Thank you for coming. Good afternoon. I'm Stacy Rasgon. I cover the U.S. Semiconductor and SemiCap equipment space here at Bernstein, and it's my great honor to introduce our guest, the president and CEO of Applied Materials, Mr. Gary Dickerson. Before we start, I wanna mention, you have, if you wanna ask questions, there should be a QR code in your program, that you can link to our Pigeonhole form, where you can put questions in and submit them, and we will have time for our Q&A at the end. So look, I love SemiCap, just trust me, it's in my blood. Like, my former life, I built plasma etchers, and so I've been in and around this for a long time.
The industry itself, I think, has been enjoying a real renaissance over the last several years, and it's really come into its own. I think it's both the industry growth and it seems the industry capital intensity have been continuing to inflect higher. I think the contributions from companies like Applied Materials are more important than ever before, as materials-driven innovation moves to the forefront of process technology development, as physical limits get closer and closer. And to that end, SemiCap has really been becoming top of mind for many of my clients. You know, there are the nearer-term questions of course, cycle and China, and I'm sure we'll get into some of those.
But I'm increasingly getting people looking at the longer-term potential of this industry as they start to view it on more secular rather than purely cyclical terms. And so to tell us all about that, it gives me great pleasure to welcome Gary. So thank you so much for coming today. I really appreciate it.
Oh, thank you, Stacy. I'm really happy to be here.
Yeah.
So let's talk about, like, maybe how we got here. And it's funny, you know, like, 2023, when you were here a year ago, we were looking for 2023 to be sort of like the cycle trough, and we were all looking. I think at the beginning of this year, for 2024, at least, we were looking for $70 billion in industry WFE or something like that. You know, we did mid-90s. Right?
I would say we were never at 70.
You may not have been, but I mean, collectively, many, many investors were, and I mean, we did mid-nineties. And not too bad for a trough WFE number versus, like, an industry that, you know, was a different place when we were doing $20 billion in trough WFE.
Yeah, yeah.
But beyond that, a lot of things that contributed to that sort of, like, differential, I mean, it was a lot of areas where you guys have really refocused.
Yeah.
It was packaging-
Yeah
... And DRAM in China and ICAPS.
Yeah.
Like, ICAPS is their lagging edge business, or it's industrial..
IoT, communications, automotive, power, sensors.
There we go. There we go. But I was wondering just if you could talk about, rather than the number itself-
Yeah.
... Like, talk about, like, the efforts that AMAT has been putting into those areas specifically.
Yeah.
Because I think those really are the ones. And, and it's not like you just, like, stumbled upon them. Like, these are-
Yeah.
... You guys have been investing in for a long time.
Yeah. So I would say, you know, the real focus for us is enabling the device architecture inflections.
Mm.
I mean, the thing I love about this industry is that there's this constant race-
Yeah.
... Whether it's in, you know, high-performance logic or memory or ICAPS around power electronics and sensors or packaging technologies, it's this constant race. And I was... I mean, AI is the biggest inflection of our lifetimes. I know inside of Applied, we're driving major programs that are, you know, really big-
Mm-hmm.
... Top-line, bottom-line contributors to the company. Lisa Su was presenting at a conference a couple of weeks ago. She said that she is driving 100 times improvement-
Mm-hmm.
... In energy-efficient computing.
Mm.
So I think that when you think about AI and the growth, and AI server has 8 times more foundry logic content, 8 times more DRAM content, and this whole focus around power consumption and energy-efficient computing, that's really the focus for us, are those big inflections. So you have, like, gate-all-around for high-performance logic, that's part of AI, 30% improvement in power consumption. Then you have backside power, where you move the power lines to the back of the wafer, 25% improvement in power consumption. You have packaging technologies, that's a $1.7 billion business for us today.
Wow!
And we say that business can more than double going forward. So you know, that is a lot of that's focused on-
Mm.
... Performance and power consumption. DRAM, there are architecture changes that are coming there, that are in the next few years, that our customers are talking about big improvements in power consumption. So I think those are these races for these architecture inflections, and ICAPS, you have the same thing. You have inflections, like I said, in power electronics, silicon carbide, and GaN, and sensors, and all of those things. So within Applied, we have the broadest connected portfolio in the industry. And when you think about those big inflections and Gate-All-Around, we can capture over 50% of the incremental spending-
Mm.
... Wor Gate-All-Around. In backside power, 50% of that spending. And, you know, we're really, really well-positioned for all of those major inflections. And we had a big change in strategy in Applied about six years ago.
Mm.
So we formed an Integrated Materials Solutions group that focused on how we connect all of those magical technologies together, to enable the big inflections. We've grown. We have integrated platforms. So in your smartphone, I tell people there's 15 billion transistors and 100 kilometers of wiring-
Mm.
... Which is mind-boggling. You have to move data through that 100-kilometer wire at very high speed, very low resistance. It's magic. I mean, the way you drive that is just incredible. So we have these connected platforms with seven technologies-
Mm.
... Under vacuum, that has grown from 20% of our revenue to 30% of our revenue.
