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

Wanted to switch to sort of the memory outlook. I know your market share is pretty even across the 3 major device types. But with DRAM spot market pricing, seeing accelerating declines over the past month or so, there seems to be increasing consternation in the market about the memory outlook, particularly DRAM. And so I guess maybe share with us your thoughts on DRAM's supply-demand. As we exit '21 into 2022, do you think there's a threat of overcapacity in DRAM? And I'll ask in a second, the same question around NAND.

Yes. So I think we have to disaggregate spot market for our customers and the devices that they make and then what we see from a capacity deployment standpoint because they operate on different time horizons. I know there's this debate around strength of the spot market, whether it's underlying demand or whether underlying demand of these devices in DRAM are being influenced by a supply chain that hasn't fully found its, [ to frame it ], equilibrium. And I'll leave that debate for the customers and those that are much closer to the spot market. As I look the supply-demand balance this year, the DRAM markets, your typically mid- to high teens supply-demand balance. This year, I would say supply lags demand. I would put demand over the top end of that range in the low 20s. I would put supply at the higher end of that range, high teens. But clearly, there's more demand than I think supply -- we'll see supply this year. You'll see a profile of investments. There's no material change to what we see happening in the back part -- in the back half of our calendar year. There's no material change to the customers' plans versus our view a quarter or 2 ago. So those plans and the investment profile, I look at sort of what happened in NAND last year. Back half of calendar 2020 was very strong for NAND, and you saw that strength persist into Q1. DRAM, you're going to see stronger spend in the back part of the year. That's going to bring your market more in balance. And I think you operate more in balance from a supply-demand standpoint as I look at 2022. Again, we'll keep an eye on it. We'll share our perspectives along the way. But that's the setup that I see from a DRAM perspective.

Okay. And any thoughts on the NAND supply-demand environment?

Yes. So I'd say supply-demand balance, mid-30s, plus or minus. I think the market is more balanced than say, DRAM, where you've got a bit of undersupply going on. And I would say the balance is more towards the higher end of that range, so call it upper 30s this year. But balance from a market perspective, I think that leads to the spend profile we talked about earlier in the conversation. And I do like the discipline that I see out of the customer set, exposed to both the NAND and DRAM market. That behavior pattern I find very healthy. And I would expect the supply-demand balance in the NAND market to persist in [ Q 2022 ].

I wanted to turn to sort of the company's outperformance relative to WFE last year and again in '21 year-to-date. And I think you stated or Gary stated that the company's mission is to be the PPACt enablement company and focused on leveraging your broad portfolio of solutions to integrate and co-optimize. Can you spend maybe a couple of minutes talking about some of the areas where you see the company best positioned to gain share? Whether it's advanced foundry/logic or project like technologies moving into the DRAM segment?

Yes. So this is where I really like the positioning of the company. Company outperformed in 2019. We outperformed again in 2020. We're outperforming again this year. If you look at the composition of that spend, let's take 2020 for instance, that was a year where WFE was growing around 18%. Our systems business was up over 26%, NAND was the driver of growth, foundry/logic under-grew the market. If you window to 2021, what you see around setup of the market, we talk about a market that's growing mid-30s, you talk about foundry logic being the strong driver of growth and NAND under-growing the overall market, growing but under-growing the overall market. It's almost the opposite, the mirror image of what happened in 2020. And against that opportunity, you see our systems business outperforming. So I like the balance of our business against a variety of spend environments. I worry a lot less about what device type happens to be spending. We're almost agnostic across those device types. What I do see is strong position at the respective customers. And I look at the node-over-node opportunity, whether it's foundry logic or it's memory, I see us well positioned to increase our opportunity node over node. And we've seen that play out over multiple years. And I don't see any reason why that momentum as I look forward into 2022 would slow down. If I were to think about the markets from sort of a rank ordering standpoint, given what I see from a DRAM standpoint, given what I see from a foundry/logic standpoint and those road maps, I would say maybe the opportunity to make incremental gains is greater in those 2 markets than I currently see from the road map standpoint in NAND. We've got a great position in NAND. We're playing a strong hand. But I see less node-over-node momentum. I still see node-over-node opportunity, but less node-over-node opportunity in that market than the strength we're seeing from a DRAM and foundry/logic standpoint. So I like how we're set up against a variety of spend environments. We've got a broad portfolio. We're well integrated to the customers across a broad spectrum of the industry, and I like the node-over-node momentum across multiple geographies. The other thing I'd say is it's from a packaging standpoint, which is going to be an increasingly important part of the PPACt road map. This is a market position we've been building over quite some time now. We've been investing in this. We saw this inflection coming. Gary has been focusing the company on this opportunity. And now you see it really beginning to take hold and drive meaningful activity at the company. This year, it's going to be an over $800 million business for us, and it's going to be growing over 60%.

