ASM International NV (ASM) Earnings Call Transcript & Summary

Good afternoon, everyone. Thank you for coming to this session with ASMI. My name is Adi Metuku, and I cover the European semiconductor names here at CS. I'm pleased to have with us today, Benjamin Loh, CEO of ASMI; and Paul Verhagen, CFO. In terms of the format of the session, I have a few prepared questions, which I'll go through, but towards the end I'll open it up to Q&A from the audience. So we will be taking questions towards the end. So without further ado, let's begin.

So firstly, Benjamin, Paul, thank you for making the time to come to this conference. Maybe to start us off, you guys updated 4Q outlook yesterday. Can you talk a bit about what changed and what drove the outlook update.

Yes. Maybe I'll take that one and maybe Benjamin, you can add to this. The biggest change was our interpretation of the U.S. export restrictions to China, which we announced on October 7. When we actually came out of our Q3 results 2 weeks thereafter, we've taken a conservative approach, in particular, with regard to U.S. persons involvement, what does it mean? And where does it begin? And where does it end? We just didn't know many, many questions. So basically, we assume that everywhere where U.S. persons would be involved in the value chain for deliveries to advanced fabs, we take it out. In the subsequent weeks, we had multiple meetings with U.S. government, with the Dutch government, with some lawyers. And now the definition has been narrowed down. It's much more clear. And based on this narrow down definition, and some organizational changes, we can actually ship quite a lot of products without U.S. person involvement, not all of them because we also have products that are developed in the U.S. So one, to support those without U.S. personal involvements like close to impossible. And two, also U.S. technology, you cannot, let's say, transfer to advanced fab. But the good news is that for the product development and all the other product center that we have in Korea and Japan and the Netherlands, we can now basically ship those, service those without U.S. persons involvement.

Got it. And...

And that is the biggest change in the outlook. So the outlook, we increased the revenue outlook. It's also a little bit the other business that is slightly better because of improved supply chain conditions, but the biggest chunk of that is increased sales to China to advanced fabs.

Yes. Got it. And so -- but underlying, there was some improvement as well. Did I -- okay, got it. Now maybe looking out to 2023, a lot of people are wondering where this WFE end up. So maybe if you could talk a bit about what your views are on WFE. And also if you could comment a bit on the different end markets around logic, foundry, memory and so on, that would be great.

So I think the whole industry is still very much trying to converge towards a view of what WFE is going to look like for next year. But if you look at the individual segments, I think you're going to see basically the logic and foundry segment still being fairly resilient. There are some estimates that say it might go down by a few percentage points, but that's about it. Memory is the one that I think has the biggest impact in terms of slowing down. However, if you look at what has been kind of forecasted or being projected, the range is too big in the sense that it can range from minus 10 to minus 50, which, to some extent, kind of says that the industry is still trying to figure that out. And of course, if you look at the trailing edge, what we terminology called the power/analog, that part is actually still strong. And I will not be surprise that we see some growth over there. So if you add that all together, I think because of the drop in Memory, next year WFE will be down. but by how much, I think that's still left to be seen. It's still a little bit too early to really give a number.

Okay. Got it. And maybe can you also talk a bit about how you see your performance relative to WFE in terms of outperformance. I think this year, the outperformance will be something in the high teens. So how you're thinking about that next year? Or maybe if you could talk a bit about where you see growth within the different subsegments of WFE?

Sure. I think in terms of growth, we are actually seeing resiliency in leading-edge or advanced logic and foundry, which plays to our strength because that is our biggest play. And actually, that is also where you find ALD being used the most. And to some extent, those are epitaxy. So for us, that's very comforting. That's a strong area. Memory, yes, it will be down, but it depends on how much it is going to be reduced compared to this year because we do have something like a compensating factor that over the last, I would say, 18 months or so, we have had new penetration wins, but will those penetration wins be enough to compensate or offset the drop in overall CapEx for Memory, again, that's still left to be seen. Power/analog, we are actually growing. We have very good traction especially with new equipment that we have launched over the last, I would say, 24 months. The reception from customers are very good. and we have had opportunities to win over some new customers as well. So if you look at that, I think next year, whether we outperform or whether we don't outperform, that's still left to be seen. It's a little bit too early. But generally, we are quite happy with how things are trending for us going to 2023. And the other factor that makes us kind of comfortable and confident is we're going to probably exit the year with quite an elevated backlog. We already have a very elevated backlog at the end of Q3. I think that will continue as we enter into 2023 as well.

