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

Hello, everyone. Thanks for joining us back for the session. I'm really delighted to have the management team from Applied Materials joining us this afternoon, Dan Durn, the Chief Financial Officer; and Mike Sullivan, the Head of Investor Relations. I'm really delighted to have them both. And Dan, I thought maybe just to kind of bring everyone on the same page, give us a state of the union. What are the top 2 or 3 things that you're finding the most interesting and exciting right now? And what are the few headwinds that you're keeping your eyes on?

Sure. Thanks for that. We appreciate the opportunity to participate. So thank you for inviting us. As we take a look at the industry, I can't think of a time that this industry was better set up for the opportunity and the road ahead. We've got demand drivers that are diversifying in a very significant way as part of the data economy cut in -- cutting in. And these strong growth drivers, they're very strong secular growth drivers, providing a nice tailwind behind the industry. And we're just in the very early innings of that demand for compute power playing out. And it's happening right when traditional Moore's law to be scaling is beginning to hit a wall. And so what we're seeing the industry do is turn to a new playbook to drive the power performance road map, and the time to market, we call it PPACt road map, turning to a new playbook that has multiple vectors of innovation. Each of those vectors, whether it's new materials, new types of structures, new types of architectures, use-case-specific architectures, new ways to connect chips together to unlock power performance and new ways to shrink. Each of those 5 elements of the new playbook are fundamentally enabled by materials engineering. So we're incredibly well-positioned against this opportunity and where the industry is going. It's led to strong outperformance in '19. We did it again in '20. We're set up well in 2021. We look forward into 2022 and beyond. We don't see any reason why that momentum would slow down. And as we look out to the 2024 time frame, where we put forward a new long-term model, we're really well set up to deliver and exceed those targets that were put out. So we feel good about where the industry is going and certainly our position against that opportunity.

Got it. Thank you, Dan, for that introduction. Before we go to some questions, there have been some media reports about perhaps some rising COVID cases in Asia, right, especially Taiwan and Iran. Just wanted to get a quick update from you if you have seen any disruptions to your business in Asia so far.

Yes. The way I would characterize it is there's no material disruptions to-date. And what we're experiencing is very similar to what we've been experiencing for the past 15 months since the start of the pandemic. You'll find hotspots in different regions. It'll create pockets of nonlinearity from a supply chain standpoint. And it's just incumbent upon us as an executive leadership team of the company to try to stay as aggressive as possible, manage the things that we can manage and sweep inventory when we can, put it inside our warehouses, increase staffing levels in our factories that we can continue to drive output as we experience those points of nonlinearity. And the company has done a great job over the last 15 months. And so it'll just be more of the same, continuing to manage in as aggressive a way as possible these situations when they arise. So a lot of the same thing we've been dealing with for quite some time.

Understood. Right. I hope the team stays all good there. Dan, anything else that you are seeing from a supply disruption perspective? And where I'm really going to is, over the last several quarters, right, we have seen very strong expansion, right, especially in your gross margins, despite all the supply challenges. So I'm curious, are there still some friction, some supply disruptions that you might be seeing that, as we get on the other side, hopefully soon, that you might be able to recover some of those one-off costs? Or do you think that those kind of costs are just part of business going forward?

Yes. So I think these costs will be with us for the foreseeable future. Certainly, the freight and logistics costs will be. The way I describe it is, is we had a strong point of view of where we thought this industry was going. We had a lot of convection around it. And you can see the uptick from an investment standpoint 3, 4 quarters ago. Made investments in terms of staffing within our factories but also expansion of our capabilities and capacity. That's served us well. We'll see more of that throughout this year as well because, again, we just have a strong point of view of where things are going. It's incumbent upon us to grow in a disciplined way but make sure we're putting those capabilities in place in advance of that demand materializing and continue to invest with our suppliers and supply chain to make sure they grow alongside with us. And I think as I reflect back on the last 15 months and reflect on comments from customers, I think we get high marks for managing the situation better than most, better than others. And so we feel good about the capabilities we've put in place. We feel good about the trajectory we're on. We'll continue to manage it the same way we have. But any time an industry will go through a positive [ area ] like it's going through right now, it'll just create those points of inefficiencies, nonlinearities, just important to stay out in front of it, manage it as aggressively as we can. I kind of call it playing offense. I think we've done a good job playing offense. We're on our front foot. We're leaning into these problems. We definitely don't want to play defense, and we definitely don't want to be on our back foot. So we've got a nice positive bias to growth and where this industry is going. We're going to stay out front.

