艾司摩爾 (ASML) 2020 Q3 法說會逐字稿

Thank you for standing by. Welcome to the ASML 2020 Third Quarter Financial Results Conference Call on October 14, 2020. (Operator Instructions) I would now like to -- for a moment, I would now like to open the question and answer queue. (Operator Instructions)

I would now like to hand the call over to Mr. Skip Miller. Please go ahead, sir.

Thank you, operator. Welcome, everyone. This is Skip Miller, Vice President of Investor Relations at ASML. Joining me today on the call is ASML's CEO, Peter Wennink; and our CFO, Roger Dassen. The subject of today's call is ASML's 2020 third quarter results. The length of this call will be 30 to 60 minutes, and questions will be taken in the order that they are received. This call is also being broadcast live over the internet at asml.com. The transcript of management's opening remarks and a replay of the call will be available on our website shortly following the conclusion of this call.

Before we begin, I'd like to caution listeners that comments made by management during this conference call will include forward-looking statements within the meaning of the federal securities laws. These forward-looking statements involve material risks and uncertainties. For a discussion of risk factors, I encourage you to review the safe harbor statement contained in today's press release and presentation found on our website at asml.com and in ASML's annual report on Form 20-F and other documents as filed with the Securities and Exchange Commission.

With that, I'd like to turn the call over to Peter Wennink for a brief introduction.

Thank you, Skip. Welcome, everyone, and thank you for joining us for our Q3 2020 results conference call. I hope all of you and your families are healthy and safe.

Before we begin the Q&A session, Roger and I would like to provide you with an overview and some commentary on the third quarter as well as provide our view on the coming quarters. And Roger will start with a review of our Q3 financial performance with added comments on our short-term outlook. And I will complete the introduction with some additional comments on the current business environment and our future business outlook. Roger?

Thank you, Peter. Welcome, everyone. I will first review the third quarter financial results and then provide guidance on the fourth quarter of 2020.

Net sales came in above guidance at EUR 4 billion, primarily due to additional EUV system revenue. We shipped 10 EUV systems and recognized revenue from 14 systems this quarter. For the 4 systems that shipped in Q2, but did not receive factory acceptance testing before shipment, we were able to complete customer site acceptance tests and recognized revenue this quarter, bringing the total to 14 EUV revenue systems in Q3.

EUV system revenue this quarter was 66% of total system revenue, which is the first time EUV was higher than deep UV system revenue. This further confirms EUV has entered the realm of high-volume manufacturing and is an integral part of our core operational activities.

Net system sales of EUR 3.1 billion was again more weighted towards Logic at 79%, with the remaining 21% from Memory. The strength in Logic is driven by the high EUV revenue.

Installed Base Management sales for the quarter came in at EUR 862 million, showing continued strength in our service and field option business from the beginning of the year.

Gross margin for the quarter was 47.5%, coming in at the midpoint of our guidance, which is a good outcome considering the significant EUV revenue. On operating expenses, R&D expenses came in at EUR 534 million and SG&A expenses at EUR 132 million, which was slightly better than guided. Net income in Q3 was EUR 1.061 billion, representing 26.8% of net sales, resulting in an EPS of EUR 2.54.

Turning to the balance sheet. We ended third quarter with cash, cash equivalents and short-term investments at a level of EUR 4.4 billion, which is the same level as last quarter.

Moving to the order book. Q3 system bookings came in at EUR 2.9 billion, including EUR 595 million for EUV systems. We saw some EUV demand reduction due to a delay in customers' node timing, resulting in a net booking of 4 EUV systems. Order intake was largely driven by Logic with 86% of bookings and Memory the remaining 14%.

With that, I would like to turn to our expectations for the fourth quarter of 2020. We expect Q4 total net sales of between EUR 3.6 billion and EUR 3.8 billion. We expect our Q4 Installed Base Management sales to be around EUR 900 million, which is driven by strong demand for field upgrades and growing service revenue with an increasing contribution from EUV service.

Gross margin for Q4 is expected to be around 50%, which is significantly higher than Q3, driven by higher immersion volume and improved deep UV product mix. The expected R&D expense for Q4 are EUR 550 million, and SG&A is expected to come in at EUR 140 million. Our estimated 2020 annualized effective tax rate is still expected to be around 14%.

Finally, I would like to talk about capital allocation and working capital. As mentioned last quarter, we are in a transition period with customer contracts as we work to move towards new contracts with improved payment terms. We do see some of these new contracts starting to materialize and expect to improve our free cash flow generation in the coming quarters.

Interim dividend over 2020 will be EUR 1.20 per ordinary share. The ex-dividend date as well as the fixing date for the euro-U.S. dollar conversion will be November 2, 2020, and the record date will be November 3, 2020. The dividend will be made payable on November 13, 2020.

Conditions in the COVID environment have improved around our ability to operate and our assessment of our supply chain. Taking this into account, along with our improving free cash flow generation, we will resume executing share buybacks this week, in line with the plans that we have communicated earlier this year for a total of EUR 6 billion over 3 years.

With that, I'd like to turn the call back over to Peter.