Mm.
And then we've also built these integration, innovation teams that are world-class. So we're really co-innovating with our customers-
Mm-hmm.
... Across every one of those areas, in high performance, logic, memory, ICAPS, packaging.
Mm-hmm.
We're working four technology nodes out into the future.
Mm-hmm.
So again, that's that race-
4, 4 technology nodes?
Four technology nodes out in the future.
Where, where does that take us?
Past 2030.
Okay. Okay, okay.
And so, you know, again, it's really those connected technologies in that portfolio that are enabling... And we're innovating with new architectures and then working with our customers to implement-
Mm.
... Those, new architectures. But that's the whole ballgame. Any one of those segments, it's always that race for that next inflection that gives you the 30% performance or power-
Mm.
... Or area scaling, all of those things.
You know, that co-development or co-innovation with the customers, is it different versus, like, five years ago or ten years ago? Like, were they doing it on their own, and you w.ere just throwing-
Yeah.
...The stuff over the wall, and?
I think very different.
Right.
I think that, you know, within Applied, we didn't have those teams in place that—so we have, again, world-class people that, you know, when we're going in and we can see, obviously, what everyone is doing, and we have firewalls a hundred miles thick, you know, between each one of—
Yeah.
... Those different customers. But yeah, it's completely different than the way we used to work. And I think the credibility comes from the fact that we've been able to innovate with new-
Mm.
... Ways to build those structures that are in chips today.
Mm-hmm.
And so the pull is getting stronger and stronger for that co-innovation.
Let's talk about a few of these. I'll maybe take one of them. Let's take Gate-All-Around-
Sure.
... First and get to the generalists in the audience that they may be for. It's a next-generation transistor structure, and we can probably talk about exactly what it is. But you guys have talked about, I think you already had something like $2.5 billion of revenue this year.
This year, yes. Yeah.
You thought it could double next year.
Yeah.
I mean, maybe just to level set, what are the types of of process and technology materials changes that have to happen in Gate-All-Around versus, say, the prior things, like the FinFET before that? Where is it?
Yeah. So it's interesting, when you look at this portfolio, for Gate-All-Around, you know, I was looking at the revenue growth for 10 business units. And, you know, this is where when you go from, like, a 3 nm chip to a 2 nm chip, the number of process steps are going up by more than 30%.
Mm.
You know, there's this, all of these different areas where you're creating new materials, you're shaping materials.
Mm.
We have very, very high share on that selective removal-
Mm.
... The majority of the steps there for the shaping types of processes. The material modification, when you think about thermal processes-
Mm.
... Or implant or CMP, again, those are, multi-billion dollar businesses. All of those businesses are multi-billion dollar businesses that are part of that flow.
Mm.
And then you have E-beam, where, you know, E-beam, where you have leadership there, about 50% market share, new technology innovation. So Stacy, it's really... And I talked about that connected platform that we have. We have those also in Gate-All-Around, so it's pretty broad.
Mm.
And I think the real uniqueness for Applied is the breadth of that portfolio, but the ability to connect it together in that co-innovation with customers. Because a lot of times when they're driving those inflections, you know, those are very hard technologies-
Yeah.
... To make work. And so just really high velocity in how we're learning-
Mm-hmm.
... With the leading customers.
Got it. Of the 2.5 incremental-
Yeah.
... How much of that is actually incremental versus, like, replacement in that revenue?
So the incremental revenue, if you look at 100,000 wafer starts, FinFET is about $6 billion.
Mm-hmm.
Gate-All-Around is $7 billion.
Okay.
So think of it, you know, there's about $1 billion incremental for 100,000 wafer starts.
Wafer starts.
And then we're capturing a little over 50% of that.
Okay, and then there's more wafer starts clearly that are happening in Gate-All-Around.
Yeah.
Okay.
Yeah.
Got it. Do you think all of the different customers... I clearly, I'm not asking you to talk about specific customer roadmaps, but they're all on, on different time frames. But you are working with all of them, I presume?
Oh, for sure. Absolutely.
Okay.
Again, these race-
Yeah.
... The race that goes on-
Yeah.
... Is life and death for all of those customers.
Yeah, yeah. Yeah. Maybe I'm jumping ahead, but like, what happens after Gate-All-Around? People have been looking at things like forksheets and then CFETs.
Yeah, I think you'll stack-
How far does that go? You said-
Oh, it's gonna go for a long ways.
Okay.
It will go for a long ways.
Uh.
I think that, you know, you'll stack the N and P transistors. And so yeah, I think that, for the next 10 years, there's a lot of fun innovation that will happen.
Okay. And I guess of the, you talked about, like, the content increase in AI servers. I can't remember. Have you ever, have you guys given a number to date for, like, where you think your, your, quote-unquote, "AI revenue" across all these different things are here?
Yeah.
I, I think one of your competitors gave a number, I think it was for every 1 percentage point penetration of AI servers was $1 billion-$1.5 billion in TAM. I don't know where the number came from or how they got it, but-
Yeah.