Got it. Makes sense. Dan, if I look historically, right, the history of the semicap industry has been that we have had 2 very strong growth years, then kind of a 20%, 25% pullback. Then we kind of stay consolidate at those levels, then we have another step function up. Now this time, we're already in the second very strong year. So what gives you the confidence to say that, look, next year could be another strong year for the industry? Because historically, that has not happened. So at the risk of saying it's different this time, why do you think it's going to be different this time?

Yes. So what I get back to is, as I think about the secular trends that are shaping the industry, strong point of view that the semiconductor industry is going to go to $1 trillion. It's going to happen over the next decade. And the industry is going to need the manufacturing capacity in place. Things like the data economy, we've seen an acceleration of those trends as part of the response to the pandemic. And we're really encouraged every day, every quarter, every year that passes, things are continuing to line up around our thesis around the data economy. But I'd also say from a data-generation standpoint, 2018, first year that machines generated more data than humans. In a very short period of time, by 2025, 99% of the data generated on the planet is going to be by machines. And so for the first time in the history of our industry, you're going to see a decoupling of growth rates between human population and consumer behavior and our industry. And so all of that thesis is playing out. We're really encouraged by it. Then you talk with customers and you get more insight into multiyear capacity plans, the partnership with customers as we pivot to this new playbook to drive the power performance road map. That collaboration with customers is greater today than it's ever been. I reflect on the infrastructure that's being put in place by our customers. We're tracking over 50 new 300-millimeter projects around the globe. It can take around almost $300 billion of wafer fab equipment. And they want to stay out in front of the demand that they see coming as part of this -- the data economy. And when you transition from consumer-oriented devices as being the primary driver of semiconductors to investments made by companies because it's the basis of how they compete in their industries. As those industries transform digitally, and every industry is going to go through this, companies are increasingly faced with a choice. Do you invest, survive and thrive? Or do you choose not to? And not only is it nondiscretionary, I think it's existential for the customer's customer. Those that don't make these investments to change the basis of competition, how they serve their customers, how they drive their road maps, they're going to cease to exist. And so we're entering an era where you're tapping -- increasingly tapping into the global economy as your economic pool. When we net all of this out and synthesize this information, we've got a strong point of view that we've got an upward trajectory as we look into 2022. And when I reflect on our business, we think this year, high $70 billion WFE, that gets you to a high 20% growth rate. In the first half of the calendar year, actual results for calendar Q1, our guided results for calendar Q2, you look at our Semi Systems business, and that's up 50% year-over-year. So strong momentum from the company, performing well against that opportunity. And as I look at the company in our reporting segments, Semi Systems, our services business and our display business, we see all 3 of those reporting segments up second half over first half, and I look at all 3 of those reporting segments as being up 2022 over 2021. So we think we're really well set up as a company to outperform against this opportunity base.

Got it. Makes sense. Dan, what is the lead time for Applied Materials right now? And I ask that because, every month that goes on, we see yet another number for WFE, right? We started the year close to $70 billion. Now I think they're already kind of knocking the doors of something in the high 70s to $80 billion like you alluded to. Is there some limit to how much WFE can be, like, just given the lead times that it takes to deliver equipment? So maybe that suggests, to your point, that maybe the cycle is perhaps -- could be stronger, right, for longer. This time is just that there's actually limitations on how much capacity can be added in a given year.