Thank you, Roger. As Roger highlighted, we had a very strong quarter with EUR 4 billion in revenue and good profitability, driven by strong growth in Logic. We expect Q4 to be a solid finish to the year in both sales and profitability. And in spite of added macro uncertainty in the first half of the year due to COVID-19, our view on growth this year is largely unchanged from what we believed at the start of the year. This is a clear reflection of our customers' drive to innovate and continue to invest in future technology nodes.

In Logic, customers continue to see strong demand for advanced nodes, in support of the buildup of the digital infrastructure, which includes secular growth drivers such as 5G, AI and high-performance compute. And as we're still in the early stages of this digital transformation, we expect Logic demand to remain healthy and continue to drive demand for our products.

In Memory, customers are continuing to indicate that they're seeing healthy demand in data centers, with improving demand for consumer electronics. With customers’ expectations for higher bit growth next year and taking into account the longer lead times and qualification schedules for advanced litho, we are starting to see a recovery in lithography demand for DRAM, with strong growth expected in Q4 this year. Based on the continuation of this improving end market environment, we expect this Memory recovery to continue into next year.

Sales to China contributes -- continue to grow and accounted for 21% of our systems revenue this quarter. We expect sales to our domestic Chinese customers to grow to above EUR 1 billion this year, which includes sales to both Logic and Memory customers in China, with the mix skewed towards Logic this year but trending to higher Memory sales next year.

Regarding U.S. export rules to China, we are aware of the requirements set by the U.S. Commerce Department for specific companies in China and as such, according to the current regulation, ASML can continue to ship deep UV lithography systems from the Netherlands. ASML requires a U.S. export license for systems or parts that are shipped directly from the U.S. to the customers affected by the rules. While it is not our policy to comment on the individual customers, we aim to serve and support all of our customers around the world to the best of our abilities, whilst being, of course, compliant with laws and regulations set by the jurisdictions where we operate.

On our Installed Base business, we still expect significant growth this year. Through the first 3 quarters, we realized revenue of around EUR 2.6 billion. And as Roger mentioned, we expect another solid quarter in Q4. Service business will continue to scale as our installed base grows with increasing contribution from EUV service revenue, as these systems run more wafers in volume manufacturing. We expect significant demand for upgrades as customers utilize upgrades to increase capacity and improve imaging and overlay performance required on future nodes.

On EUV, although our customers are still climbing the maturity curve, we continue to see increasing customer confidence in the technology, which is translating into expanding layer counts in Logic, initial deployment of EUV in Memory and increasing service revenue.

With 10 shipments this quarter, we have shipped 23 EUV systems year-to-date. With completion of customer site acceptance test and revenue recognition of the 4 systems shipped last quarter, we achieved a remarkable EUR 2 billion of EUV system revenue from 14 systems in Q3. We're still planning to manufacture 35 systems this year, but due to the pace of customers' node ramps and their fab readiness, a few systems may end up being shipped early next year. But despite this potential shift of shipments to early next year, we're still targeting EUV revenue to approach EUR 4.5 billion this year.

We continue to drive profitability of our EUV systems and service business. We are on track to achieve at least 40% system gross margin, and we started to breakeven on service business this quarter. We will continue to drive margin improvement in both systems and service costs via cost reductions and delivering more value. And as we said before, we expect EUV margins to -- we expect EUV to reach margins comparable to deep UV margins over the next 2 to 3 years.

We are on track with our EUV cycle time reduction plan to get to 20 weeks by the end of the year, enabling a capacity of 45 to 50 systems. With respect to demand for next year, we currently have an EUV systems backlog of EUR 6.2 billion exiting Q3, with around 65% of this backlog planned for shipment next year. While we expect more orders in Q4, we did see some EUV demand reduction for next year due to a delay in customer node timing, which resulted in net bookings of 4 systems in the quarter. Although there is clear uncertainty due to the current macro environment as well as exact timing, slope of ramp and the ultimate size of the customer nodes, we currently expect EUV systems revenue growth of around 20% next year.

In our deep UV business, we qualified the first NXT:2050i in Q3, which shipped early Q4. This immersion system is based on a new version of the NXT platform where the reticle stage, the wafer stage, the projection lens and the exposure laser all contain performance enhancements. With these innovations, the systems deliver increased customer value via improved performance in overlay and productivity and are, therefore, critical in support of their next node introductions.

To summarize 2020, in spite of macroeconomic uncertainty in the first half of the year, we see the year playing out quite similar to what we saw at the start of the year. We expect to end another strong year in Logic, Memory growth of over 30% and significant growth of over 20% in our installed base business. With this, we expect double-digit growth in both sales and profitability, leading to estimated revenue of at least EUR 13.3 billion.

As we look to 2021, it's too early to provide a detailed guidance as we are working with customers to determine demand plans going forward. While there are still significant uncertainties, we expect another year of low double-digit growth, largely driven by our current view of expected EUV systems revenue growth of around 20%. There are a number of elements that will determine the degree of growth and uncertainty, of course.

First is the macro environment because nobody can predict the global economic impact of COVID and how this will impact the end markets that we serve. On top of this, there is also the geopolitical environment, predominantly the U.S.-China dynamics, that creates additional uncertainty.

In Memory, demand will depend on bit growth next year. Customers seem to broadly believe the inventory issues will have been normalized by the end of this year and expect stronger bit growth in 2021. We currently see strong litho demand for Memory next year, which is consistent with customers' comments. However, the degree of growth will, of course, depend on continued technology transitions and how much capacity will be added.