... Have you guys ever sized it or in some way?
Yeah, I think we think that. So we had an event early May, where we talked about some of these inflections. And we said data center wafer starts will pass-
Mm.
... PC and smartphones-
Mm.
... Going forward. And then AI servers, you have really the training on the data center part-
Mm.
... That's about 5% of WFE today.
Mm.
And then there's all the data generation that's going into-
Mm.
... The AI models. We think that ICAPS portion is about 20%.
Mm.
We think on the data center portion, that's a 30% compound annual growth rate going forward.
Mm.
So that's kind of the, you know, the shape of that.
Got it. Got it. So that grows, takes share, you take share within that-
Yeah.
... New processes, new materials.
Yes.
Let's talk about packaging. So this is $1.7 billion. It was what, it was $1 billion last year or-
Something like that.
Yeah.
A little over $1 billion.
Okay, and it's been growing.
Yeah.
And maybe talk about some of those inflections because, I mean, clearly, this is one of the things that has to happen as Moore's Law is slowing, right? Right, and-
Absolutely.
... Moore's Law doesn't mean you can't continue these technologies. It costs more money, you have to pay for it somehow. I think the other area where this is happening is, you know, people are no longer content with single buys. If you hit the reticle limit, you can't get it any bigger and-
Absolutely.
... Larger transistors, you got to stitch it. So, so you got to stitch them together. And so I guess, what is AMAT doing within this advanced packaging space, and how do you see that going? And what are the... Again, what are the kind of materials and other kinds of innovations that you're bringing here?
Yeah. So I think packaging is one of the most exciting areas in the whole industry.
How long have you been, like, at a packaging business? Because you didn't use to talk about it nearly as much as you do now.
Yeah. You know, we have a full-flow packaging lab-
Uh-huh.
... In Singapore, and so we have customers, leading customers, working with us on new innovations there, like hybrid bonding. So, you know, we've been investing in packaging for a decade. I would say that we've really accelerated investment there over the last, you know, 5-plus years, something like that. But we have, just like on the leading edge foundry logic or DRAM, you know, we have very broad set of products there.
Mm.
We, I think we said, out of overall packaging spending, we have about 30% share.
Is it all advanced packaging, by the way? Do you, do you do any legacy packaging?
We do have legacy packaging-
Uh-huh.
... But more of the money is going into-
Yeah.
... Advanced packaging. I think high-bandwidth memory, we said is growing 6x for us this year.
Yeah.
That's about $600 million.
Uh-huh.
We have a very broad portfolio, so PVD, Plasma CVD, ALD, plating, etch. We have hybrid bonding.
Mm.
Digital lithography, and then there's other new-
Mm.
... Inventions that we'll bring into packaging-
Okay.
... That will grow our share there.
You want to talk about any of those today?
No.
Let's talk about that. Let's talk a little about hybrid bonding, because this is something new that is, that is coming.
Yeah.
And for our audience, not everybody may be familiar with exactly what that is.
Yeah. Well, and so it's basically when they're doing the stacking.
Mm-hmm.
And a lot of this is about how do you drive power, you know, efficiency-
Mm.
... Energy-efficient computing? So in die-to-wafer bonding or die-to-die bonding, you're basically stacking the chips and connecting directly from one chip to the next. And so that improves that energy efficiency. And so when you look at the you know hybrid bonding, of course, I mean the high-bandwidth memory is now fusion bonding. That will go to hybrid bonding in the future because you want to increase the I/O density, and again that improves overall efficiency, and you're improving the power consumption.
Mm.
So it's really a way for you to stack vertically with better power-
Right.
... And better performance.
Got it. Got it. And you have a JV with, with, with Besi on-
Yes.
... Copper-to-copper hybrid.
Yeah. So, so again, when we're thinking about these inflections, so we have a really great position with a broad connected portfolio, but in that case, we saw a great opportunity to take all of the innovations that we have within Applied into a platform, partnering with Besi as the leader in hybrid bonding.
Mm.
So that's that combination with them.
When do you think that we are going to start to see this?
It's not a large amount of revenue.
Mm.
But it'll ramp to be pretty significant over the next few years. We haven't given an exact shape of that, but it'll definitely ramp to be meaningful.
Got it. And I guess, again, related to packaging, the Backside Power Delivery that you mentioned, maybe you could talk to our audience about what that actually is for, what it-
Yeah. So-
... I think it's very, very cool.
Yeah. So wiring is an area where we have very high market share.
Mm.
Again, I talked about the 100 km of wiring.
Yeah.
So it's basically taking the power to the backside of the wafer. So with that, you can enable about a 25% improvement, which is what customers have talked about, 25% improvement in power consumption-
Mm.
... And you can get up to a 30% improvement-
Mm.
... in area savings. So one of the things TSMC has talked about publicly-
I'm amazed you haven't said PPACT yet, by the way.
Yes. So one thing-
Yeah.
... That TSMC has talked about is their roadmap for energy-efficient computing. So what they said was that it not just on power and performance, but on area-
Mm-hmm.
