Applied Materials, Inc. (AMAT) Earnings Call Transcript & Summary

I'm Vivek Arya. I cover semiconductors and cap equipment at BOFA and really delighted to have Brice Hill, the new Chief Financial Officer of Applied Materials, along with Mike Sullivan, who we know, Head of IR. And this is Brice's -- I think your first kind of sell-side conference?

At Applied.

Yes. At Applied.

At Applied, right, yes. Thank you.

Right? And I thought it would be great to have Brice here because, first, I think kind of new, fresh perspective, that'll be good to hear. And then also, in his prior life, Brice was at Xilinx, and I think before that at Intel, where we actually met. So I think he brings a very interesting kind of customer perspective to the role. So it will be really useful to get his views on the industry. So maybe let's pick up from that Brice. This is, I think, what your 3-month anniversary.

Yes, that's right.

So give us your perspective so far, the good and the bad, hopefully, no ugly is in there, but what have you liked so far? What do you think you need to be working on?

Well, great. And thanks for the invitation. This is fantastic. Appreciate it. Good to see the -- all the friendly faces here. 3 months at Applied, I think it's obvious what the principal challenge has been, and that's been supply chain challenges across the world. The shutdowns that were COVID-related, especially in Shanghai, happened on March 28 and really affected the ability to finish the last quarter. April is the last month of the quarter for Applied. But on the positive side, if you can imagine, rescheduling and reworking your production plans almost every day and surely every week. As your perspective of which supplies come in and how to optimize your output for the customers. We're trying to serve our customers. Our customers are asking for us to speed up and get them tools as quickly as possible. And so I think in a good way, the agile reaction and the execution focus of the company, despite the challenges has been amazing. I mean we were super pleased even -- we were super pleased with the output last quarter despite those challenges. And I think it's the execution focus and the ability to be agile. We have factories with tools that are partially built And we're shipping tools partially built to customers so that they're ready to go once we get that final part for those tools. That's a huge logistics and operational challenge. And I think the company has just been heroic in responding to that from an execution focus over the past few months. And then what's going well, this industry and our company has grown a lot over the last 3 years. We're trying to grow capacity. So we got frustrated by that supply chain reaction. But we are making significant progress with the supply chain and the ability to grow output once we solve this more short-term capacity issue. So I think that's also been going well. If you listen to our CEO, he's been working with all of the key suppliers on getting to this new level of output. We grew our inventory last quarter by $500 million. So we've got the orders, the inventory. We just got to close the circle to raise the ceiling going forward.

Absolutely. So Brice, would you describe the challenges as more component availability or logistics? In that, there is a perception that as China starts to reopen, that could help right the environment, at least from a logistical perspective, do you think it could make the component demand situation perhaps a little bit tighter because now there is maybe incremental demand, which is not there right now.

Well, it's interesting, whether it's component or logistics. By the way, when you say component, do you mean like tools?

Exactly. I see that you need to complete your tools.

Yes. I wouldn't really pick processors or chips out as the target for these shortages. I think for us, it's more think subsystems. These are factories that also have their own suppliers in the network. And that's really what has been affected, at least with this result. There are chip shortages, but we're generally able to work through those in a reasonable amount of time. And the ecosystem is sort of on our side with those because they know if Applied is able to ship more equipment, then that helps the ecosystem recover from a chip perspective. But the last thing I'll say is just restarting those factories and getting their supply chain to normal because they all have their own suppliers, there's just a start-up phase. So there's a little bit of an air gap for us going into this quarter.

Got it. Would you describe -- how would you characterize the situation as it exists today? Is it still kind of a firefighting mode? Or is it kind of the -- I'm trying to get you to kind of characterize the situation. I don't know if there's an easy characterization, but is it a firefighting mode, is it a consolidation phase. Like do you have line of sight into when we get to the other side? Or is that still work in progress right now?

