you know, 1,000 ft, we've said that we'll return 80%-100% of the capital to shareholders. And when we're not doing much M&A, it, you know, over time, it'll be closer to 100%, obviously. And in the short term, we try to be a little bit price sensitive. We use a grid to guide our buybacks each quarter, but I think you said it right. We're always in the market. And if you see the price be more favorable, we're probably buying a little bit more and vice versa. And so that's a dynamic each quarter.
Got it. And can you just go back and talk about service? And have you ever talked about how much China is as a portion of your service business? In other words, I'm sure that it's a lower portion of your service business than it is of your equipment business because it's been, you know, in the 40s of your overall revenue. But is there some metric that you've given us that maybe might help us figure out how much China is as a portion of service?
We have. We said that the attach we think is approximately equivalent to the attach in the rest of the world, so you know, somewhere in the 30% range would be our view of the attach for services.
Just in terms of the types of customers, is your service discussion with, like, the big customers, the largest spenders, is that discussion changing, or is the growth in service predominantly it's the smaller customers that are waking up to the realization that they have to sign up to long-term agreements?
No, it's a great question. And it kind of, what I was alluding to on the location of some of those larger customers, when there's incentives globally to move the production to, you know, either new cities or new places that a company, a large company hasn't operated before, they're more likely to tap us for services help, and the labor force that goes with that. So I, I do think that, even the leading edge nodes and the places where some of those factories are being invested, there's increased demand on the services side. And that's part of our planning process.
Great. And maybe just last question, Brice. When you sorta look at your stock and you look at how, you know, the questions you get from investors, what's one thing that, you know, you leave these meetings and you say, "You know, people just are not getting this piece of our story." Is there something people are missing?
I think, you know, my perspective is longer term. And so a lot of the questions we get are very short term, you know, what's gonna happen next quarter, etc. And we tend to back out and say, "Oh my gosh, look at AI and look at the need for device performance and improved efficiency of devices going forward." And that's what guides our investment decisions. So if you're inside Applied and you're thinking about, "How do I choose how much R&D to invest?" We're thinking about a five and 10-year roadmap, ever-increasing demand for semiconductors, more capacity being put in place every single year, and then developing that leadership development position with our customers that's collaborative. I think the collaborative nature of that probably isn't obvious to people outside the industry.
It's not like speculative R&D where you just do the R&D and hope that there'll be a solution. It's like we're working on something that both we and the customer want to solve.
Got it.
Yeah.
That's the time we had. So thank you.
Great. Thank you very much, Tim.
Thanks a lot.
Appreciate it. Thank you.