Got it. So when I hear all of that, it sounds to me like you'll still outperform. It's just a question of if you don't know where WFE will end up and...

That's one. And it's a little bit too early for us. We normally do not give this kind of guidance at this stage. So you'll be better when we get into maybe the next earnings cycle, perhaps that we have a better view, maybe we can share that.

Okay. Got it. And maybe a question for you, Paul. Just on the book-to-bill comment you made yesterday in the press release. You talked about it being just a little over 1. Is that being driven by the Chinese orders being rebooked into the backlog? Or is that driven by underlying demand that you're seeing at this stage?

It's a little bit the same as for the revenues. The biggest part is indeed from rebookings of China orders, but also the other business is maybe slightly stronger than we anticipated. But again, the biggest part is China.

Got it. And would it be fair to say that China is more profitable for you -- domestic China is more profitable for you than the rest of the business?

The products that go to advanced fabs that we actually took out with our Q3 earnings update are indeed to have above-average profitability, and above average means we said it's north of our guided range of north of 50%. We haven't even said how much, north of 50%, but yes, so this is also somewhat of a positive to the margin. Yes.

Got it. Now as we look into 2023, now you haven't been able to deliver everything you wanted to deliver this year because of supply constraints. So how do you see that changing in 2023? Do you think you will still be supply constrained? And if so, are there any specific bids in terms of semiconductors that are constraining your deliveries?

I think we are seeing improvements in the supply chain. Unfortunately, there are certain buckets which are still constrained. And our view is that, that is going to continue into 2023. For how long into 2023? That's still left to be seen. But for sure, the supply chain has been improving. It's just a certain -- to maybe give you an example, microcontrollers and stuff like that, still in a tight situation. I think if we are able to see even more improvements in the supply chain, of course, we'll be able to convert more of the backlog into revenue. So it could be that at some point next year, you might see revenue getting higher just because we are able to convert more of the backlog. Yes.

But it sounds like your premise into 2023 is or at least based on what your suppliers are telling you, you will still have supply constraints.

We will still have some constraints based on the visibility that we have some of these components will still have some constraints.

Got it. And then maybe just coming back to the backlog. You've had the CapEx cuts that have been announced and then Micron announced them twice. So what have you seen in terms of -- in your order book, are you seeing orders getting pushed out or the delivery being rescheduled or you think any concellations.

We are seeing some pushouts, but they are limited to Memory, which is not surprising given what has been announced so far. So we do see some pushouts not cancellations but pushouts from the Memory segment.

Yes. Got it. Now slightly changing pack and coming to LP. So now if you look at -- you said that LP will contribute a little over EUR 100 million to revenue next year. And that feels like a very, very steep ramp. Now having said that, when I look at the silicon carbide market, demand has been continuously surprising to the upside and supply as well. So do you think there could be some upside to that EUR 100 million number in a maybe 12 months down the line when we're looking back, is that a possibility? Or is there a set plan and your capacity will mean that it will be hard to deliver much upside to that?

Sure. We are in the process of also trying to increase let's say, capacity to manufacture even more. So a little bit of background. LP is a family-owned business, very much just based in Milan with a secondary process developments in Catania. So they don't really have a lot of infrastructure that enables them to build even more. So we are in the process, having taken over now to try to see what we can do. I think we are very confident of what we can do next year to hit the target of north of EUR 100 million. We see actually a strong demand coming in as to how much can we deliver. I think they also have, to some extent, some supply chain, plus we need to expand our capacity. So I think how much we can stretch that are still left to be seen. But we are very confident of hitting the target of north of EUR 100 million.

Got it. Now the growth that LP is expected to deliver, it feels a bit like it will be far greater in excess of the market growth. Would you say LP has some secret source that will help it gain share? Or is it mainly driven by the transition from 6-inch to 8-inch within silicon carbide driving more demand for single wafer...