I think it's work in process, but we would say also that we have line of sight because we are working. One of the things that's happened over this last year is we are working with all of these suppliers to understand their build plans, to understand how the people, if they're in that region, if the people are coming back to factories, what their output levels are. And it's just a matter of making sure there's no more hiccups in that to refill the supply chain and then allow us to -- like I said, we have more inventory. We have the orders. So it's on these bottleneck components getting those into inventory position. For example, I just -- let me go one more detail just to say, this quarter, our Q3, we will finish this quarter with 0 inventory on those bottleneck components because we'll have used all of those to ship every part we can this quarter. So you start Q4 still without inventory on those bottleneck components. And we hope during that Q4 period, that will begin to normalize because their output will catch us up once the factories are up to full speed.

Got it. You think you're competing for those bottleneck components with a number of other peers in the equipment industry? So how do you then ensure the right allocation to apply it? Or it doesn't matter because if the customer, they need a tool from you or one of your peers if they can't get all the tools, then it doesn't help.

It probably doesn't help like you suggest. But in general, I don't see a competition. We're working with the suppliers. They have teams that are dedicated to production for us when those people come back to factories they're working on our subsystems. So there hasn't been that dynamic in the equation.

Got it. How much of the tools you were planning to ship this year is for end demand that is for this year versus end demand that they were kind of preparing for next year, i.e. is the fact that the equipment industry, right, is facing these headwinds, is that kind of hurting demand in some ways? Or was that always expected to be demand that would have translated into next years?

It's definitely hurting supply is the way I would say it because when we look at our utilization, you've heard us talk about we see -- partially because of our service agreements, we see the tool ecosystem, and we see how utilized it is. And right now, it's at record levels of utilization and customers -- their end customers, our customers' customers want more product at this point because they're even calling us to say, "Can you speed up?" So I do think the supply is crimped relative to demand. We think of the market as underserved at this point. So it is having an effect from that perspective.

What would unconstrained WFE look like at this point?

We've said it's definitely over $100 billion. And we're probably still operating at the current level in the $90 billion or low 90s level. So we've got the ability to jump up once that supply chain normalizes. And we have the capacity to do that.

Right. Which would make sense because last year, WFE was whatever, $86 billion, $87 billion, which means the exit run rate was in the low 90s and industry sales have kind of just hung in, right, at that level. So does it mean potentially that WFE instead of being $100 billion this year is something in that, whatever, mid-90s plus/minus, and that we potentially have a 100-plus year next year. So we have another year of growth. Is that a reasonable scenario?

It's absolutely a scenario. But the first -- what I thought you were going to ask is, can $100 billion still be achieved.

Lets ask that question first then.

Right. We still have our -- for us, our Q1 ends in January. So we still have our Q4 and our Q1 that we count in the year's production. So we do think there's -- it's going to be dependent on that supply chain. And whether that supply chain can help get us into a healthy position in Q4 so that we can raise the ceiling on output and have a better Q4 and a bigger Q1. So it's still alive from that perspective. But then going back to the way you asked your question, when we think into next year, regardless of where we look -- where we stand this year, where we finished this year, we've got a significant backlog. It's multi-quarter. We've got really strong orders coming in, in each quarter, so it's adding to it. And then we also track the number of factory projects across the world. In our last quarter when we did earnings, we said that we see 68 projects. Now we see 80 projects. So for us looking forward into the next year with that strong backlog and with the contact that we have with customers, we think next year will be stronger. And so I would look at it as there's some buffer in terms of pent-up demand that will help drive next year.

Got it. What's the cycle time conceptually for Applied? And that at what point of the year, will you say, you know what, even if I get those missing ICs today, this is just a tool that I just cannot ship this year, right, just given the cycle time to actually build that.

Well, unfortunately, it's close to 0 because -- and the reason I say that, if you can picture when we finish a quarter, we actually have tools at customers that are waiting for a component or we have tools on the floor that are 99% complete, waiting for a component or two. So what's happening when these bottleneck components show up, they immediately get installed and either sent to a customer for a tool that's ready to go or installed in a tool that we'll ship the next day or will even ship that day. So -- and that's where I went back to this, the agile nature of the execution team managing all that. That's obviously something that doesn't happen in normal times. And that's been going very well to keep the machine running despite these choke points.

