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

I appreciate that. Good morning. Thank you all for attending Day 3 of the tech conference. I'm Toshiya Hari. I cover the U.S. semiconductor and semi cap equipment space at Goldman Sachs. Very honored and very excited to have Gary Dickerson, President and Chief Executive Officer; and Dan Durn, Senior Vice President and Chief Financial Officer, from Applied Materials. Gary, Dan, thank you very much.

Well, thank you for having us.

After a very busy night. I can only imagine. So thank you. I've got a bunch of things that I'd like to ask you guys on, but before going into specific questions. Gary, I was hoping you could give a general update on the business.

Sure.

What you're seeing in the core markets, what the key initiatives are for Applied in 2020. And as a leader of the business, how are you spending your time these days?

I spend a lot of my time with R&D leaders and CEOs really across the entire ecosystem. I think from a big picture standpoint, what is happening in our industry, we communicated on the call yesterday. The business is larger than it's ever been in the past. Basically, you have IoT, AI, big data, layering on top of mobile social media and also PCs. So you see the semiconductor equipment business, again, much larger in the range in the last 5 years of 10.5% to 12% semiconductor equipment versus semiconductor revenue and also much less volatile. So again, it's really driven by technology transforming every industry: education, health care, transportation. You see retail. You see all of the really data-centric companies. Market caps are rising the fastest. And we're really in the early innings of this big wave for the entire industry. If you look at -- 2018 was the first year machines generated more data than people. By 2025, that's going to be 10:1. And people talk about 1 trillion connected devices by 2030. And really, at the foundation of all of that is the semiconductor industry as the infrastructure for the data economy. So that's really driving our business. At the same time, you really have the end of classic Moore's Law: the 2D geometric shrinks. So the new playbook that we've been communicating and I think really if you listen to chip companies, you listen to many people throughout the ecosystem, there's pretty good alignment that the old way doesn't work, that's not going to enable the scale and the data economy for the future. So it's really about new architectures, new structures, new materials, new ways to connect chips together and new ways to shrink. So for Applied Materials, the opportunities have never been better because a lot of that's about materials innovation or innovating with integrated material solutions for new structures. So for me, every week, I'm with R&D leaders for our customers. In fact, tonight, I've got 1 of our biggest customers that I'll be with. And again, it's really through the entire ecosystem where we have those connections. So that's really what's driving our business.

Great. Thank you very much for that. Dan, on last night's call, you spoke to your expectations around 2020 WFE up 10% to 15%. I think collectively, we went into the call expecting you to sound good on memory. What was kind of surprising was the optimism around logic and foundry, some of the trailing edge of process nodes. Can you speak to your expectations around both memory and logic and foundry for the full crowd?

Sure. Thanks, Toshi. I think it's important when we put context around what WFE is going to do in 2020. We firmly established what the baseline upon which that growth will be built. And so we do know that WFE in 2018 was $56 billion. It's Gartner number. It's validated by a third party. It's a good number. From there, the industry was off about 10% to 12% in 2019. That gets you your baseline upon which 10% to 15% growth is built. And we think it's going to be at the high end of the range, given everything we see. Embedded in that is strong pool and foundry logic that continues throughout the year. And we really like the setup in what we see. Multiple nodes, multiple customers are in the process of going larger. And so we think that's healthy. Every quarter won't be a record. Q1 was a record. Every quarter won't be a record, but we expect continued strong, sustained pool and foundry logic throughout the year. And then we look at the early signs and stages of the memory recovery. And if that takes hold and builds momentum throughout the year, there's an opportunity, maybe our fiscal Q4, fiscal Q1, which would be the back half of the calendar year, looks even better than we currently anticipate. We'll keep an eye on it. We'll go 1 quarter at a time, but we really like the setup. And then when we take one more step back and we think about that $6 billion, $7 billion incremental transition into 2020, that's the aggregate size that the growth represents. When we look at that, probably 40% of it is incremental foundry strength. 60% of it is incremental memory strength. And in memory, you see balance across device types. So I like the fact that it's not single-threaded through one device type, foundry, logic, NAND, DRAM. It's broader-based than that and I really like that setup. Not only for the rest of this year, but particularly when we look into 2021, we really like that view as well.