In DRAM specifically, we see support of this expectation through high utilization of our litho systems in the field at the moment. While we expect to see more EUV systems go to DRAM next year in support of 1-alpha node, Memory is still a key driver for our deep UV demand.

In Logic, we expect demand will remain healthy. However, final demand will depend on timing and the slope of node ramps, driven by the end demand curves. Customers are continuously recalibrating their road maps, leading to changes in their shipment requirements, which will likely have an impact on our demand next year. The other aspect of demand timing is around the slope of the node ramps. And this will be determined by how many wafers will move into each of the foundry nodes and is, as I mentioned before, a function of the health of the end demand next year.

Lastly, on Installed Base business, we expect continued growth, but the degree of growth will be more dependent on upgrade businesses as service business is pretty predictable and grows with Installed Base.

In summary, there are a lot of dynamics at play, both at the macro level, geopolitical level as well as market-specific circumstances. However, the ongoing transformation of the digital infrastructure, along with the secular end market drivers, such as 5G, AI and high-power compute, will continue to fuel demand for advanced process nodes, both in Logic and Memory, which drives the demand for our products. Therefore, although we are currently going through a period of near-term uncertainty, the long-term demand drivers only increase our confidence in our future growth outlook towards 2025.

With that, we will be happy to take your questions.

Thank you, Roger and Peter. The operator will instruct you momentarily on the protocol for the Q&A session. (Operator Instructions)

Now operator, could you have your final instructions and then the first question, please.

(Operator Instructions) Our first question is from Mr. Krish Sankar.

It's Krish from Cowen. I had 2 of them. First one, Peter, on the 20% EUV growth for next year in terms of revenue, what does it imply for units? Is it around 40? I'm just trying to figure out what the unit number would be for next year for EUV? And then I had a follow-up call?

Yes. Krish, you need to understand, I mean, we always get this question on the unit numbers, but, at the same time, we always get the question on what -- why is the ASP in that quarter higher, lower. For instance, last quarter, it was like EUR 145 million, which was last quarter, the result of the specific configuration sets and repair, upgrades that were included in the shipments. So this ASP, because what you're asking is the number of units times the ASP, gives you the euro sales number. And this is why we just give you the euro number, yes? And it depends on the configuration and the richness, you could say, of the configuration that actually determines ultimately how many of those units we will ship.

So we're not guiding any units. We just give you a euro number that we believe is going to be 20% up from this year, which is around EUR 4.5 billion this year. So I think it's going to be at least a 20%, as I think we've looked at the demand picture. I think we derisked it for some of the recalibrations of the nodes and the nodes timing, and we're coming up with that euro number. And I think going forward also, we will do that. Just like we do with deep UV and with the total company sales, we just don't tell you how many units we sell. We just tell you what we think how much euros we will sell.

Got it. Peter, that makes sense. And then as a follow-up, on the DRAM side, clearly, you're seeing continuing strength from DRAM for deep UV. There were some concerns in the marketplace that the DRAM makers might scale back some of their near-term CapEx because of Huawei going away. Have you guys seen any of this nuance? Or just because of your long lead times, you're kind of agnostic to such noise?

Yes. I think you know, we had this question on the impact of a potential drop of Huawei as a key customer for our customers. When Huawei doesn't sell those smartphones or the digital infrastructure, somebody else will. So I think that in itself is not going to be a driver. It can only be a short-term disruption in timing. Demand might fall away, it might move to somebody else, but that probably takes some time before it pops up. So I don't think that's a structural issue.

No, I think we just need to look at what the situation is. We have got a push from our DRAM customers. We corroborate that push with what we see in utilization of our tools. And our -- the utilization of our tools is high. And it's grown, I would say, relatively gradually to this level throughout the year, yes? And I think if you then listen to our customer, they said, well, we think that the inventory situation will normalize towards the end of the year with a utilization position that we're currently seeing that is pretty close to what we think is kind of the maximum capacity, then it's not strange to conclude that they need more capacity next year, especially if you look at the demand coming out of the server business. And also, we have the uptick that we're seeing in consumer electronics. So all in all, that makes us more optimistic, together with our customers, on the DRAM business going forward.

Our next question is from Mr. Joe Quatrochi.

It's Joe Quatrochi from Wells Fargo. Just back on the 20% EUV growth for next year, can you help us understand, is that assuming that a few tools slip from 2020 and are shipped actually in early 2021? And then secondly, on that, is there a scenario that as we get into January and you get better visibility, there's actually some upside drivers of that 20% growth, ex the systems that could potentially fall from 2020 into 2021.

Yes. That's a good question. Well, as we said, 2020, we always said about EUR 4.5 billion, which we stick to that. And then -- so the guidance of the 20% up from that EUR 4.5 billion, of course, includes all shipments that go from 2020 into 2021. You can draw the conclusion from that, that our forecasted ASP at the beginning of the year, i.e., the configuration richness, yes, that's been a bit better than we anticipated. We've seen that through upgrades, yes? So yes, that moves into next year. And that's why units, I said it on -- as an answer to the question of Krish also, units are not that important because it's -- really it's the ASP that can really vary because of the configuration. So we guide you euros. But yes, it will move into 2021.