I see. So I guess to kind of make sure we understand that very important point, which is we should not be giving up hopes of the industry being able to achieve, right, the $100 billion because let's, for lack of a better phrase, call than the magic screws. If they do come that you can actually put them in the right place, so the industry achieves that number.

Right. We get those fasteners and then we'll close the loop, yes.

Got it. Okay. The next thing from a kind of near-term perspective is demand from Chinese customers, right? It's been an important part of the industry for the last several years. Just in the near term, there has been some concern about restrictions and so forth that the government might place on them. Have you heard anything along those lines? Do you think might Applied might be restricted? Or what's the right way to first tell us how big is China domestic for you? And is there a threat of those restrictions?

Sure. Well, we report the total China number. So 34% of our most recent quarter revenue was shipped to customers in China. And the way we think through the question you're asking is we kind of break that business into its component pieces. So a portion of that is our display business. A portion of that is our services business. And then the rest is split between multinational companies that have factories there and then local companies, which are largely ICAP's investments, in other words, more mature technologies that they're building. So first, we do get questions from the government. We actively respond to questions. There are trade rules in place today that we're compliant with. We apply for licenses, and we work with the government on that. And we've been successful at getting various licenses. So I think our perspective is if you think through each of those businesses, you would have to handicap whether you think there's going to be any controls or restrictions on those. Maybe it's less likely if you think through the ones that I named. But even if there is, the way we do our demand testing, as we try to triangulate our demand, is we look at the overall semiconductor industry and we look at the growth of the semiconductor revenues and we test whether the WFE, the equipment spend makes sense relative to that demand curve. In our perspective, that would be that if we're, for some reason, restricted from shipping some of that equipment to China that it will go elsewhere because of that demand curve. We believe in that. We think it's happening. It's been consistent for the last few years, and it's forecasted to continue for all the underlying drivers that we know.

Yes. Got it. Is that China demand a big part of growth in the next handful of quarters? Or is that a small part. So let's say, hypothetically, if it were to go away or restricted or whatever, I imagine it doesn't just shift over to some other customer, right? There's probably a lag in that.

I would characterize it as steady. So we saw an acceleration over the last few years in a number of new businesses and new projects. But as a component going forward, it's just more of a stable part of the WFE demand curve.

Got it. Have you seen any impact of slowdown in certain consumer markets? Because this is one thing that because of all the macro headwinds and crosscurrents in the market there is, again, this pervasive feeling consumer is under a lot of pressure, PC, smartphone demand can weaken and ultimately, it does impact your customers. So have you seen any change in their behavior to reflect that.

We definitely have -- you probably heard us talk about our display business. So that's the place where it seems to be more high beta with consumer demand cycle. So the TV, consumer PC, smartphone demand purportedly, those are lower and driving a lower what we call DFE forecast for next year. And so on that business, in particular, with a lower forecast for next year. We decided to reshape our road map. We delayed some of the products on our road map or reschedule them, which allows for a lower spend. And our perspective, just to tick in on that business for a second. That's -- we've largely made the IP investments. We've largely made the factory investments and so that is a strong cash return business for us. And so the resizing that we've done, I think, puts us in a position where we'll still have a healthy cash return from that business. And it gives us the option to grow the business if LCDs start growing again very quickly or the OLED technology takes off. So it's kind of like an annuity with a strong option if that happens in the near term. But that's the place where -- and then some people have asked us, well, okay, if you saw that in the display business, why aren't you seeing that in the semi business? And just the reality is because we kind of went through the stats. We got more orders, we've got super high utilization. We think it's just so underserved at this point at this very moment in time that if you take some of the demand off the top because there couldn't be some off the top, unfortunately, it's still underserved. So we're just not seeing a change in the equation at the short term.

Got it. So no slowdown or impact from memory customers or...

We have not seen that yet. So...