Toshi, one thing I'd also add, when you look at this AI/big data era and trillions of -- $10 trillion being generated as industries are transformed, you really need to drive a significant improvement in performance per watt. People talk about 1,000x improvement in performance per watt. And so one of the things we're also seeing is capital intensity is rising. We see that in terms of equipment as a percentage of revenue over the last 5 years and we think that's going to continue. But that gets back to this new playbook, yes, where I spend my time. It's really how do we enable those new structures. It's absolutely crucial. New ways to connect chips together. We had a record revenue this last year in packaging. This whole trillion connected devices, specialty devices, IoT, communication, auto, power, sensors. Again, that's a different kind of innovation. So really driving that 1,000x improvement in performance per watt, and I deeply believe the future will not look like the past and we're not going to -- for sure, when I meet with all of these leading R&D customers, we're not going to get where we need to go, doing what we've done in the past.

So it feels like collectively as a group, we spend a lot of time thinking about litho and etch and depth. But to your point, there's more than more and...

And when you think about performance per watt -- I was with one big tech company, the R&D leader, and we talked about one of the initiatives we have. We talked about gate leakage, 1,000x improvement in gate leakage. And the direct comment from him was the value of performance is unlimited. So when you're driving many of those types of initiatives on integrated material solutions, you can even trade-off power for performance. So if you want to increase drive current for certain applications, high-performance computing, you can do that. If you're on an edge device or a mobile device and you want to reduce leakage and power, you can also drive that. But really, there's tremendous, tremendous value in the data and building out the edge in the cloud and having those new capabilities are enormously important.

Got it. I'll definitely come back to the tech inflections, but just bringing it back to...

Okay. I'm still excited about that.

No, no, that's awesome. Just bringing it back to some of the shorter-term issues. The coronavirus obviously, evolving situation, difficult to handicap. I think you guys gave a great color on the call last night, $300 million potential hit to revenue in the current quarter across all segments. Gary or Dan, I guess, operationally, what are you guys going through today? What's sort of the nature of conversations between you guys and customers? What are they working on? If you can provide some color as to what's going on in the field, that would be helpful.

So what we see -- let's talk about what we said last night and then break it down to what we see to provide context around what we said. As you point out, we said $300 million derisk to the revenue in the quarter that we've guided to, and you see that split across all 3 of our reporting segments. In rank order, you'll see the biggest hit in our semi systems business, followed by display, followed by service, and service would have been our first $1 billion quarter in the current quarter. And so that gives you a magnitude of the derisk on services. And then you can pretty much get a sense of where the other 2 businesses are relative to that. So that provides color on what we did. When we breakout coronavirus into 2 buckets and the impact, what's going on out in the field, and we talk about customers, and then you talk about supply chain, in customers, we're seeing strong pool. The customers want to ramp. They want to get the capacity in place. They want to fight through the logistical challenges and the travel challenges, and they want to keep the production going because they see strong end market demand. That side, very, very healthy. The other side is very healthy as well. We've got suppliers who understand their position to have ramp plans around getting their workforce back in their factories and back to production. We have daily and weekly calls with the entire supply base and our organization understanding inventory in the channel, understanding ramp of supply, understanding trajectories around where this goes over the next couple of months. And make sure that we have workarounds, plans in place where we see signs of constraints and work very, very closely with that supply base to make sure the strong demand we see from our customers and our customers' customers gets met by the industry. And when we net all of that out, we felt the approach we took in the current quarter was prudent, represents everything we see and is knowable today and looks recoverable. Vast majority of it recoverable in the next 2 quarters. So we feel good. Looks temporary. Looks like it's recoverable within the current fiscal year. So no change to our outlook. And we're going to continue to manage it and drive it. What we've seen in the first 2, 3 days of the workforce returning in China, it's a fluid situation. Too early to tell, but we see some good indicators of return to normalcy in certain areas. I don't want to extrapolate that. We're going to stay close to it and continue to monitor it. It's still too early to tell, but we see a couple of indicators that I'm encouraged by.

Okay, great. In terms of your longer-term WFE expectations, it's a tricky thing to discuss and predict. But I think as a group, we tend to be a little too bullish at the peaks and a little too bearish at the troughs. I wanted to get your kind of updated view on through-cycle growth for WFE. It's clearly a growth market. It is cyclical. It's, again, very difficult for us to predict. But what are your thoughts on...