... They have said that the design technology co-optimization has grown, where materials innovation now is about 50% of the area scaling.
Mm.
So things like Gate-All-Around, things like backside power, moving the power lines to the backside of the wafer, those are really, really important to hit those energy consumption targets.
Mm.
But also, you have area save, in, you know, basically scaling advantages-
Right.
... By going to those different architectures through materials-
Got it.
... At the same lithography.
Got it. So if you can get 30%, yeah, that means you take 30%. It's not that simple, but-
Yeah.
... Give or take.
You can put more transistors in there.
Yeah. So let's talk a little about the markets. So AMAT has, I'd call it, a more balanced revenue profile maybe than some of your peers. And maybe you could just talk a little bit about those, so your near- and long-term expectations for the different pieces of the market, like foundry logic and DRAM-
Yeah.
... And NAND. A nd, and where, what is AMAT sort of bringing to each of those, each of those pieces?
Yeah, so foundry logic. So when we look at the overall market, we think foundry logic will be about two-thirds in the future, one-third memory. Now, foundry logic is a bigger percentage versus memory, but we think-
It's horrible, right?
It's well, especially NAND-
Yeah.
... Has been horrible. DRAM's not so bad, actually.
Yeah.
But.
NAND, in your numbers, is not, it's spitting distance of zero at this point, right?
It's not significant. It's not a big number, for sure.
That's been a problem, actually, like, like on an overall basis. Yeah.
Yeah. So I think that, we're very bullish on foundry logic.
Mm.
Going forward, I think, again, this AI driver for the overall market will drive foundry logic content in leading edge for sure. Those die sizes are very, very large, so that's gonna drive more wafer starts.
Mm.
ICAPS, we're also bullish on ICAPS as you go forward. If you look at the transition to electric vehicles or renewable energy or, you know, all of those data generators, as you're transforming everything, the digitization of every industry, we think that market's gonna be pretty healthy. DRAM, compute memory, again, you have eight times more DRAM content-
Yeah.
... In an AI server. In high-bandwidth memory, the chip size is much larger-
Sure.
... And you need three times more wafer starts for high-bandwidth memory. So compute, memory, we're very, very bullish on. I think on in packaging, obviously, we see an opportunity to more than double the size of that business over the next few years. I think that keeps going at a high compound annual growth rate.
How much of your DRAM business today is high-bandwidth memory, by the way?
I don't know that we've broken out that as a percentage. It's not a significant percentage, but obviously-
Yeah.
... It's growing at a very, very high compound annual growth rate. And Stacy, the thing I would say also, is that, again, this race for the architecture inflections is there in every one of those different segments, including ICAPS.
Mm.
We'll talk more about that over the coming quarter.
Mm.
But, you know, if you think about power electronics and silicon carbide and GaN and sensors and RF and all of those different areas, they're not driven by shrinking feature sizes, but there are significant architecture changes.
Mm.
... In those kind of ICAPS markets. So again, that's the key thing in all of PC-
You're actually still innovating on the trail-
Oh, absolutely. Absolutely.
Got it. Is that semiconductor customers don't usually like change, right? And so if they're already using a process, I'm gonna make it up, I don't know, but if they're using a process on, you know, 90 nanometers or whatever it is, and you come in there, so you have some tape, some innovation, like, what gets them to adopt that?
It really is, how important is it from a system standpoint, if I can enable power electronics with, you know, faster charge time and longer battery life, and that-
Uh, yeah.
... Is really important to the end customers, then, you know, you work on those inflections. So it really is looking at it from a system to materials-
Got it.
... Type of a perspective. We'll share more of that over the next few quarters.
I can't wait. I can't wait to hear it. I want to talk about China a little bit. China was one of the things I think came in quite a bit stronger than we had expected.
Yeah.
It looks at the end, I'm gonna talk about China Foundry Logic, because there's something else going on with the China DRAM for you guys. We can talk about it in a moment.
Sure, sure.
On the foundry logic side, it does seem like they are adding a lot of capacity and maybe continuing in. I know you, as well as many of your peers and others in the industry, have suggested that you think, at least at this point, this is sustainable.
Yes.
Um, why?
You know, I think that, number one, we look at what our customers are forecasting and in those markets, and you know, we see that remaining healthy over a longer period of time. If you look at drivers for those ICAPS markets-
Mm.
... We think the compound annual growth rate. We formed our ICAPS group April 12th, 2019, five years ago.
Okay.
And, you know, we did that because we knew that that CAGR was gonna be a very healthy CAGR. And when I look at electric vehicles, that's really big. China's big in electric vehicles. If you look at renewable energy, China's big in renewable energy. You know, a lot of those ICAPS markets, if you look at the CAGRs just within China, the second largest economy in the world, and there is a lot of domestic demand, and you have inflections there.
Mm.
The whole foundation of the automotive industry, it's like a data center on wheels, 7,000 chips in an electric vehicle, that's inflecting. And so those markets are good growth markets, but there's a multiplier in any of those different-
Mm.