Isn't that surprising, though, Brice, I mean like one would think, again, this is more an external perspective, right, that it's so apparent that many of these consumer markets are slowing down, consumer is under a lot of pressure because of inflation or other reasons, right, China weakness, Europe, turmoil, that isn't it inevitable then why isn't that slowing down, like?

I don't know if it's inevitable. The way I think about it is there might be some buffers. There might be some lags in making adjustments if you're on the build side of this equation. So if you have construction labor, skilled labor in place to do factory projects, and you're ordering equipment to build out those factory projects. That may be a little bit more sticky than the monthly demand signal. The other thing that we think about is when you look at the wafer starts, the wafer starts haven't gone down. So you should see companies asking if that were happening, you should see companies looking to reposition orders, and you should also see them cutting wafer starts because they'll get into an inventory position in their components. It can happen at any moment, as you're suggesting, but that's not what we've seen so far.

Understood, yes. Next thing is, I think on the last call, there was a suggestion from Applied that '23 WFE could grow off of '22 that -- '22 again, if the industry gets the supply that it needs, $100 billion and then '23 could potentially grow off of that. What is that confidence based on?

Yes. I think it kind of goes back to those indicators that I've been mentioning, but the strongest would be new orders and a multiquarter backlog and a signal from customers. The one new thing in the environment is customers are starting to give us longer-term visibility like what they actually see into '24.

Right.

So we look at all those indicators, more fab projects underway larger backlog, strong orders, customers telling us what their plans are into '24. It gives us a lot of confidence into '23.

Got it. Do you think some of that visibility into '23 or '24 is predicated on your customers, getting this funding from U.S. CHIPS Act or EU CHIPS Act or whatnot? Or is that incremental to the signals that you're seeing from them?

I don't think it's incremental. I think it adds confidence to the customers in terms of their ability to start the projects and make the investments. So I think it is adding confidence for them. I don't picture it as -- at least relative to the demand set we see today is a trigger that will even lift the WFE higher. I think it's factored into some degree. And part of the reason I say that is, again, we triangulate it with the demand function itself on semiconductors. And so we do not think that's getting out of line. We use the syndicator. We talk about as WFE intensity relative to the semiconductor revenues. And it's increased a little bit, but we think it's still in the same band that it has been. And there's a little bit of a wild card because the semiconductor companies will probably raise their prices or are in the process of raising their prices. So you might see the denominator in that equation go up a little bit. But anyway, back to your question, we aren't thinking that those projects would significantly change the nature of that relationship, meaning a huge spike up in capacity that doesn't make sense relative to the demand function.

Absolutely. So when you talk about the growth side. Last year, it was all about how fast can these companies grow and now this time, right, what will they do in the case of a downturn, right? The discussion has -- pendulum has shifted, right, 180 in that. And this concept of maintenance WFE, right, comes up that if, let's say, if we do get into a downturn, right, what is that minimum level of spending your customers need to do to run these -- what is the conceptual way to think about that Applied thinks about that?

Well, it's interesting. We talked about that before and this concept of maintenance. I guess if you look at it like -- if there is a slowdown, what's the level of WFE in the ecosystem, we've looked at our curve over the past couple of days, it looks like over 10 years, there were a couple of years where there were single-digit downturns, I think, is what we were seeing in the data. So maybe from that perspective, that's one answer. When I was thinking about the term maintenance and trying to sort of get into the mindset, I just want to share that the customers -- I think the way they think about the business is, they have to build new products. It's fresh fish, right? They have to build new products. New products mean more transistors, 30%, 40% more transistors. And so when they think about it, even if they have a stable base of factory equipment, they're probably going to have to grow that amount of factory equipment because in order to keep up with the amount of transistors, they want to put in new product. So my own view of maintenance is there is growth in floor space and there's growth in tools that just go along with putting more transistors in the product, which equals more performance. So it's a way of sort of scratching what's the maintenance level I think it's tough to get at. I don't think it's as easy -- I think it's complicated because they're planning road maps and they have to deliver against those road maps.