Yes. I think if you look at the drivers for the industry over time, you had PCs where you're waiting for an operating system upgrade and then you went to mobile social media, which was much more pervasive. We're all carrying around cameras and data centers in our pockets or computers in our pockets. And then if you look at the future, the pervasiveness of technology, you're transforming all of these major industries. In the future, if you're not a tech company, you're going to lose. Again, and you look at the value in data and transforming, again, transportation, education, retail, health care, enabling better outcomes for people. There's so much of a profound impact that we can have with this type of technology. So the pervasiveness is dramatically more broad than it's ever been. And you can see that, again, over the last 5 years, you see the intensity -- capital intensity as a percentage of semiconductor revenue in a pretty tight band between 10.5% and 12%. They average around 11.5% in that time period. So again, it really has dampened out. I mean certainly, if we look at what we're seeing in 2020, we communicated that that's up last year. Dan gave the math that it's over $50 billion the previous year in 2019. 2021, we think a lot of things come together in a very, very strong year for Applied Materials. We think the semiconductor industry is still going to be very strong. Again, as you're driving these -- this big wave of AI big data, generating data -- again, I talked about IoT, communication, auto, power, sensors, that part of the market is bigger than it's ever been. And if you look at some of the fastest-growing segments, sensors for sure fits into that category. So we have a very, very big focus there. That's bigger than it's ever been in the past. You look at generating data, storing data, processing the data to unlock the value connecting data. These are just massive trends. And so when we look at WFE, and Dan could give statistics around the percent of semiconductor revenue, we think capital intensity is still going to be very strong. The value -- again, as I go meet cloud service providers or people in the ecosystem, there is enormous amount of money in transforming these industries. The value of the data is really significant for these companies. The numbers are very, very large. So enabling the power and the performance for all of those companies, whether it's on the edge or the cloud is really, really tremendous. Then, Dan, maybe you could give statistics around capital intensity and the semiconductor revenue.

Sure. And so as we look over time, and you can take this back over probably 4 decades, there's a fairly tight band. You'll get year-to-year volatility, WFE divided by semiconductor revenues, but when you take averaging over periods of time, 3-year moving averages, we've got a method we use in the company that's actually showing some really nice trends as we smooth out that year-to-year volatility. You see definitely a center of gravity of 10% to 12%, as Gary talked about. We're seeing an average of about 11.5% today. And so you ask yourself, where do you think semiconductor revenues are going to go in the future? You look at demand drivers for semiconductors. PCs took the industry to about $25 billion on average. Mobile social media took it to low $30s million on average. The data economy has now taken it to new normal of $50 billion. And I suspect if we're on the stage a couple of years from now, we'll be talking, "Is the new normal $60 billion?" And what you see embedded in this is the trends that are driving and shaping how people live their lives are enabled by semiconductors. And the technology of semiconductors is enabled by companies like Applied Materials. So there's no way for the semiconductor to be -- semiconductor industry to be those key enablers of these trends without growing structurally larger. Today, it's a $450 billion industry. And you don't need significantly higher growth rates than global GDP to get to a $750 billion industry and ultimately $1 trillion industry. And then if you take a look at our industry, normalized and sized to the overall semi industry, because to grow the semi industry, you need manufacturing capacity in place to do it, you can see very quickly our point of view, our industry is going to grow structurally larger as a key enabler of that technology. Last thing I'll say is mobile social media took our industry to the low $30s million. That's entirely a consumer-driven choice. What we see unfolding today, I think, makes that -- is going to make that pale in comparison from a sheer magnitude because what's driving the data economy are the largest companies exposed to the largest industries chip-enabling the physical infrastructure of this world, so that you can get better productivity and better returns on the physical assets of the world. And we're really encouraged by the fact that what this leg of growth is much larger than past legs and nonconsumer discretionary. The setup I don't think has ever looked better for our industry.

Just one more data point. So if you look at data exploding -- again, I said in the 5 years you'll see 10x more data coming from machines versus people. And if you look at how much power it takes to generate and store and process and connect all of that data, if you take -- we've done a lot of modeling that's been echoed by large tech companies. If you take what power is needed for a certain amount of data in different applications, if you just scale what we have today, between now and 2025, you're going to need to generate an additional 10% power in the world. That's -- I mean, that's a really enormous challenge. And so you can't -- you have to innovate. You can't get to where we need to go by scaling what we have today. And a lot of people talk about you need to improve performance per watt by 1,000x versus what we have today. When you think about building out this data economy, the infrastructure, again, that's where it comes back to this new playbook, where you have to innovate with new architectures, new structures, new ways to connect chips together, new materials because again the value is so enormous, but there's also big challenges that we have to overcome as an ecosystem.

And Gary, I guess, related to that, just given the breadth of AMAT's portfolio, you've been really good at identifying key inflections in the past.

Yes.

As you speak to these R&D leaders and your customers, what's next? What excites you when you think about your business over the next 3 to 5 years?