Yes. So, hard for me to speak for others in the industry, so I'll limit the comments to Applied. So we talk about investing for 3, 4 quarters now to get out in front of this. And we've seen no material change to the lead time of our portfolio. We've got a broad portfolio. The best way to think about it is kind of 6 months plus/minus, depending on which product line you're talking about, some will be on the short end of that time line. Some will be on a little bit longer side of that time line. But as a broad general category across the broad portfolio, it's 6 months plus/minus. I think we've responded better than others. We're going to continue to make investments to make sure that we can continue to grow alongside with our customers. But you see where we are in the year from a time line standpoint. I think that gets you a good sense of sort of how things play out for the rest of the calendar year. But again, I think we don't see a material change one way or the other, given the rate of investments that we've made around our business.

Got it. The last question, Dan, just kind of broad WFE related. Is there a way to think about WFE and WFE growth in terms of intensity, right, just the ratio of fab equipment spending to semiconductor sales? And we have seen that ratio between 9% and 16%. And despite this very high spending levels, we are still kind of knocking on the doors of 15%, right? I don't think it's exploded any higher. Where do you think that number goes over time? And I ask that because your 2024 model, right, which had $85 billion to $100 billion, I mean, we could be at $85 billion start of next year. We don't have to wait till 2024 to get there. So is it accurate to think about the industry in terms of WFE intensity? What do you expect that intensity to do over time?

Yes. So like most metrics, it's a good metric to track, but it's certainly not the only metric that, as you know, that will characterize the industry and how it performs over time. But you're touching on a good point. I'm going to make 2 comments on it. I'll first start with our approach to the investor meeting and where we set expectations, and then I'll double-click on our view institutionally on where we think this goes over time. So my personal view from an investor meeting standpoint, I think it's prudent where you set expectations. You do it in a way that allows the company a high probability to hit and exceed targets. And I think you see that approach playing out with respect to the investor meeting. We've got a high degree of confidence that, as you think about the probability distribution between the base case, upside case, downside case, you're going to see a skewing of that probability of outcomes towards the higher end of the range. That's what we said at the investor meeting, but we also want to set a set of objectives that gives the company a good probability of meeting and exceeding. So it's just a philosophy around how to be prudent and where we set expectations but certainly doesn't lessen our confidence of where things are going and the company-specific momentum we have around that opportunity. Our strong point of view on this is, is you're going to see upward bias in capital intensity over time. We talk about the new playbook playing out, the importance of time to market and the handoff of demand from consumer devices to the largest companies exposed to the largest industry. That complexity, that return on the incremental investment by our customer's customers, it goes to the foundation of how they compete in their industries and the return on that investment is significant. I do think you see profit pools shift over time to the electronics industry and to the semiconductor industry as this plays out. The economics around tapping into the global economy versus a consumer-oriented device are going to be orders of magnitude greater than what we've seen historically for an industry. So that's going to accrue benefits to the industry. The other trend that I think leads to a nice upward bias of capital intensity over time, governments, in a very logical way, are making the connection between the semiconductor industry, economic growth, productivity gains and national security. At the core of those 3 pillars is the semiconductor industry. You can't have a modern economy without a secure supply of semiconductors. So what is today a very capital-efficient way to deploy capacity globally, large factors -- large factories clustered geographically to take advantage of scale benefits of an ecosystem that's been built. The model going forward, you're going to increasingly see a shift. It's going to be to more regionally distributed capacity. So you're going to end up with a larger number of smaller-scale facilities. You won't end up with more capacity. The capacity that comes online is just going to be distributed geographically. Those larger number of smaller facilities is going to lead to an upward bias in WFE as that trend plays out. That's going to benefit us well. Governments will step in and make our customers economically neutral between how they deploy capacity today and that more globally distributed, geographically distributed model. That's where governments are going to play a role in making our customers economically neutral, but it is going to have the net effect of an upward bias of WFE intensity over time. So I think it's a good metric to track. And I'm not smart enough today to say what number it asymptotically approaches over time, but I do see enough headroom in the way we're framing this to get to high teens. Does it get higher than that? I don't know. I'm not smart enough to know, but I do see enough headroom as we sit here today and look out into the future to go from, call it, 14% today to high teens over time.