Right. And to your point, we saw this yesterday with Apple right, when they launched the M2 that die sizes are 25% plus larger. So even if they make the same number of max, I mean, die sizes of those processors are larger, right, that's a lot more.

And this is a really interesting complexity there in that because the shrink level of new process technologies maybe isn't shrinking as fast as it did before. You're ending up needing more silicon real estate to add that many more transistors to a new product. And this gets into a space that Applied is investing in, which is sometimes those trips will just get too big all by themselves to build and yield successfully. So another way a customer might go at that is by splitting up the chip into chiplets or different components. And then the problem used to be, if you can't connect those at high speed, then it runs really slowly, and it's sort of -- it's not a good solution. today with some of the modern package technologies, the hybrid bonding, the interconnect capabilities that Applied is working on, you can split up your large device into chiplets connect them at high speed and solve the problem of evolution, if you will, or increase performance by doing that. So we think the industry will find ways around this, but you hit the dynamic. I mean, there is going to be more transistors because there's more performance needed right now, that means more fab equipment and also means probably new technologies to solve some of those problems.

All right. This might be going a little bit into the weeds, but does this move to chiplets? Because I imagine with chiplets, you're able to constrain the die size, right, but you just need a lot more of them. But as you constrain the die size, yields are better, but because you have more of them, right? Is it a net accretive proposition to somebody like Applied? Or is it neutral to your business?

We do think it's accretive because there's still going to be the need for monolithic chips. So it's -- there's not going to -- at least in my perspective, now this is opinion, in my perspective, there's not a one-size-fits-all solution. Your design engineers will think about what the optimal solution is for their particular application. And sometimes it's going to be a monolithic chip and sometimes it will be a set of components. In any case, even those templates, even if they go down the chiplet path, some of those chiplets will need to be on leading edge. They'll need the performance, the power efficiency -- that's right.

So speaking of trailing edge, if you dissect your foundry and logic exposure, how much of that would you classify as kind of leading edge versus -- where do you draw the distinction because there are so many different numbers. Where do you draw the distinction between leading and trailing -- and what's that split for you today?

It's about half and half, and we're drawing it at 7-nanometer. So 7-nanometer and I think everywhere that I know of is...

Intel is all trailing edge then?

I don't think so. I don't think so. I mean there -- you'd have to keep up with the names and their processes. But I think the way we think about it is the 7-nanometer node generalizing 7-nanometer as a type of node applies EUV technology. And the mask sets are very expensive. It's a higher type cost. If you go back to mature nodes where they don't have EUV, the mask sets are cheaper and the yields, in some cases, it's a lot easier to work with from a design perspective. That's where you see products that don't need the bleeding edge or don't need the leading-edge performance, we'll find better solutions in that space. And the way -- one of the examples I'd like to give, Vivek, is if you look at 28-nanometer, 28-nanometer has oceans of capacity. It's the largest node in the world, and we're adding -- as a world, we're adding a lot more capacity to 28. One of the key products on 28 is video processor and a video engine. And when you think about Machine Learning and putting that capability in so many different places, that's one of the key drivers of that. So I think that's a good example of trailing edge use.

Got it. Speaking of trailing edge, there is also this concern that maybe the industry is starting to overinvest because for several years, whether it was because of trade tensions or other reasons, we saw the Texas Instruments, right, analog devices, right, these kind of IDMs constrain capital spending. But now we are seeing their CapEx go on a much larger revenue base to a much higher level of capital intensity. Does that worry you that the industry is on the trailing edge over investing at a time that at a point where it may be because of a cycle or whatnot that we could be headed for a deceleration.