... Segments that you're serving. So again, we look at those markets as being healthy-
Mm.
... Relative to the demand drivers going forward, and then domestic versus, domestic demand versus supply.
Got it. Is all this going to serve more domestic demand? Like, I was like, who are these customers who are actually building these fabs? They're new, many of them.
Well.
Do they know what they're doing?
Yeah. I would say that the higher majority of our revenue is coming from people that are more experienced.
Okay.
But there are a number of those smaller companies, not as big a percentage of the total revenue, but there are companies that have emerged that have less experience. Obviously, the yields are much lower for those customers, but we track wafer starts for every fab around the world every month. We pretty much know where the yields are at for those, all of those different companies.
Mm.
And then when we look at the market, we basically try to start from the end-use demand. What, you know, what do we look at for end-use demand? And then, you know, back that into what the CAGR will be. And so we think ICAPS, over time, will grow pretty much in line, you know, with the overall market.
Okay. You ever guys ever given a, a view of the overall market growth? I mean-
Well, I think, you know, we think that-
Trillion dollars.
$1 trillion by 2030, we think that's definitely a reasonable estimate. And, you know, I think if you look at the big drivers like AI and some of these inflections that consume an enormous amount of chip capacity, we think that's gonna be a great growth driver for a number of years. And then you look at capital intensity-
Well, that was my next question.
Yeah.
What do you think as we go to $1 trillion, what do you think capital-
Yeah, so capital intensity, it's interesting. If you look at 2000, capital intensity was about 17%.
Yeah.
Then, you had the shift from 200 mm-300 mm wafers-
I believe that-
2.3 times the number of chips per wafer. You had the emergence of foundries, where you had a lot of used equipment. And you pretty much had, you know, no growth, you know, for that entire period of time. But, you know, and then the capital intensity with the used equipment, the wafer prices went down to about 9%, now it's back to 17%.
Yeah.
So, you know, I think that capital intensity is gonna remain very healthy. I mentioned, like, the 3 nm-2 nm-
Yeah.
... It's over a 30% step growth. So, you know, it's very hard, and this race for power consumption and performance is trumps everything.
Yeah.
You know, because, again, when you're growing so fast with those data centers that are consuming a larger and larger, and larger percentage of the world's power, or edge devices that are consuming a lot of power, you know, that's a huge driver for complexity and for the industry.
Does it really feel like there's any real structural reason that one should, I know in any year, anything can happen, but over the cycle, any reason that capital intensity shouldn't go down? It doesn't feel like it.
I think it's gonna—I think, again, it's really like magic, what we're able to accomplish. And so, no, I think it's, you know... And we can see this out many technology nodes into the future. I think it's going to become more and more complex, but I'm also very optimistic that we're gonna be able to deliver those innovations.
None of these nodes are getting cheaper anymore.
No. No.
Okay. Going back to China for a second. So, you know, lots of the industry is running at a very high mix of China revenues right now, ranges anywhere. You guys are actually, people don't realize, you guys are on the lower end of the big five. I think you were, like, low 40. I think, like, mostly China.
Yeah.
It was, like, 41% or something last quarter. I think ASML was at 49%. And now, you've suggested that China mix is coming down to-
Yes.
... Is that primarily the DRAM piece that's driving that?
Yeah, I think the DRAM is coming down. Again, what we had in 2023 was 27%, 2022 was 28%. You know, I think it'll come back down into the 30% ZIP code.
Uh.
And as you mentioned, you have display, you have AGS, and then DRAM, for sure-
Yeah.
... Is coming down, a significant amount.
Yeah.
So that's gonna bring it back down into the, kind of that normalized range.
That was just sort of related to the export controls, right? I mean, you were not able to ship to the Chinese dealers. You realized you could ship more than you thought, and they ordered, and now they've got enough? Is there anything more complicated than that or?
I think it's just based on those customers' demand for that point in time. We're not giving an estimate of what does that look like going forward, but, you know, I think overall, when we put it all together, we think it's around 30%.
Okay.
Something like that. There may be high 20s or 30%.
What, what are the implications-
20% for equipment, something like that.
Oh, I'm sorry. Say that again?
Well, the equipment we said was, again, 27%-28%.
Yeah.
Then 23% and 22%.
Yeah.
Semi equipment's 20%, then you have AGS and display on top of that.
Oh, okay.
Maybe low 20s for equipment, and then the rest, AGS and display, something like that.
That's... And AGS, are they over-indexed to China, like? Oh, display certainly is, but I mean-
Display is over-indexed, for sure.
Yeah.
But AGS is, our penetration there is similar to other regions.
I want to get to AGS.
Okay.
But I wanted to ask, so clearly, the Chinese customers, it would mean that the margin's been better with the Chinese customers, right?
Well, I would say that margins on... If I look at ICAPS-
Uh-huh.
... They're in the same zip code, and those are smaller volume customers.
Mm.
So whether they're China or a different region of the world, they're in roughly the same zip code for those ICAPS customers.