Now is there upside, that's your second question. I think what we have -- what we clearly said also in our prepared remarks that we have adjusted our demand picture based on the recalibration that has happened within the customer base on the node timing and on the ramp of that node. That, of course, happens first because we know that. But it's also like you have to look at this almost in a way of communicating vessels. What goes down also goes up at the other side. Now in a communicating vessel, it all happens at the same time, whereas there's a bit of a time lag between when you adjust downwards and when you see that demand later in time coming back up, that is the upside. That is not what we've put into our 20%, yes?

So you could actually say, we've taken all the information of today, taking into account all the uncertainties that we see, taking into account the node calibrations and this is what we'd actually tell you what we qualitatively see today, that there is upside is also clear. Because like I said, it's a communicating vessel, what comes down comes up somewhere. There's going to be time difference. And that's why we're preparing for a potential upside also in 2021. That's why we still talk about a build capacity that is higher, it's 40 to 45 systems, or is that 45 to 50 systems even, yes? And that's what we're organizing for in the supply chain, yes? So yes, there could be upside for the reasons that I just mentioned, and we're prepared.

Perfect. That's helpful. And then just a quick follow-up. On the low double-digit revenue growth for next year, how do we think about OpEx growth?

OpEx growth. I mean on the SG&A side, I think it will continue to develop, as we've seen in the past quarter. So no major uptick there. On the R&D side, we are pushing down the accelerator quite a bit in terms of the programs, both on High-NA, also on EUV. I mean we're also obviously looking at and working on the successor of the D tool, which will have some demonstrable progress and value to our customers, which obviously requires quite some R&D work.

And also on the deep UV road map, there's still quite something to be done. So there, we talked in the past about adjusting the number that you've seen for this year with salary increases. In fact, if we look at the magnitude of the program, it's probably going to be a bit higher. We're in the process of finalizing that right now because we're right now in the finalization of the budget. So more precise guidance, we'll be able to provide you with in 3 months' time. But I think it's fair to assume that the increase in the R&D budget will be a bit higher than just the salary uplift.

Our next question is from Mr. Sandeep Deshpande.

It's Sandeep from JPMorgan. My question is regarding a couple of things. Firstly, you've already indicated that you've seen some pushouts, and you've reported a net EUV order intake in the recent quarter. I mean there has been a pushout from one of your customers, but then that same customer could place orders at another customer of yours. And so have you seen any movement on that front? Or do you expect to see movement on that front in the next few months?

And the question there which arises is that can you satisfy short-term demand? Because you've talked about in the past that EUV tools are 12-month lead time. So whether that sticks.

And my quick follow-up to that question is on the gross margin. You're indicating this 20% revenue growth number into next year for EUV, which is a good number. But clearly, I mean, it implies some amount of unit cuts because of this pushout from one customer. Does this have an implication on gross margin at -- for ASML next year on EUV from the 40%, so this progress from the 40% towards a much better EUV gross margin?

Okay. So you guys are all squeezing 5 questions into one, which we haven't been a little bit prepared for. So Roger is going to answer that one.

On your question of the pushout of the net 4 systems, have we then see other customers then stepping in? Well, it's a bit the story of the communicating vessels. And this is what you're asking, in fact, yes, which, by the way, is not retail. When you cannot get it with retail shop A, you move to retail shop B. It's a bit more difficult. Not so much for us because we could probably do that. But it's for the customers.

If one customer adjusts their node timing, it takes time for others to see what that means and how that needs to be digested. So this is what I said in an earlier answer, yes, I see upside because there's a timing difference, if you think about the communicating vessels between things going down and the other part going up again. So I definitely see that. This is why we're preparing also to be able to ship more because to your point, if you would really only follow the order intake and then add 12-month lead time, it's never going to work. So you have to prepare for more because you would expect that there is a decent chance that you see some upside next year and you want to be ready. So this is where the 12-month lead time is actually -- it's -- the integral lead time is more than 12 months, it's more like 18 months.

But if you then prepare yourself because you do expect this, then you could, of course, accept POs with a delivery time that's way within that 18 months. So yes, we are preparing for more. And time will tell whether customers need it in the time frame 2021. I think they could, but time will tell. There's a real opportunity there.

Sandeep, on the gross margin for EUV and then maybe a little broader than that. So on EUV, you're absolutely right. So the 3600D will see an increase in gross margin. And I think we have said in previous calls that we hope with the 3600D to approach the corporate gross margin. So an improvement over the 40% that we targeted for this year. And as we already said, we're probably going to even slightly exceed that target for this year. So yes, from that vantage point, there will be an improvement in the gross margin for next year.

The downside on EUV is that in the gross margin for next year, you will also see that we already have some operating expenses for High-NA. That is a little bit accounting. And with the CEO and the CFO of this company both holding degrees in accounting, we say with a bit of pride. But the downside is that, even though we're not selling High-NA next year, we already have some operating expenses in there, which find their way into the cost of sales, and therefore, in the gross margin. And that, we assume, will be about a 1% reduction in comparison to the gross margin that we have this year, just to give you an order of magnitude.

So a clear plus from the increased gross margin on the D. Also, an increase in the gross margin of EUV service margin that we talked about and then a little bit of a negative there. And then if you look at everything also into the deep UV business, for instance, that then becomes a little bit of the swing factor for next year. The mix in deep UV for next year and also the level of field upgrades.