Got it. And Dan, just you mentioned this additional spending force of government wanting, right, to have this more geographically dispersed capacity. When do we actually start to see those benefits? And is it possible that, while that is going on, maybe intensity goes towards that high teens or even 20% number but then perhaps goes to whatever is the long term, but at least for the next handful of years that we do stay in this elevated state. But when do you think we start to actually see that spending happen? Because so far, it's been more talk than actual spend dollars as far as I can see.

That's right. So here's our view on it. As we absorb announcements, whether it's countries in Asia, Europe, the U.S., there's going to be more announcements on this going forward. Governments typically don't act quickly. So our view is, is that the first probably insertion point you see of government activity is probably 2023 and beyond, maybe it runs for a handful of years. We're not factoring any of this into our expectations around 2022. Right now, the model, as we've put forward and the expectations where we're setting them, are all secular trends playing out. That's where we've set the expectations. Any influence of this earlier than 2023 would be upside to where we're setting the expectation. And I do have a strong point of view that the role governments will play will not be to overbuild capacity. I just don't think that's where they're going to go. But I do think they will insert themselves into the process to take our customers and make them economically neutral on where they build capacity. So the net impact of that is, is I don't think we're going to see a spike in WFE intensity, but I think you will see a steady upward bias over time, certainly going to serve us well. And as those trends play out, not only does it benefit us from a WFE intensity standpoint, but the service entitlement around that capacity that's built outside of home geographies and home ecosystems that service entitlement on the capacity is higher than we would see clustered into their home ecosystem. So I see the benefit of this across a couple of dimensions of our business. And I also take comfort that governments are waking up to the strategic importance of semiconductors in this industry to really trends that shape how people live their lives and economic security and national security. It's a special industry, and governments are telling us that.

Got it. Makes sense. Now Dan, what I also find kind of different about this particular cycle has just been the strength that we have seen from foundry and logic spending because prior cycles seem to be a lot more exposed to just memory spending volatility. This time, we have kind of memory discipline. And on top of it, we have very strong foundry/logic spending, but we have already had 2 very strong years, at least a very strong foundry spending. Hopefully, there is logic spending to come after that. How is your visibility with your foundry/logic customers? Are they signaling to you a good spending trend over the next 1 or 2 years? What are you hearing from them overall?

Yes. So we do see foundry/logic strength continuing as we look forward. We fully expect '22 and '23 to be strong investment years as customers continue to bring on compute capacity. And then when we think about our opportunity node over node, we saw a nice bump, 5 nanometers over 7 nanometers, companies performing well in this environment. As I compare the transition to 3 nanometers to the 1 to 5, there's a significant increase in our opportunity on a relative basis. So we like the opportunity as we look forward into 3 nanometers, and you're going to see a similar dynamic again in 2 nanometers for our business. So we like the momentum we see on a node-over-node basis and our opportunity. From an overall market standpoint, as you think about this dynamic I shared on data and data generation, and you see this really significant ramp of machine-generated data, data for data's sake is interesting, but it's not very relevant. So then the question is, is how do you turn that data into insights that companies can take action on to drive improvements in their business, how they compete, how they serve their customers, how they drive their road maps, how they drive efficiency into their operations and supply chain. It's those insights that matter. And the only way to get those insights is through compute power. So you're going to see strength in this area of the market as the data economy continues to build up momentum. The other thing I'd say that's driving strength is not only do we see it at the leading edge, but we see it in trailing node geometries, and we're seeing strong growth across that node continuum. The breadth of that demand, the breadth of that strength, I think, served this industry well. And as you think about the Internet of Things, pushing Intelligence to the edge, everything is going to be chip-enabled over time. That intelligence being pushed to the edge is giving you a diversity of demand that serves the industry well. That's not to say every year will be higher than the last, but I do think there is an upward-sloping trend line. And any cycle overlay that we see over that trend line, you're going to see the amplitude of that begin to compress on a percentage basis, and you're going to lead higher highs and higher lows. And foundry/logic will be right alongside as this industry begins to mature, stretches to $100 billion, $120 billion, $140 billion over time. It's a good dynamic and a good setup, and we see the strength continuing based on a number of factors.