The ICAPS portfolio is another position where Applied has a pretty high market share. Can you spend a couple of moments discussing the ICAPS business? And I guess the big question for ICAPS is, since it addresses more mature foundry logic nodes, given how tight those nodes are today, do you think that the demand we're seeing would be "extraordinary"? Or do you think that this demand is sustainable into 2022 and beyond?

Yes. So our current view is, is we're going to have sustainability and follow through at these levels. And let me share with you a bit about what we've been seeing playing out over time that informs that perspective. So if I were to go back, say, a decade in foundry/logic, what I would observe about the industry back then is 80% of the spend was on the leading-edge and 20% of the spend is on what we're calling the ICAPS nodes. And from a definitional standpoint, just to make sure we've got a dividing line on that, today, we define leading-edge as 7, 5 and 3 nanometers and the logic equivalent of that, so the most current logic node. And once 2 nanometers comes online, the leading-edge definition, it's a rolling definition, will be the 5, 3 and 2. Everything else today, 10 nanometers and above, is part of the ICAPS definition, market definition. So if I were to go back in time, 80-20, then went to 70-30, then went to 60-40. Last year, it was balanced between leading-edge in our ICAPS nodes. This year, it's balanced. At our Analyst Day, we would expect between now and 2024 for that balance to continue between those 2 markets. What we see happening, and it's why we reorganized the company about 3 years ago, we saw a market trend where innovation was coming back onto a broad mature node profile. And we call that market segment ICAPS for a reason. It's not just a collection of letters. IoT, communications, auto, power, sensors, what we observed in those markets, with those specific market focuses, our customers were innovating again. And they're innovating across a very broad node profile. And so as part of this organization, we've got very talented engineers, who are very focused on those road maps, and we're working alongside of our customers, bringing those new technologies to market that serve a very specific market need and that innovation engine is really getting geared up on that broad node profile. So you've had a broadening of the demand profile and you've got a broadening of the innovation between leading-edge and trailing node. And at the core of that renaissance of innovation, is pushing intelligence to the edge. And Gary talks about machine-generated data producing over 99% of the data in the world by 2025. A lot of that is going to be driven by intelligence at the edge. When you generate machine-to-machine data, it's going to create a virtuous cycle of demand for compute power and storage in the cloud. You're going to turn that data into actionable intelligence, actionable insights. Companies are then going to tune how they serve their customers, how they compete with their peers, how they drive efficiency into their operations and produce more value for shareholders. When you get better outcomes, it's going to drive more intelligence at the edge. It's going to drive more demand for compute power and you get a bit of a virtuous cycle built up around this explosion of data that's generated machine to machine. So there's a variety of factors that come into play that we're taking a point of view that we're going to see follow through in sustainability at these levels, and it's going to grow in line with the leading-edge. And I think if we look at comments coming out from other participants in the market, I think you see very similar messaging coming out from multiple places and participants within the market that sort of mirrors our perspective on this.

One of the segments where Applied has also grown faster than market for the past couple of years at least, is your process diagnostics and control business, I think on track to grow over 60% this year. I know it's a split between sort of your e-beam portfolio and optical inspection. And I guess maybe my first question is, are both segments within PDC optical and e-beam growing at similar rates? Or is the growth really being driven by one area or the other?

Yes, sure. So here's how I'd shape the performance of that business. So the business has had a lot of momentum over the last couple of years. And so we like how we're positioned against the opportunity. Our PDC segment is going to grow greater than 60% this year. So a very, very strong grower. As I look at the process control market, I wouldn't say every segment of the market is going to be a robust grower this year. But as I take a look at our business and how we're positioned and that greater than 60% growth for that segment of our business, I think the strong driver of growth is going to be our e-beam business, and its e-beam metrology. We've got some of the best resolution. We got the best resolution in the industry, and it's super important to our customers and their PPACt road maps, dialing in the road map, getting their integration scheme crystallized as quickly as possible and getting new technology node online. It's also one of the capabilities that we're leveraging that's driving some of the momentum we see from a DRAM standpoint in the current environment. So it's a great business for us. It's driving industry-leading technology and capabilities to the market. Our optical pattern wafer inspection, we've got great traction with our new product in foundry logic space. It's around EUV applications. And now we're broadening that outside of foundry/logic into memory. And I would say that the momentum we see around that business since introducing that product, it's the fastest ramping optical tool that we've had in our history. So we feel good about the differentiation and the way we're ramping it. But if I look at it relative to the overall segment where e-beam is growing above that 60% rate, our optical business will grow under that segment average, but still a nice business showing some momentum for us as well. So that's how I see that market shaping up and the relative momentum of the technologies that exist within that portfolio for us.

You mentioned the new Enlight optical inspection tool and the growing number of engagements. Can you maybe quantify sort of what -- how that number of engagements has grown over the past year or so?