Definitely not worried about that at this point. I don't want to dismiss the concern on him, but just if you take apart what's happening, at least the way I think about it, you did have a lot of mature companies who are using mature capacity, and they were able to serve their business on mature capacity and have low capital. But if you look at the components themselves, analog components, MCUs, sensors, devices, et cetera. You look at the stack of those components and the demand forecast for those components over time. And it's 6%, 7%, 8%, 9%, 10%, 11% CAGR for 10 years for those components. So I think what's happening to those mature companies is they're having to really add capacity to serve a growing demand. And then when we think about the intensity of that ICAPs business, it is a lower intensity than the leading edge. It's less than that 15% that I said, which is the average intensity, and we think that reflects a good portion of their revenues are served by existing capacity. But as they add new capacity, that's looking more like leading edge did a few generations ago. They have to make new investments. And so -- anyway, it's a long winded answer, but we see high utilization there. And those companies are smart companies. It's hard to believe that they would get too far over their skis from a capacity perspective. And when we look at the utilization, we're not seeing that.

Got it. Because that, I think, again, is this disconnect where from the outside, we look at just raw CapEx, right, going up 30%, 40%. And we say, well, this is in an industry that hardly grows, right, maybe 1 or 2x GDP. And then the companies themselves are saying, well, CapEx doesn't equal capacity. That capacity is not growing as much. And I think investors that have a tougher time aligning the two. That's what I'm trying to understand that is there a difference between their CapEx requirements and the capacity that it's going to output?

Yes. This would probably be an interesting concept back to your maintenance capital question because I assume -- I haven't studied in detail those companies, but I assume those types of companies, were spending capital to keep up their equipment sets over time, but maybe it was largely an existing base. All of a sudden, when you have to add new capacity and new floor space, those investments are a lot higher than what they were 20 years ago or whatever the aged portion of that. Just the skilled labor, the components that go inside, fab costs more, even if it's not leading edge than it used to in. So if you have to -- if all of a sudden, you're in a position where you're adding space, that's going to be capital that gives you 0 wafer starts until you start to equip it, right? So without digging into it, I can imagine that the capital looks unusually high relative to their history because the price level is a lot higher and then also they have to probably add space.

Got it. The last place you were at Xilinx there. I imagine you were used to -- 69% gross margin would have been a bad year. I think I think right -- so now it's probably in the 70s, I imagine, somewhere. So why has the semiconductor industry being more successful at passing along cost and logistics inflation versus the equipment because one could argue that yours is also equally in some cases, may be more important right? So why isn't this industry being more successful in passing along cost inflation?

Right. Well, I think there is a range in the semiconductor side, but focusing on our business part of it is, I think, that there's just sort of this long set mentality that Moore's Law and in this ecosystem, cost of transistors are going to go down. And it was a big part of the focus of the industry, and there wasn't a lot of inflation. And so companies built the muscles to try to get in that equation if they're putting steps in a process within a certain cost envelope to have that process still makes sense so that Moore's Law still made sense, right? If you think about that cost per transistor curve, it was a kind of a constant curve in the companies that are building those processes, every single step, they're trying to fit into a cost envelope. Unfortunately, the dynamic has changed, and it's become more complex to get to the next levels. And so I think that equation is changing, which opens the door to a company like Applied. We're trying to solve problems working with the customers, and that's one of the things that attracted me to the company is really knowing what problem is the customer seeing? What is the next-generation high-tech problem that Applied can solve, that provides a lot of value to your point. And I think changes that old dynamic of maybe sort of a blind selection, let me find a supplier to do this one step and only the customer knows the whole picture. Now it's like the problems are much more difficult. We need to work together on the problems. I think that changes the dynamic and helps us have a value discussion with the customers.

Got it. So as you look at all these new projects, right, you mentioned 68 projects going to 80 projects and so forth. If you were to look out 3, 4, 5 years, do you think gross margins kind of stay in this 46%, 47% range? Or do you think that there is ever a scenario where, dare I say, start with a 5? Or is that too much wishful thinking at this point?

Well, the only thing that we've said so far is we do expect -- we do maintain our perspective that we'll be able to get to our model. We had one model that was 48.5%. We had another scenario that was 48.8%, and we think those are reasonable midterm targets for us. And just really quick, when we don't need as much overtime in the factories when this air freight for every single component goes away because the supply chain starts normalizing. And when -- to the extent that we have to pay some agent fees to find chips, if we're in those sorts of situations, when those things fade away and volume goes up, we get some immediate improvements from a margin perspective. Pricing will also help us move the needle back to that 48.5%. And then I won't make any specific comments yet, but I'll take that as an action item from when we do our next Investor Day.