I think there's a combination of both. If you look at 6-inch per day, they have, I would say, a good position, and that is where a lot of the high-volume manufacturing ramp is being done. And they have a good position there. Over the next 2, maybe even 3 years, you see the transition from 6-inch to 8-inch. And what we're seeing today is very, I would say, healthy engagements with almost all of the global players in working together, whether it's for evaluation or whether it's -- let's say, for R&D, including some which are already potential wins for the transition to 8-inch. So the transition to 8-inch will be, I would say, a big positive for the LP business. timing wise is a little bit difficult. I think the industry has to solve the substrate problem first. Whether it's going to happen in '24 or '25, still left to be seen -- but generally, the industry agrees that, that is the path forward to produce even more silicon carbide, the semiconductors.

Yes. Got it. So it sounds a bit like even within 6-inch, LP is going to gain share in the near future.

I would not go to that extent yet because we do not know what is the share situation today. For sure, I would say that they have a very strong demand.

Got it. Now maybe slightly changing tack. We talked about the China rules. There's been a lot of talk about the Dutch government or the EU imposing their own rules. It's been in the press and the Dutch foreign minister has been talking about it. So can you talk a bit about any preliminary discussions you're having internally? Or how do you see this affecting you guys going forward?

No, I can say something. I mean, one, start with -- we don't know, of course, what the ultimate outcome will be. But what I can say is that I think the Dutch government agrees with the objective of the U.S. government, but there's a different view on how to achieve that objective. And that's where the discussion is about. So I think that the Dutch government. And again, if they will be able to get that through, I don't know, would most likely prefer a more specific and more targeted approach to achieve that objective. While the U.S. has taken a more broader approach at even non-advanced products, as an example, can not be delivered to advanced fabs under the current U.S. restrictions. And that's why you could ask. So that's not really serving the objective of the U.S. government. And so that's where amongst all the discussion is about what the ultimate outcome will be, I don't know. Of course, we have many discussions with them, like our neighbor in the south, also discussions with them, we educate, we try to influence. But yes, what will ultimately be decided, of course, we don't know.

Got it. And would you say it would be a Dutch initiative? Or would it be an EU-wide initiative so there could be more players?

We tried to make it an EU initiative as well because ultimately, this is relevant for the whole EU and not only the Netherlands to what extent, again, the Dutch government does that. And we know that they do talk within the European community about this, it's difficult for us to see and to tell, but it's definitely on the EU radar as well.

Got it. Maybe then changing tack and looking at the financial model for next year. This year, if you take OpEx, OpEx is continuously surprised to the upside versus what people were expecting. So next year, how are you guys thinking about OpEx to the extent that you can comment?

Yes. So what we're planning to do next year is maybe step backwards. So this year, R&D but also SG&A year-to-date, let's say, year-to-date September, year-to-date Q3 grew by whatever, close to 50% [indiscernible] that is a significant step-up, and that was needed for a number of reasons. So next year for R&D, we plan again a meaningful step-up, not to the extent you've seen this year, but still a meaningful step-up, even knowing that there is a level of uncertainty in the market, but there are so many opportunities out there. We just want to make sure that we try to capture those and continue to invest there. For SG&A, we made a big step-up. And based on, let's say, Q3 run rate, I would expect a marginal increase a little bit and to a certain extent, also dependent on revenue development because yes, the revenue development is a little bit uncertain. But -- so R&D continue to invest, and we will pause in SG&A.

Yes. got it. And would you say those R&D plans are fixed? Or would you be willing to look at them if demand turned out to be a lot weaker than in a...

It's -- of course, we will look at the market. We will look at our revenue development, but it would have to become pretty extreme before we would deviate from what we're actually planning now. I mean, again, there's so many opportunities. And what we should not do is because maybe of 1 year potentially a different revenue figure, slowdown on the innovation because so far, why are we where we are because of our investment in technology, and that continues to be very important to the company. And we also want to continue that next year. But having said that, of course, we will look at the P&L. We will try to properly manage this and no worries. But this has a high priority at least. That's what I want to say.