Got it. And then, Dan, moving to the memory side of the business. What is the right way to think about DRAM versus NAND growth half on half? And I ask that because it seems the DRAM industry has always been extremely disciplined, right, and you see that. And one can kind of track that, right, it's computing. We can look at cloud CapEx. We can look at 5G, right, smartphones. We can look at PCs, et cetera. With NAND, there's always a substitution effect, right? So sometimes, that elasticity can actually create a lot more volatility on the NAND side. But if you set that preamble aside, when you look at your anticipation of growing second half better than the first half, what does that contemplate in terms of your DRAM growing first -- second half versus first half? And the same thing for NAND?

Sure. So let me take a step back, draw some contours around the overall industry and then provide insight into what we see in NAND and DRAM, and that gives you good context for how we're seeing things. So we said high 70s for the overall industry, which gets you high 20% growth rate. Against that market growth, foundry/logic will significantly outgrow the overall market. Then DRAM, which will be around the industry average plus/minus, maybe a little bit more. And then NAND will grow, but it will significantly undergrow the overall market growth. So you see that spread across end markets relative to market growth. Within the memory market, we've got a strong point of view that DRAM is going to be a second half weighted market. As we look at 2020, I would say that market coming out of 2020 from a supply-demand balance was a bit undersupplied relative to where the true market demand was. NAND exited last year far more in balance as an overall market. You saw a strong investment cycle in NAND last year. That strength continued into calendar Q1. Our NAND business in calendar Q1 was up almost 110% from a year-over-year basis. A near competitor was up over 90%. So calendar Q1 was a strong quarter. 3 months ago, we would have said NAND was a first-half-weighted market. In the last 3 months, we've seen the overall market strengthen. As I look at the delta between low 70s and high 70s, I would say most of that increment is in foundry/logic followed by DRAM and the smallest impact in the last 3 months would be with NAND, but you see that uptick in NAND as well. So given where we sit today, the question is, what is the shape and profile of the NAND market? We see year-over-year growth. You've got a very strong calendar Q1, and we see the year-over-year growth rate moderating significantly throughout the year. And the question between first half and second half where 3 months ago, we said first half weighted, we're putting a question mark on it today. I think it's too early to tell whether it's first half weighted, balanced between first half and second half or second half weighted. We want to see more data points, both from a market standpoint, competitor standpoint. We've got a point of view of what our business is going to do. But I think we're shaping it now with a question mark, and we just want to make -- see more time elapse and more data points before we refine the perspective that we see. But DRAM, clearly second half weighted, foundry/logic, second half weighted.

Got it. And Dan, just to kind of continue that forward, if you look in 2022, I think you mentioned the anticipation, right? The expectation is that overall industry spending grows, as you mentioned, you expect foundry/logic to stay strong next year. Any color around how do you expect DRAM versus NAND to do year-on-year? Because you would be exiting the year now at a very high DRAM position and a very low NAND position. So does that inform us in any way on what DRAM and NAND will do next year on a year-on-year basis?

Yes. I think it's premature to be point specific, certainly by device type for next year. I do think what we will see is strength in both markets. Growth rate, as you know, is going to be a function of what actually gets done in calendar Q4 versus pushing to calendar Q1, and that gives you a better sense of how that transition will look from a year-over-year growth standpoint. As I take a step back from the near-term environment and you think about the shape of the market in our investor meeting, what we shared was, today, we have over 55% foundry/logic, under 45% memory. It's been the sort of relationship in this industry for a long period of time, last 2 decades, the last decade. And we see that continuing through 2024, foundry/logic greater than 55%; memory, less than 45%. Within foundry/logic, we see trailing node and leading edge roughly growing in line, where leading edge is 65% plus/minus of the market; trailing node, 35%. Within the memory, if I were to unpack memory, I would say you've got balanced growth, both are going to grow nicely, but DRAM probably noses out NAND. There's not a big separation. So I would expect these markets over time to roughly profile in line with each other and maybe I give a bit of a nudge to the DRAM market versus the NAND market. So when you take yourself out of the second half of this year or 2022 and think about it a little bit longer term, one might be better in 1 year than the other. But over time, I think you would see those markets grow in very similar ways. I don't know if that helps, but that's sort of what we see here over the near to mid-term.