Start with the 5. This would be better.

Maybe one quick comment as well. So that's a corporate blended average, right? And the services business would be on the low end of the spectrum and the display business would be the next up on the run. And within the semiconductor, which is the highest level, there's quite a range...

Process controlled, I imagine being much...

It's a good example and not the only example.

Got it. So between where 46% or so where you are right now to the 48% or so, if you set aside pricing because I imagine that takes place over a period of time. Is it a matter of just getting revenue to a certain point?

That helps a lot because in order to get the revenue, you need supply.

Like if you do $100 billion WFE, does that?

It helps because if you think about capacity, your absorption of overhead in the factories, it immediately improves when you have more volume. And so that will be significantly more volume and it helps from that equation. And then if the supply chain starts to normalize, and we're able to use more ground, ocean freight than we are currently, and we have lower expedite fees that also immediately helps I kind of gave you a picture of the factory situation. There's a lot of overtime in the factory to manage in this environment, so we can also manage that. So we think there's things that we can quickly adjust once the environment normalizes from a supply perspective.

Maybe let me ask the question in a slightly different way, which is, do you think there is further inflation that we have not thought of yet? Because labor, you brought up labor. Like are there other -- is it possible that as an industry gross margin maybe take a step backwards before they start moving up?

Well, I think if there's a surprise, that could happen. But companies like Applied, to take us specifically, we're not anticipating that inflation is 0 going forward. So our plans will be made in context of studying what inflation we expect and delivering the results that we just talked about comprehending those inflators.

Got it. And then finally, Brice, on use of cash. If you look out the next several years, do you think the top five guys still maintain 70%, 75% right, of the industry? Do you think there is more diversification? And then conversely, how do you see Applied's role, right, in that? Like are there opportunities for further consolidation at all?

Oh, I see. So it's been a tough road for Applied on the M&A side, and we think the environment is not really conducive to a significant M&A at this point. But having said that, it's an important part of the toolkit. Applied has made some acquisitions, tech and talent, different technologies, one in packaging that happened not too long ago. So I think we'll continue to look for opportunities that will advance the core capabilities and also be strategic for customers and solving customer problems. That said, with respect to use of cash, no significant changes from a use of cash perspective. We'll invest in the business. We'll be opportunistic on M&A, and then we'll intend to return a significant amount of cash to shareholders through dividends and buybacks. And right now, we're -- we did $1.8 billion last quarter. We did $1.8 billion in the prior quarter. We certainly think that the stock is below intrinsic value at this point. So we'll continue to operate that way.

What do you see as intrinsic value?

I can't tell you that.

How do you calculate it?

We -- just like everybody else, I think use a football field and use 5 or 6 different methods to calculate. And of course, we have a fantastic DCF since we've got more information than other people. So yes.

Got it. Finally, the services business, AGS, how does that do in case -- let's say, WFE is down 10% next year, right, hypothetically, how does your services business do in that kind of a scenario?

Well, we think this quarter exactly with our semi business being constrained was a good example. We had 5% growth quarter-over-quarter and 15% year-over-year, but 5% growth quarter-over-quarter in our services business. And we think that's partly -- it's really for two reasons. One, 60% of the spares and services part of that business is under contract. So we view that as recurring revenue. And the other part that's transactional like spares, there's high utilization, and so that operated very well in the period. But we do think it will be lower beta, if you will, with respect to whatever is happening in the semi business.

And we have a very high percentage of parts and services in our reported services business. So it's 87% per third-party researchers. We have peers where it's actually 50% WFE and we were able to produce sequential growth when others weren't. So there is a difference.

Got it. Absolutely. Terrific. With that, we are at the end of our time. Thank you very much, Brice.

Thank you very much. Appreciate it. Thank you.

Thanks, everyone.