Yes. So I would say that we see strong traction from a foundry/logic standpoint. That was the initial insertion place for that technology, and it's primarily around EUV applications and high-volume manufacturing. And so we feel good about that product positioning. And now it's a process of diversifying that initial insertion point into the adjacent markets in memory as well, and we're in the process of broadening that footprint out. And so we feel good again about that momentum and the way that product is performing against the opportunity.

Good. And then just looking at your Actionable Insight Accelerator platform, can you talk about the capabilities leveraging in situ and in-line metrology e-beam and how that's helping you accelerate time to market for your customers?

Yes. So this is one of the more exciting parts of taking the capabilities that exist within the company and bringing them together in an integrated, more holistic way to address the customers' highest-value problems. When we talk about the PPACt road map, time is a super important part of bringing your next technology node online. I think the industry has a renewed appreciation of how important that is, given the robustness we see in the end markets get that technology node online and start differentiating yourself relative to the competition in those markets. And so the AIx, Actionable Insight Accelerator, is a way of bringing a diverse set of capabilities from across the company, whether it's sensor technology and our chambers, whether it's metrology, and that's both in situ and in line and it's getting algorithms working on the problem and getting it done in a big data with high-speed compute environment, you're synthesizing this perspective of what's happening in a multiple step flow as well as what happens inside a chamber. And it's combining that with a digital twin capability, so that you can synthetically model how different types of setups on an integration flow will influence on wafer performance. And what we've seen is it dramatically speeds up dialing in of process recipes and also the process windows widening them. Gary will talk about speeding up the integration scheme, the process recipe by 50% and increasing the -- or actually making it 2x faster, sorry, 2x faster to dial in the recipe and making the process windows 30% larger, which yields to more stability in high-volume manufacturing but also more yield. So it's a very important value proposition for the customer and bringing those capabilities together in an integrated way, we're seeing strong traction from the customer base to pull this capability into their operations. And we really like the momentum around driving the road maps even faster in this industry.

Last question for me and then I've got one from the audience. Dan, if I look at the company's guidance for the second half of the calendar year, especially for SSG, it looks like if you run rate that, you're already sort of at the base case model you gave us for the 2024 target. Does this sort of lead you to sort of think that the business and perhaps the overall industry could be tracking towards the high case that you had set forth at Analyst Day?

Yes, absolutely. And what we said at the time we put the model out base case with an upside scenario. If I were to put a probability distribution around those scenarios, I said they would skew to the upside. And so what we're seeing in the current environment only reinforces that perspective. But as I look at it from a market sizing standpoint, but also the performance of the company at each level underneath that from a target model standpoint, I think you can see the progress we're making, top to bottom, towards those longer-term targets. And at the time, we also said that we see an opportunity to potentially do even better than those targets that were put out there. So we feel good about where we sit relative to those targets, and we do see that bias skewing to the upside.

Great. And a question from the audience is sort of, I think, a big picture for the current capacity-constrained environment. Do you see specific catalysts that will likely happen for supply/demand such as Intel moving into foundry segment and shipping leading-edge chips to help alleviate some of those constraints?

Yes. So no comments specifically on any -- what any one customer is going to be doing. But I do think you see an entire focus of an industry coming to bear to address those potential issues. But what I would say is, is longer term, what I see is an industry that's taking a different approach to how they think about their business, given the strength of these secular trends and whether it's companies for the first time coming out with multiple year CapEx expectations, whether it's an alignment around how the power performance road map is going to be driven going forward. Whether it's us, whether it's our customers, we're all talking about the same 5 vectors that are going to increasingly define the power performance road map, whether it's a broadening of demand across a greater spectrum of node profiles with innovation across the entire spectrum, that diversification of innovation, I think, serves the industry well. And so I think there's a number of factors that are coming together that not only address the near-term set of environmentals in the industry, but also begin to speak to a larger opportunity as we go forward. One of the things that came up in one of the earlier conversations, is the number of 300-millimeter factories that are in the early stages of being built. We're in the early stages of accepting equipment. Today, that number is 57 300-millimeter projects around the globe spanning foundry, logic, NAND, DRAM. And if I were to go back a short period of time, say, to 2018, that number was half that. I think there were 29 projects at the time around the globe. It gives you a sense of the view the industry is taking towards these strong secular growth trends the infrastructure that's being put in place to absorb that manufacturing capacity as the end market demand materializes. And so I like a number of aspects of the way the industry is thinking and acting around what I see are strong secular growth trends.

Well, that's great. I think we've reached the end of the time. So Dan, Mike, thank you very much for joining us at the Needham SemiCap and EDA Conference. Really appreciate your time this afternoon. So thank you, and look forward to seeing you again soon.

Yes. Thanks, everyone. Thanks again, Quinn, for the opportunity.

Thank you.

Bye-bye.