All in all, target that we had for -- or that we have for Q4 and also the target that we have a little bit moving forward to 50%, with all the pluses and minuses, we're probably going to be somewhere in that range of the gross margin that we had for this year, for the full year, and then the 50% target that we're envisaging. That's where you're going to find yourself in that mix with all the pluses and minuses that I just discussed.

Our next question is from Mr. Andrew Gardiner.

It's Andrew Gardiner from Barclays. Just another one on EUV, please. Just in terms of some of the distinction over what you're seeing with customer plans for next year, I think we can all appreciate some of the, as you say, node, sort of, migration, recalibration. But in terms of the commitment to the technology, particularly by those who are leading the way in deployment, are you -- where are we in terms of EUV layers layer count looking to the next node sort of in 3-node relative to earlier expectations.

Peter, you made a comment that customers are more confident in the technology. So are you, even if ultimate capacity may be taking a little longer to ramp in terms of wafers, are you seeing the layer count continuing to creep higher? And so that's Logic, but also similarly, in Memory with DRAM press reporting one of your major customers visited you this week, talking about EUV and commitment there. So similarly around DRAM, where are we in terms of layer count expectations?

Yes. I think on the layer count, I mean, it's clearly evidence given to us that the customers have embraced the technology, and that the technology gives a lot of advantages in simplification, in lower WIP, lower work in process, faster turnaround times in our R&D. So that's -- when you take it all together, then I think on N5 in Logic, we're over 10 layers; and in N3, we'll be over 20. And we actually see that creeping up. So -- and that has just a fact that it gives so much more advantage to go to single patterning and take away these multi-patterning deep UV strategies, which is also true for DRAM.

Of course, DRAM is a bit behind in terms of HVM in introduction as compared to Logic. But also there, we see the same trend. And it's the process simplification that's actually driving the additional layers. And in DRAM, although we start with one, yes, we can see -- clearly, we can see strategies going forward 1-alpha and beyond that will add up to 5, 6 layers. And that's -- but it's really dependent also on how much productivity can we get out of the tools.

You have to realize that customers are currently pushing us extremely hard to get more wafers out. I mean we are introducing EUV tools with more productivity, with more wafers per hour. We have the C version introduced. We're now looking at the D version. And beyond the D version, there's going to be another version with increased productivity. But today, when we have our customers' discussion, it's really a discussion about, could you please speed up the maturity of the tool, get your EUV tools at the same level of productivity and of reliability and uptime as you have your deep UV tools and they know it is going to take time because we need more EUV wafers. We need more. That is the consistent question that we are getting.

So -- and I think it's driven by the fact that they see the big economic and technical benefits of using EUV. And that will drive the layer count up, both in Logic and DRAM. And like I said, N3 of over 20 layers. So that is only creeping up.

And just quickly, Roger, if I could just ask a follow-up to the prior question. When you talk about gross margin, I just want to make sure I was clear on what you were saying. You're saying gross margin for next year somewhere between the 2020 level, which, based on the guidance you've given us for 4Q is going to be around 48%. And the 50% that you're aiming for in terms of fourth quarter, is that right, somewhere sort of in that range?

So that's right. So 48% is -- if you do the math and you take the 50% guidance that we gave for Q4, then you would probably see this year, for the full year and -- for the full year, you would see about 48%. What I'm saying is if you take all the pluses and minuses that we just talked about, you're going to be in that range for the anticipated gross margin for next year. Of course, more guidance on that, where the year is firming up, so more clarity on that in 3 months' time. But directionally, I think that's what you -- that's what you're looking at.

And that's based on the, let's say, low double-digit growth as compared to this year. So any upside on additional business, of course, will have an impact on the margin profile. And as Roger said, we have to just have to take care of this accounting issue on High-NA, where we don't sell High-NA, but we simply cannot...

[We've seen less for] than the growth volume.

We cannot capitalize anything in inventory. We just have to write it off. So this is what is, which, by the way, will give us better High-NA margins once we start shipping High-NA.

Our next question is from Mr. David Mulholland.

It's David Mulholland from UBS. I just want to follow-up on the comments you made around China, Peter. Obviously, it's clear you're still able to ship, particularly if it's out of the Netherlands, to particular customers. But for those that have been facing sanctions from everything we've read, are you assuming any shipments into those customers for 2021. Because whilst you may be able to ship yourself, obviously you face some indirect impact if they're unable to buy from some of the other semi cap equipment space. And then I've got a follow-up afterwards.

Yes. I think it's a good question, David. When we talk to our Chinese customers, and you have to understand that when they look at some of the other process tools, there are some alternatives here or there. There's alternatives in Japan, when you talk about ALD or you talk about deposition tools. There's a European company, which is in Singapore, that can be an alternative. So on metrology, there are alternatives, both in the Netherlands and outside the U.S., in Japan and even increasingly in other parts of the world.

So if you're a Chinese company -- and there's one thing that the alternatives is more difficult on and that's on litho, yes? So you just want to make sure that you get your litho tools. And of course, companies that are on a list where under the current agreements, we cannot ship to, we will not ship to. I mean if there wasn't our agreement, that means our -- we just have to go for -- and we have to look at those lists, and we follow those lists. But currently, the customer that we've been talking about, our Logic customer, and one of the major Logic customers in China, actually allows -- or the rules allow us to ship litho tools.