Again, it's really about innovating. As I said in this new playbook, new architectures. Longer term, you have things like in memory computing and other architectures. But again, there's a huge focus on how to enable those new capabilities, new structures, I talked about the gate leakage where we're engineering a different structure for the transistor that you can trade-off, drive current or low power, and we're doing the same thing in interconnect. We have these integrated material solutions where you're combining different technologies in creating materials, and we have the broadest portfolio by far of materials. So creating materials, shaping materials, including selective removal, where we have a leadership position of 1 material versus many others with 1,000x difference in selectivity, modifying materials and analyzing materials, where we also have leadership. So that combination in the portfolio, we're engaged with every single customer on these new integrated material solutions to engineer new structures, implement new materials, whether it's in the leading foundry logic, where we have incredibly deep engagement with all of those different companies. We have real line of sight to what those structures will look like more than anyone because, again, we're involved in creating them and enabling those structures with these new solutions. And memory is the same thing as how the companies are going to scale those memory devices. It's all about materials innovation. And the same thing is true in the specialty nodes, I talked about IoT, sensors and all of those areas. So those are the areas we're really focused on. And again, I have great confidence that we will meet those challenges, but it's not going to be doing what the industry has been doing in the past.

Right. That's great. Shifting gears a little bit. I wanted to talk about China a little bit. Based on your comments, based on our math, I think China or local China was 10%, maybe 15% of WFE last year, maybe towards the low end of that. You talked about $2 billion to $3 billion in incremental dollars from that market or from that customer base in 2020. Can you talk about the growth drivers there? And I guess more importantly, when you think about the region long term, do you see incremental WFE spend from China as being additive to global WFE? Or is that upside, downside elsewhere?

So I think the best way to come at this, let's talk about what we see in 2020, and you laid out some of the facts. And I think that's a good framing. And then let's take a step back and talk about what we see longer term in China. So what we see, you rightfully pointed it out, 10% to 15% probably towards the lower end of that range in 2019. We see $2 billion to $3 billion of incremental growth in domestic China spend. And as we break that down, 1/3 of that incremental spend is 200-millimeter, which is all trailing node foundry logic. Of the 2/3 that's 300-millimeter related. Half of that's trailing node geometry, foundry logic; half is memory. And there's balance between device types embedded in that spend. So take a point of view of where we are in that range, you're probably 1/3, 1/3, 1/3. Trailing node 200-millimeter foundry logic; 1/3, trailing node 300-millimeter foundry logic; 1/3, memory related balanced across the device types. That gives you a good framing. And when we look at that level of spend and normalize it to, say, what are memory factory costs or foundry logic factory costs, a new memory factory today is $7 billion $8 billion. Foundry logic, 100,000 wafer starts is $15 billion to $18 billion. So even though the step-up year-over-year is large on a percentage basis, it's still very small when you think about the capacity statement that sits behind it. And so taking a step back and talking about what we see, we've been talking about a slow steady ecosystem build in China for several years now. And that's what we've seen historically. That's still what we see. We actually think they're being smart and disciplined about investing in the technology road maps, building the ecosystem with modest capacity adds in support of that technology road map. And so I do think the semiconductor industry will operate as a global marketplace where there is true end market demand, and then that demand is going to be met by semiconductor manufacturers. And you will see that equilibrium over time. It will operate as a global marketplace. So there won't be repetitive and additive builds of spend. Building factories is far too expensive for that capacity deployed to be long-term inefficient. It will find its point of efficiency, but capacity adds to date are very modest as the ecosystem is being built in a slow, disciplined, managed way over time.

Got it. Investors are increasingly focused on how the geopolitical landscape could impact your business, the broader semiconductor industry. Gary, as a leader of Applied, and obviously you spent a lot of time in Washington, what are some of the conversations that are going on between the industry and the government today? And does it feel like your argument is being received well?

Yes. I think if we look at our China customers, certainly from that perspective, it feels like business as usual. And you're right. I mean I'm in China a fair amount in discussions at different levels of leadership there and certainly in Washington in January and December. I'm part of the U.S.-China business council. So definitely have a strong engagement there. We are actively engaged in discussions with key decision makers, and I would say those are constructive discussions with the key decision makers. Relative to -- I think, people wonder about export controls or any of those types of things. We can't speculate on any of that. What I would say is that we're continuously engaged. We monitor the situation and then we'll react appropriately to what we see. I don't know if I can say too much more than that. Dan, do you want to add anything?