Got it. Very helpful, Dan. The next thing, a little bit of a technology question. We saw logic transistors move to 3D, right, structures very early on. Then we saw NAND, right, move to 3D NAND, right, and vertical structures a few years ago, and that led to a big jump in intensity in NAND and the spending on NAND. When do you think DRAM makes that transition to these new kinds of structures? Because people always talk about Moore's laws is "dead," but I don't think that quite pays attention to how the industry is evolving to different kind of structure. So when do you expect DRAM to take the move towards a 3D structure? And what role would Applied play in that transition?

Sure. So rather than give a specific timing because, as you know we spend a lot of time with customers. I'll let them talk about their roadmaps and when that inflection happens. But it'll be a couple of nodes from now, but I'll let them talk about that. We're going to play a very large role in driving that inflection into the industry. And you rightfully point out, the transition to 3D structures in the NAND market was very good for our industry. You'll see a similar impact as the DRAM market goes to vertical structures over time. So as you think about the shrinking on a planar 2D basis, at some point, the shrinks run out of steam. You get less bit growth out of it. The feature sizes become so small, you can't hold the charge or you run out of electrons in the memory cell, and the industry is going to have to innovate. You're kind of at that place from a DRAM standpoint in a few nodes. And what's going to happen is, is you're going to see vertical structures get created, where the catalyst to vertical structures will be similar to the NAND market. The way it's going to be different in how it's implemented. You're going to see a lot of the things that we do historically, our leadership position is playing an important role. You're going to see a lot of conductor materials. You're going to see conductor etch. You're going to see selective removal. You're going to see epi play an important role. And e-beam technologies measure how those structures are coming together. A lot of the things we do that are traditional sources of strength for the company are going to play out and play out nicely for us as that inflection hits. So we're going to play a very large enabling role when it does. It's going to be a nice impact, an upward bias from a complexity standpoint and take our industries larger than they are today, and we feel good about that. You could say something very similar around gate-all-around and foundry/logic business. And so as we think about those major inflections by device type on the horizon, I think we're incredibly well-positioned to enable those and drive those road maps for our customers. So we feel good about it.

Great. And Dan, just in the last few minutes that we have available, just talk to us about your display business, what the outlook is for the second half and then next year. Because, again, right, the demand signals seem to be a little mixed coming from Asia, but the perception is that your business has possibly troughed and could potentially grow from here. But just give us what your kind of realtime thinking is on the display business on the demand drivers.

Yes, sure. So from our business perspective, what we said is we expect this year, fiscal 2021, to look and feel a lot like fiscal 2020 in terms of overall size and where that business comes from, whether it's TV or handsets. I think when you take a look through that lens, you can see that the second half is going to be up over the first half. So we expect the second half of our fiscal year to be up over the first half in display. And then as we look forward into 2022, we would expect it to be up year-over-year. So what shapes that perspective is, is we think fundamentals in the market are strengthening with our channel prices, average area size of screens continues to increase. You're seeing increasing penetration in the handsets. And then as you think about the inflections on the horizon, things like OLED coming to larger-format displays, whether it's IT or the TV market or foldable displays in the handset market becoming more commercially pervasive and widespread commercial adoption, the fundamentals in the market are strengthening, and we think we take a step up from an investment standpoint. We haven't quantified that yet. It won't be the 2018 levels. The initial stage will be more modest as we go from bouncing along the bottom to something that's a little bit more attractive from -- a little bit more attractive in the investment cycle. And so we're expecting our business on a year-over-year basis to be up. So up in the second half and then up again on a year-over-year basis into 2022 as those industry fundamentals continue to strengthen.

Excellent. Thank you so much, Dan and Mike. Really appreciate you taking the time and sharing your insights with us. And thanks to the audience for joining, but we can conclude the call here. Thank you so much, again.

Thanks again, Vivek.

Thanks, Mike.