And it's for those Chinese customers very important to actually get their litho tools and then find a solution for the other process tools. But litho is critical. It's the most critical tool in your fab. So that's why it's not changed -- that's why it's not strange that they still talk to us and say, "Hey, just make sure that you ship us the tools that we need," because it all starts with -- in terms of CapEx, it all starts with the most expensive tool in the fab and that's litho.

That's great. And then just a quick follow-up. You've obviously said you're keeping the door open to more shipments potentially for next year, if you do see upside elsewhere. How long and how much balance sheet are you willing to keep that door open for? Obviously, you've got the funding to do it, but are you making customers commit by the end of this year? Or how long are you leaving that flexibility?

Yes. I think it's a good question. Of course, we try to push it. But on the other hand, and I know customers are listening into this call also, so I'm not doing myself a favor. So -- but part of the difference between the bottom-up numbers that we currently see under these risky -- under these circumstances that are of any more risk and our build capacity anywhere in between, yes, we're going to make sure that we have some extra leeway, some extra room. So we're going to go all out and trying to completely maximize our output between what we currently think in terms of units and that maximum number, somewhere in between, we're going to make sure that we will be able to react.

And like I said, I'm not doing myself a favor because customers are listening in. So it will probably take a bit of time before we get to PO. But if they need it, they will come, and we will ship.

Our next question is from Mr. Amit Harchandani.

Amit Harchandani from Citi. Two questions, if I may. My first question relates to the ramp on the Logic side for EUV. Could you give us a sense for where we are in terms of technological readiness, whether that is playing a part in terms of how customers are thinking about ramping, say, beyond the N5s towards the N3? Appreciate the demand dynamics is one factor. But if you could help us give us a sense for the overall technological readiness of the ecosystem.

And secondly, if I may, on a separate note, you have given us some perspective on contribution from China. Could you maybe quantify that and give us a sense for how derisked that number is? Or what's the puts and takes, if there was a potential change in U.S. legislation?

Yes. On that last part, Amit, I'm not speculating on the behavior decisions of governments because that's basically what it is. You cannot derisk. The only thing that you know, if you look at the current situation, you have to look at the specific situation that this company is in, in the geographical area that we operate in, which is Europe, yes, and the ability of us within the rules and the regulation to ship tools. That's what we're planning for, yes? Now there are a lot of powers at work, which are beyond ASML, and it's very difficult. I'm not going to speculate on what that is, what that upside is, what that down -- what a downside is. I'm trying to just deal with the reality as we see it today. And that's what we do. It's how we derisk or that's what how we look at the Chinese business, yes, which, of course, is just more than -- there are several customers there. We have 4, 5 major customers there. So I'm not going to speculate there because I don't know what governments are going to do.

On the EUV, the technical readiness of EUV, it's technical-ready. The ecosystem on N5 and N3 is there. It's got to do it. It has to do with how quickly can we bring, for instance, the EUV maturity up to the maturity of our deep UV immersion tools, which in -- some of our customers are reaching 98% to 99% uptime. Well, we are not at 98%, 99% uptime. And it by the way varies. We have tools that are over 90%. We have tools that are under 90%. So it is the stability, which has to do with the maturity of the tools. So this is why our customers, in the discussions that we are having with them, are really saying, "Okay, first things, focus on the maturity of those tools," which we will do, which is normal. I mean this is a new technology and high-volume manufacturing, completely normal. We saw the same thing when we introduced dual stages. We saw the same thing when we introduced immersion. It's very logical. It will take some time before we get all the solutions on lifetime of parts all implemented in the field. That's not an issue, just customers want it faster.

Why do they want it faster? Because their customers are just pushing them on more EUV wafers. That's what we see today. The only -- the key discussion that we're having with our customers, give us more EUV wafers, yes, and preferably, by improving the maturity of the tools faster, yes? And then on top, if the end market turns out to be okay, give me more capacity. And that's the order. So this is how it actually works. I have no doubts about the technological readiness at all. This was a question that I -- we could go into a lot of detail a couple of years ago, but not today.

And Amit, if you look at typically the things that are being referred to in terms of readiness for the ecosystem, it's the typical list, right? So you're talking about mask and mask inspection. I think it's pretty clear that the technology is further developing there to cater to that. You're looking at pellicle and actually there, I think you've also recently seen some good breakthroughs in transmission, they are both for ourselves and also for some others.

Photoresist, which we all know is only -- is always going to be there. It's just that it's -- the massive development there will happen at a point in time where the whole thing is going into HVM. That's what you see. So on the typical elements there, you see that good progress is being made. And I think that's also echoed at the EUV conferences by our customers. And they show these slides where they talk about the progress in the ecosystem. And to Peter's point, if you look at that where they had their assessment 2 years ago, there was actually quite some amber and red on the slides. And if you look at the slides that are present today about the readiness of the ecosystem, the red is gone, and it's all green.

And you have to also -- there was just the advantage of being around for a little while. I mean we've seen this also when we did dual stages, when we did immersion, when you have a wavelength change, things are very different for our customers in their production process. And that needs to mature also. It's not only our exposure tool. It's everything that goes with it. It's the pellicles where we started with 82% transmission, now we have qualified 88%. And we have, like we said, there are in research institutes, the transmission that is far over 90%.