Okay. Got it. Maybe then changing tack and you guys have provided medium-term targets. You've talked about defending your share in ALD and gaining share in epitaxy. Now some of your competitors have been very vocal about doing -- gaining share in the ALD space. So I just wondered how much visibility do you have in terms of what's going to happen at the 2-nanometer and 1-nanometer node. Are your design wins already done. So it's just a case of ramp? How much confidence do you have in defending your share in ALD and gaining share in epitaxy.

First off, maybe a general situation of where we are with 2-nanometer gate all around. I think you'll find all the 3 players are still in the process of finalizing or developing the technology, and they are heavily engaged. And for us, we probably have not seen a certain note, where we are also that heavily engaged. I mean we have a lot of engagements with all 3 of them. So that's actually one of the reasons why just alluding to Paul's point, we need to continue the R&D.

And this is on ALD or...

ALD while epitaxy to some extent as well. But a big part of it is ALD. So if we look at that, and the 3 of them are varying stages they have -- none of them have fully completed the process and planning, let's say, the next step, they go into pilot or high-volume manufacturing. So some of the, let's say, applications, we can see where we stand. We can see, for example, okay, how do we spend, how do we stand, if there's competition, no competition. So we have a pretty good idea. And of course, our competitors have been kind of making announcements, some wins that they may have and so on. Of course, there's going to be some wins, some losses. But overall, if we look at how 2-nanometer get all the ones going to play out for us, we are very confident we will maintain our leading position.

Yes. Got it. And maybe -- can you talk a bit about what key differentiating factors are when your customers affecting your tools, both on the ALD side and then on the epitaxy side?

On the ALD side, if you look at our history, we started in ALD much earlier than anybody. And over the 2 decades or so, we have developed a lot of expertise in terms of chemistries, precursors, materials. So we have become the broad-based ALD player. In other words, we can play in all the different applications, the ALD serves today. Some of our competitors are, in fact, most of them came into the game late. So they have been able to develop applications that serve, for example, a specific part of it. So very often, you'll find that we are the board-based player, and we have to compete with all 3 of them. but not for the same application but for different applications. So if you look at that one area that we also have significant strength because of our early involvement is also installed base. So we have the largest installed base. And from node to node, not everything changes. So if those things that don't change, they're automatically or to some extent, they just carry forward, that is already secured for us. So the competition is really in all the new applications that comes about at a new node. Epitaxy is a kind of a different story for us. Here, we are the follower. So we are up against a competitor that has a significant part of the market share. And what we tried to do here was to -- when we started developing our new equipment a couple of years ago to make sure that we focus on developing a good technical product because we know the only way that we can win is we have a good technology that can compete because our competitor is much bigger than us. There's no way you can compete with them commercially. And that's exactly what we did. We came out with a product called Intrepid. We had our breakthrough with the leading family at 7-nanometer and that's continuing. Last year, we announced a second advanced CMOS win. We have a second customer win for gate-all-around. And so the advanced CMOS area, we are continuing. So today, we have 2 customers, but we are involved with all 3 of them at 2-nanometer gate-all-around development. And we do expect to win some of the layers. That's the part of the epitaxy market is involved with advanced CMOS. You also have the part, which is more on the trailing edge but it's very much in the power electronics and wafer manufacturers. And there, again, last year, we launched a new product called the Intrepid ESA and the reception by those customers has been very good. We have won the new customers as well. And that part of the market will continue to grow. So the 2 of that added together give us a lot of confidence that in a couple of years by the time we exit 2025, we would have at least double our market share.

Got it. And maybe just briefly on gate-all-around, would you benefit more on the epitaxy side or on the ALD side?

I think the ALD part is trigger. There will be also more epitaxy layers.

In terms of growth, would it be bigger on the epitaxy side, absent the actual growth in maybe delay accounts, et cetera.

It will still be on the ALD.

Right. Got it. Got it. And then maybe before we move to Q&A from the audience, just one question maybe for you, Paul. I just wanted to understand how much success are you having with pushing price increases to your customers in this inflationary environment? And is that in any way putting you at a competitive disadvantage wanting to extract more pricing?