Okay. DRAM customers? I mean, is it... I mean, I guess what I'm asking is, as it comes down, how should we be thinking about even other margin drivers as, as, as well that are there?
Yeah. Yeah. So I think, yeah, China recently being up, you know, in the 40% range, definitely has been a adder to margins.
Yeah.
So, you know, we think the normalized rate is around 47%-
Yeah.
... right now. We were up at 48% before we had all the COVID supply chain headwinds-
Yeah.
... and you know, cost increases. We're still driving towards the 48%-48.5% that we committed previously.
Yeah.
It was delayed with all of the COVID-
Yeah.
... Cost headwinds. But I think there's tremendous opportunities for us to continue to drive margins higher. You know, I always tell people, "We don't want to waste a crisis."
Yeah.
The COVID supply chain challenge, we've really strengthened our operations and supply chain, and then we have been implementing price increases for customers. Then the key thing, overall, is enabling those architecture inflections uniquely.
So our ability to do that, we're creating a lot more value, we need to capture more value.
And, like, this is a question I asked Tim, but I mean, I'll ask you. Given all that value, like, why do the gross margins have to start with a four, like? I'm not asking for margin, like, guidance or anything, targets.
Yeah.
Do they have to, like, fundamentally?
No.
It doesn't feel that way.
No. I would say the only other thing, for us, the AGS margins are lower.
Uh.
Now, AGS is growing. You know, we've said we'll grow at a double digit-
Yeah, let's talk about AGS.
Yeah, double-digit compound annual growth rate. That is somewhat dilutive.
Mm.
But the operating margins are in the same zip code as the rest of our portfolio within Applied. Yeah, that business, we have a tremendous opportunity. It's at a $6 billion run rate right now.
Yeah.
And so you think about a double-digit compound annual growth rate, you could add $1 billion in growth just in our service business.
Mm.
And, you know, for customers with all of, all of this complexity, they're constantly. We have 200,000 chambers in the field.
Okay.
They're constantly ramping all of these new technologies across all of those different device types. And so for them, ramping to high yield as fast as possible is really valuable to them. And then high volume manufacturing, yield, output, and cost, with all that complexity, that's valuable to them. So that's driving this double-digit compound annual growth rate. Over the last 10 years, we've driven our percentage of subscription agreements as a percentage of our service revenue. It was about 40% agreements, now it's two-thirds-
Wow.
- agreements, and we're also shifting customers to higher value agreements.
Mm.
That's part of that double-digit compound annual growth rate.
Mm.
There's a lot of service innovation that we're driving also that will fuel that growth.
Like what?
Well, a lot of AI applications, when you're thinking about ramping new factories, matching chambers, you know, optimizing the yield, output, and cost, high volume manufacturing, new sensor technologies that feed the data into those AI models. We have remote connectivity with thousands of tools in the field already, so... And that percentage keeps increasing, so then we can deliver service faster-
Mm-hmm.
... Better. And then, you know, with those technology innovations, we're able to help them ramp faster with bigger process windows for higher yield, and then optimize edge die yield, which is really hard for those complex processes and output and cost.
Yeah. This business is great. Even in down years, this business can-
Yeah, we grew kind of mid-single digits-
Yeah.
... This last year.
Yeah, I guess the installed base still grows even in a-
Yeah, and again, we are shifting. We're growing the percent of service agreements, and we're, and there are higher content agreements, so we're shifting people to those higher content agreements.
Okay, sweet. Customers, a lot of companies used to just do their own servicing. Is that just... As these tools get more complicated, is that just really not viable anymore?
You know, we just have unique-
Uh-huh.
... Access to that data. We also have unique sensor technologies that we're delivering for the customer. So, you know, all of that asymmetry-
Mm.
... In our understanding of those tools and technologies give us an ability to deliver that yield, output, and cost faster and better.
Got it. Now, you have your 200 mm tool business in the AGS segment?
We do.
Okay.
Yeah.
So-
That's kind of mid-teens, so 85-
So it's not that big of a... Okay.
No. You know, the service business is, you know, around 85% and the 80- above 80%. Mid-teens is the 200 mm business.
Got it. Got it. Okay. Yeah. I want you to talk about some of the other areas that I traditionally maybe you weren't as strong as you've been doing well. The process control is one where I think you've been talking a lot, and you've done very well there, like, you know, in things like defect inspection review and that sort of thing, in E-beam.
Yeah.
You've been moving much more lately into optical.
Yes.
Maybe you could talk a little bit more about some of the opportunities you see there.
Yeah. So-
Cold Field Emission.
Yeah. Yeah. So that is a really important growth driver for us. I think, you know, going forward, that will be a high compound annual growth rate. It has been a really strong growth business for us. We have about 50% share, close to 50% share-
Mm.
... Of E-beam, and E-beam is really important. The key thing here, and this is part of this connected portfolio, when you're enabling nanosheets, or Gate-All-Around transistors or any of these big inflections, capacitor scaling and DRAM or so E-beam is a very, very, very important technology to drive the learning rate.
Mm.