The same is true for -- like Roger said, for photoresist. This is very normal. It's very normal. The entire industry, the ecosystem is up, basically climbing a maturity curve. And that's going to be there with us for the next 1 or 2 years before it really starts hitting home on maturity levels that we saw with deep UV. It's just going to take a bit of time.

Our next question is from Mr. Alexander Duval.

It's Alex from Goldman Sachs. You talked about the progression in terms of EUV models and how you continue to invest in that direction. And I wondered if you could talk a bit about the average selling prices. Effectively, where will they go up to on the D and E models? And is it fair to assume 50% of the mix in the next year is going to be coming from the D model? And when do we see the E model roughly enter numbers?

So Alex, in terms of the ASP increase for the D, what we've indicated there is somewhere between 10% and 15% is a good estimate of the increase in ASP. You see -- in terms of productivity, you see a 19% increase in the productivity of that tool. And also in overlay, you see an improvement of 1.5 to 1.1. So some pretty good value that is being provided to the customer, which justifies the uptick in ASP that I just talked about.

On the E, that's quite a while out. So we're not going to -- we're not going to give any indications on the ASP development there. But as I said, there are some pretty significant improvements and value enhancements that are being planned on that tool. But more to talk about that when we're approaching the launch date of that.

On the D model and the composition of next year in terms of units. As we've indicated, we plan on really introducing the D model for HVM at the midpoint of next year. So that kind of gives you a bit of an indication of how the spread is going to be for next year.

Our next question is from Mr. Mehdi Hosseini.

Yes. Mehdi Hosseini, Susquehanna. I have 2 questions with a few follow-ups. It would be great if you could give us the mix of EUV upgrade that is embedded in the field option? And also, if you could give us an update on the multi-beam e-beam. I believe the beta tool was shipped early this year. What's the update? And how do you see acceptance into next year?

Okay. Now the -- I don't think we're going -- we can give you, at least I don't have it, the mix between upgrades and the system price. Is that what you're asking, Mehdi? I mean...

I just want to understand...

Yes. So (inaudible) of our EUV sales, how much is upgrades business and how much is system business. Is that what you want?

Yes. Yes, you have highlighted upgrade opportunities with the older models. And I want to see how much of that is driving -- is in the field option.

Okay. Okay. Yes, you have these -- you have different types of upgrades. You have an upgrade, like for instance, this is why ASP in Q3 was a bit higher. We've upgrades that are already included in the tool that we ship, so that is one. But you also have upgrades of tools that are in the field. And that depends on what upgrade do you want? Do you want an upgrade from a B to a C, that is a different upgrade than one you have a 3350 going into a 3400C. I mean that's almost like a complete refurb in the field. And that's significantly more expensive. And customers are looking at all of them. So it's -- there also, you have the mix of what kind of upgrade do you want? And you have 3400B upgrades, which actually are in the -- from a productivity point of view, that are in the midpoint between the original B and the current C. So this is also what you can choose. So it's very difficult to give you as a forward-looking outlook on what that is because it very much depends on the customer plans and on the capacity planning in terms of wafers per day.

So we're just going to give you the total number, yes, and -- when that happens. And these things, these update plans also change. I had a very recent discussion with a very large customer. Well, he really wants to look at an upgrade plan where there was different scenarios. Well, depending on how that scenario pans out, the upgrade revenue is going to be different than a lot of scenarios. So I wouldn't even dare to give you any guidance there because I would be wrong, yes?

So the MBI, the multi-beam tool. Yes, I mean, we're going to ship another 2 beta tools this year. And like I said, it is really -- customers are going to use those tools, because not just the tool, it's also how do you integrate a multi-beam wafer inspection tool into your entire yield management process. And that will happen over the next half year to 3/4 of the year, and that will determine the uptake of that tool and the POs for that tool.

So what we're seeing is that, at least that's what we believe, and we are now in the first integration phase of the first tool that we shipped last quarter, we just have to see that what we believe the advantages of a tool that is 6x faster, how much of that can translate into real deal value for the customer. That's going to be a bit of a wait and see, but we have people there, and we're optimistic. But ultimately, that optimism needs to be confirmed by the customer that starts putting in the POs. So you have to be a bit patient, Mehdi, on that news. So it's going to be another 1 or 2 quarters before we get some clarity on this.

Our next question is from Mr. Aleksander Peterc.

This is Alex Peterc from Societe Generale. Just one on the balance sheet first. Could you now give us an idea of where we will land in terms of net cash for the year and what will happen with working capital in the fourth quarter, if you could quantify that for us?

And then just secondly, coming back to your comment, Peter, on the communicating vessels in EUV orders. So given the pushouts by one customer, is there a realistic scenario whereby we get a big jump then in orders as that node progression starts to materialize. You get then a very strong '22. Is that a realistic scenario? Or do you see things more smooth out in the future?

Aleksander, on free cash flow, we're not guiding free cash flow. I can tell you that there will be a very significant free cash flow generation in Q4 for a couple of reasons. One reason is the free cash flow generation this quarter was rather low. As a matter of fact, it was slightly negative. And one main reason there is that many, many tools shipped in the last month of the quarter, as a result of which the cash generation of that will fall into Q4. So there will be significant free cash flow generation in Q4. That's also the reason why we communicated that we will resume the share buyback, why we announced the interim dividend. So we're very confident on the free cash flow generation in Q4, but we're not guiding that.