Yes. So we're -- when it depends on the contract structure that we have with customers, which is different. Some it's by order base, annual or whatever. So where possible, we have already increased prices with good success, I would say, which is good. For most of the orders, we have announced that we will increase prices. And where possible because we have, of course, a few large customers, they're not easy and they're not really applauding of course, when we come with this message. We try to do that with what we call customer improvement projects. So do it hand-in-hand with a value creation opportunity for them and for us, a win-win for both where we take, of course, the biggest chunk of that to make sure that we can offset the inflationary pressure. That's not always possible. So even if it's not possible, we still try to increase prices, and then we have announced that we will do that. And yes, based on everything I know today is that we're pretty confident that we will be able to offset the cost inflation.

Got it. And maybe at this stage, maybe we'll take some questions from the audience. If you guys have any questions, please put your hand up.

I have 2 questions. The first one just about the revised China guidance. So now at 15% to 25%, I'm curious on the remaining amount, the 75%, 85% you can ship because the restrictions on Memory and advanced logic. So I just want to understand why 80% of the tools? Is it mostly mature tools or you're seeing certain memory or advanced logic you're able to ship?

So basically -- so we announced our China sales with the Q3 update, which is 16% year-to-date Q3. We said approximately 40% of that in a conservative scenario, we believe we can ship. And we have the latest announcement we said more or less half 15% to 25%, let's say, 20% for now. so half of that. So we can still ship to non-advanced fabs, basically all products; and to advanced fabs, it's mainly products developed in the U.S. because one of the U.S. persons involvement and two, because you cannot transfer technology to advanced fabs. We produce outside of the U.S. So in theory, we can actually ship every product, also the ones that we think we will not ship simply because we cannot service them thereafter. We cannot update through software. We cannot, let's say, transfer of knowledge, train, do other upgrades that are needed. And that's, of course, a problem for our customer. So we would think that customers would not order these products. Of course, that's up to the customer to do whatever they like to do, but that's the current situation. So -- there's a big chunk still non-advancement you can ship and then a big chunk of advance, we can now ship without U.S. personal involvement and then the remaining part is what I just mentioned, U.S. products, which most likely cannot be shipped.

To clarify, is there any risk like because you have U.S. suppliers going into that, that might be restricted that now you're shipping and [indiscernible]

We -- every product that we have does not -- has less than 25% U.S. content. So if and when produced outside of the U.S., you can ship into China with less than 25% U.S. content.

Okay. I was just saying like you might have some of your competitors or like U.S. suppliers, they can't ship the tools. So then our customers still taking delivery of the products they can get from you. Like do they have a risk that can't complete the line and eventually have to...

So one of the things that we are now doing is, of course, reconfirming with customers if they still want the tools, they ordered, let's say, 1 or 2 quarters ago amongst others for this reason. Because, indeed, they will need all products to complete the line. And if you cannot get certain products, then they might also don't need our product. We don't know yet. But there is this risk that there's an indirect impact that is correct.

And the second question I wanted to ask on gate-all-around because my understanding was one company was already going ahead with gate-all-around, but I think it's a preliminary process. So are those ALD decisions done or it's more of the performance version 2024, so there's still work or design in.

So I think if you look maybe a little bit further ahead. So all 3 of them have announced that they are going into high-volume manufacturing for gate-all-around in 2025. And if you work backwards, then what is -- it varies from customer to customer. But you could say, by early 2024, they would have to have made all the -- or finish the development, made all the decisions and started all the equipment, so that they get it probably sometime within 2024, and then they take probably the better part of a year or so to start 9 to 12 months to start tuning and are preparing for the high-volume manufacturing ramp. So that's where we are today.

[indiscernible]

We already are inside there with certain applications. I would say that, that is a very preliminary version of gate-all-around. And I'm not sure how much of a high volume they are planning to do, but I do believe that the bigger high-volume manufacturing for them is still going to occur at what is called L2 or 2-nanometer. Yes.

Got it. I think we're out of time with that. So let's end it here. So firstly, thank you very much for making the time and attending our conference, Ben, Paul, and thank you to our audience for your patronage. So we'll leave it here. Thank you, guys.

Thank you, Adi, for having us.

Thank you very much.

Thank you.