So, you know, you're trying to... you have 100 billion transistors on a GPU, they all have to work. So the ability to image uniquely, we have leading electron optics with cold field emission. We're about 10 times faster in imaging, highest resolution. So that business is growing for us. It'll grow about 4x our revenue in cold field emission to about 50% of our total E-beam this year.
How big?
Over $1 billion. So, that business will grow for us. And again, the demand, as you go to these more and more complex structures, will increase for E-beam overall.
Mm.
And then in optical inspection, we see opportunities to grow there too, to expand the TAM. Especially with new technologies that we're bringing to market. So that one can be a, a significant growth rate for us. But again, there's also the multi-billion dollar synergies on time to market for those new innovations. So that's incredibly synergistic for the material-
You talked about leveraging some AI methodologies between optical and E-beam to drive, like, the classification-
Yeah.
... And identification. I thought that was really interesting.
Yeah.
Can you talk a little bit? I understand it's probably dumb about it, but, I mean, I thought you were using E-beam to identify specific defects, and then using optical. Optical, you can scan the whole wafer, but you don't pick everything up. And using that through a neural network to identify areas-
Yeah.
... Where there are likely to be defects without actually having to go and find them first. Is that, is that kind of how it was working, or?
Yeah. I think there is another aspect of AI-
Uh-huh.
... Inside Applied that, you know, gives tremendous benefit. So, both in the detection, the review, measurement of key structures, we're using AI very, very heavily, and so that'll-
Mm.
... Accelerate the value from that business for sure.
Got it. Got it. Now let's talk about Sculpta.
Okay.
This is a modification tool that AMAT sells to reduce the need for multiple patterning on extreme ultraviolet lithography. But I thought the way-
Yeah.
... It works, it's slanted edge and everything.
Yeah. So it's a directional pattern shaping technology.
Yeah.
So it's a new technology for customers, and when we first introduced it, we said you could use it to reduce EUV double patterning. And it's, you know, a $200 million business today. We said it would grow to $500 million over the next few years, but customers are finding more applications for Sculpta.
Like what?
So you have certainly the EUV double patterning, but as you're, you know, directionally removing material, you can improve line edge roughness, you can eliminate bridging defects. One of the key technologies, even, you know, for EUV or high-NA EUV, is tip-to-tip spacing.
Mm.
Sculpta is just very, very precise-
Mm.
... In being able to move that tip-to-tip spacing to smaller and smaller feature sizes.
Mm.
So there are multiple applications that we're seeing customers adopting with this new tech, directional pattern shaping-
Mm.
... Technology.
Okay. Can you talk a bit more about your, your integrated solution?
Yeah.
You have a new platform that actually, I think, really... I can't remember, Vistara, is that the name of it?
Yes.
Number one, tell us about that, and number two, there, there was something else you, you talked about, you could incorporate third-party chambers into this platform, and I still am unclear on how that would actually work. So like-
Yeah.
... Like, what does this platform bring to the table?
Yeah. So in terms of power consumption and then, your output per sq ft, throughput density, there's about a 30% improvement in that platform. And then the other, there are other two aspects of it, besides this, sustainability improvement in power consumption and throughput density, intelligence built into the platform. So getting back to all of the sensor technologies that help you optimize the yield output and cost-
Mm.
... You know, that's built into that, Vistara platform, and then flexibility. So Vistara can improve all of those metrics, but you can also, we talked about before, I talked about before, that seven technology in one, platform, and our percentage of integrated growing from 20% to 30%.
Yeah.
I think the first thing for us is integrating our own technologies into this new platform with this increased flexibility. Vistara is designed in a very flexible way-
Yeah.
... So that we can plug in
It was linear, right? It wasn't-
It... Yeah.
... Perpendicular. Yeah.
Yeah. So you can plug in all of these different parts of the Applied portfolio. So that's our primary focus-
Okay.
... With that kind of flexibility.
Okay. I'm assuming all the integrated stuff, like, do you sell an integrated per chamber, or do you sell, like, the solution? How does that-
The solution.
Okay.
It's a platform.
Presumably, you can get higher value-
We do.
Yeah. Okay. And that goes for Vistara as well. Have you ever just started to ship Vistara yet?
We have, yeah.
Yeah. Okay. I don't know if I can ask how many you've shipped or delivered.
I don't think we've talked about-
Yeah.
... You know, it's mostly in memory right now.
Yeah.
That's really the first place that we've had adoption.
Okay.
With Vistara, Vistara.
Okay. Got it. You know, I always wonder about competition in this. Funny, like everybody said, they put out targets, and everybody there has always been that they're gonna take share. The market shares in general are fairly stable. I guess two ways, how do customers, like, when think about when they're evaluating a, I guess, for a new bake-off, how do they like, where does AMAT differentiate on a specific like, bake-off, I guess? And then number two, specific to China, how do we think about the rise of local Chinese semis, 'cause I actually am of the opinion personally, that they will take and are taking more than their fair share, I think, just given the nature of the situation they built?
Like, how does AMAT think about how, that situation, how to deal with it?