Yes. And on your communicating vessels question, well, it actually was my introduction on the communicating vessel. But the question is, could that then give a big jump in orders because there's a time difference. And you basically asked the question, how big could that time difference be? And I know, like I said, we're preparing for more output in '21 because I think it could happen in '21. But you asked the question about what about '22? Well, you could say, well, we're in 2020 and so what do I know about 2022. But in the context of your question, I do believe that, in any case, 2022, we will see the effects of that communicating vessel movement. I think there's also an opportunity that it can be pulled into 2021. That's why we're preparing for more shipments and from a -- let's say, from a manufacturing and a supply chain point of view.

But more importantly, in 2021, I think we will see then the introduction of the next nodes in Logic and in DRAM. I think 2021, in any case, will be, you could say, the recipient of the node transitions. There will -- 2021 will be a year where you see another node transition happening, which, as we talked about, N3 will have all the 20 EUV layers. So that would then be on top of what I would see as a normal progression of the EUV volumes.

Okay. We have time for one last question. If you are unable to get through on this call and still have questions, please feel free to contact the ASML Investor Relations department with your question. Now operator, may we have the last caller, please.

The final question is from Mr. Adi Metuku.

It's Bank of America. I have 2 questions. First question, just on the assumptions you have embedded into your low double-digit growth guidance for revenue in 2021, can you talk about what you're assuming in terms of domestic Chinese demand? Secondly, also foundry plus Logic demand and Memory demand and services, just to get a sense for what exactly you're embedding.

And secondly, now there's been a lot of discussion in the market around how the number of customers at the leading edge could go down over the next few years. If you could give us some sense based on your discussions and any order trends, et cetera, that you might be seeing, how likely do you think that is, I think that would be much appreciated.

Yes. Your answer -- so what do you want to know exactly, the number of customers in the leading edge? You mean -- are you asking a question on how are this trending?

No, if you look out 2 to 3 years out, do you think the number of customers producing leading edge devices could go down. If not, how do you think about that? How do you think about that when you do your planning, et cetera?

Yes. I think on the...

On the first question...

I'll take the second one.

On the first question, Adi, I think it's -- we've given you the top line figure, both for the total business and for EUV. As Peter mentioned, of course, there are many -- there are quite some uncertainties that have been factored into that number and there are quite some swing factors going left and right. I think what we should do is, in 3 months' time, when we have a discussion on the full year's results, but also going to have our first real insight into next year, I think that's what -- what I suggest is that, at that point in time, we're going to talk about what are the main drivers.

So at that stage, I think it would be good to say, just as we did at the beginning of this year, if we look at the different segments of the market, how do we look at those and what growth drivers do we see there. I think that's the way I would suggest we do that and leave it at the top line indications that we've given for now.

Yes. And I think I will answer your 2 questions on domestic China. I think it's almost not relevant. It sounds a bit arrogant. But you have to realize that the Chinese customers in Logic and in Memory, they are in ramp. They have customers. All their wafers are sold. They go to their customers that put them into devices. So if for whatever reason that China business does not exist, those customers still need wafers, and then they will buy those wafers somewhere else.

And I think where we just said, in Memory, for instance, in DRAM, we see the very high utilization towards the max. So where do those wafers come from, they need to come from added capacity somewhere. And I think this is, again, is a question of this communicating vessel. When one set of customers might not, for whatever reason, need those capacity tools, then another one needs it. And I think with deep UV, that time period is much shorter than with EUV. So I think that is -- it will resolve itself relatively quick.

Now on the customer, the number of customers in the leading edge, there's one thing that's absolutely certain. Things are not getting easier. The next nodes will see increased complexity. And I think only the very large customers can deal with that. We've seen over the last 20 years a significant reduction in number of leading-edge customers. And I think it's going to continue with less than a handful, less than a handful. And I'm not going to speculate who those are going to be because I -- they are all dear to my heart. And I think we have to figure it out who has the best designs, the best production technology with the most efficient and the lowest cost. But in the end, it doesn't matter. Because in the end, it's what we always talk about, is the number of wafers and devices and chips that are needed in the digital transformation that need to be made somewhere on the planet. And that means that I do believe that our large customers are going to be larger, and they're going to be dominant in areas of chip production, which are only going to grow.

It is why I said at the end of my prepared remark, our confidence in our 2025 outlook has only grown. As always, ultimately, you could argue also a bit as a result of what we've seen as a part of the COVID crisis how important this digital infrastructure is. So in that sense, yes, I think it will be fewer and fewer customers, but it will be much bigger than they are today.

Before we sign off, I'd like to remind everyone that due to COVID-19, we have moved our Investor Day to June 23, 2021. The event will be held in London, and we hope by that time, we can have a face-to-face meeting. More details will follow in due time. We hope you'll be able to join us.

Now on behalf of ASML, I'd like to thank you all for joining us today. Operator, if you could formally conclude the call, I'd appreciate it. Thank you.

Thank you, sir. This concludes the ASML 2020 Third Quarter Financial Results Conference Call. Thank you for your participation. You may now disconnect your lines.