Yeah. So I think, you know, if you look at 2011 to today, we're up from 16% to almost 22, 21.6%. So we have gained share, certainly DRAM, 10 points of overall, WFE share.
Yeah.
But the key thing is, again, how you're positioned for those big architecture inflections. So you know, for us, I talked about Gate All Around, backside power, 4F squared, any of those big architecture inflections. We're working out the 4 technology nodes into the future. And you know, for us, it's really shaping the portfolio, that connected portfolio. So we're improving our competitiveness in all aspects of the portfolio, from critical applications to semi-critical to non-critical. Tremendous innovation going there. But then, even more important than that, is how we connect that all together to enable big inflection. So when you think about any of those architecture inflections, whether it's in ICAPS or the leading edge or packaging, there are tipping points.
Mm.
For those architecture inflections. If you're enabling that architecture inflection, it puts you in a completely different position. Because, again, you have a portfolio of highly differentiated or competitive or less competitive parts of your portfolio, but if I'm really providing what makes you competitive and, and really helping you win that race to that new architecture inflection, then, you know, I'm strategic in how we allocate our innovators, our innovative technologies, all of those types of things, that's a completely different discussion. And I would say that's changed a fair amount.
Mm.
As we talked about. You know, the way we engage with customers, how many nodes we're working on with them, and our role in that co-innovation with customers have changed a dramatic amount.
Mm.
So I think in any of the markets, Stacy, that is the most important thing. And then, so for us, we've outperformed 5 years in a row in terms of wafer fab equipment, and we're well positioned. By looking at these big inflections, we talked about 50% of the incremental spend for many of those big inflections. And again, more and more of these inflections are about these magnetic materials and those architectures. So that really, absolutely, by far and away, the most critical thing for our customers. And then if you think about our customers in any region, you know, we've been competing with domestic companies in Korea for 2 decades, more than 2 decades, and they've made progress in some of the non-critical types of applications.
Mm.
But for our customers, what they care about is their position versus their competitor.
Mm.
And so they never, ever compromise-
Mm.
... That competitiveness. So again, building that portfolio, shaping it, connecting it, and then being essential to those inflections is really how you grow share. That's by far and away, what's most important.
Got it. Got it. So a few minutes left, we've got some audience questions.
Oh, sure!
Ooh. Is there anything that would change your view about capital intensity remaining high and making 9% a feasible scenario again?
No.
No. We're in a can't get enough situation.
I don't see that. I mean, you know, there's not gonna be 450 mm wafer inflection. You went from 200 to 300, 2.3 times the number of chips per wafer. Die sizes are getting bigger. HBM required 3 times the number of wafer starts. CPUs, they pack as many transistors as they can in reticle field. You know, I don't see that.
Okay. What are investors not talking about today on AMAT that they will be talking a lot about in 2030?
I think it's really, again, our role in enabling these inflections. I think people, again, we see it because we see that four technology nodes out in the future. I think people really, the amount of those inflections that will be enabled, the contribution from materials innovations, is something that is growing. And again, I talked about TSMC, the roadmap for energy efficient computing, and more and more of the innovation coming from these magic materials. And that's true in ICAPS, too. That's something people talk about, trailing edge technologies. There's innovation happening there-
Yeah.
... That people really don't-
It's so hard, right?
No, it's really hard. And how they compete, people really don't understand that. That's something we need to do a better job of helping people understand.
Can you speak on the opportunities that excite you the most in advanced packaging, and how much do you think the front end versus back end spend might get long term?
So again, that's a $1.7 billion business for us today. And if you, if you look at what people are talking about, this, these roadmaps for dramatic improvements-
Mm.
... In power consumption, energy efficient computing, there's gonna be incredible innovation in packaging technology. So we already have the broadest, most connected portfolio. We have a unique packaging lab, we're co-innovating with customers on those new technologies. We have some new opportunities to expand our TAM. You know, some that we've talked a little bit about, some we haven't talked about. But, you know, I think those architecture inflections are absolutely crucial to the whole industry hitting their goals for energy efficient computing. So, again, this is one I'm spending... When I go back Saturday, I'm-
Uh.
... Meeting on this topic. We spent, we're spending an enormous amount of time here.
Uh.
Partnering with customers and also ecosystem partners.
Got it. That's helpful. So Gary, we've got about one minute left, and you've been doing it all along, but, I will give you your one-minute soapbox. Succinctly, like, why should investors buy AMAT stock today?
Yeah, I think that our markets have never been in a better position than today. You look at the drivers in this AI era, they're bigger than that compound annual growth rate is gonna be more significant, more pervasive computing than we've ever seen. Energy-efficient computing is a big drive in all aspects of the market, and Applied is just incredibly well-positioned for those future architecture inflections with a very unique portfolio. So I think that's the main thing. But on top of that, you've got a $6 billion service business that has a great opportunity to grow at a double-digit compound annual growth rate going forward.
Got it. I think that's the best place to leave it at. Thank you so much.
Okay, great. Thank you.