Or should we expect a fundamental line growing line because you have a lot of wafers to migrate anyway? And therefore, we won’t see so much lumpiness. Do you see what I mean?
Peter Wennink: Not entirely but I do believe that what is important to us as you look at these memory cycles as this game of chicken. So it’s — you build the capacity, there’s always big step-ups, the underlying growth pattern is also more regular. So you have these times where you have overcapacity. I think that’s just part of it today. We are going through this cycle. We just see the utilization rates of our memory tools going up. I think EUV is always in that sense, cash. You don’t buy loosely extra EUV systems in dynamic space. That’s always going to be the gating item. Now if you then come to the conclusion that you need more EUV because DDR5 and HBM is where the external demand is, then there’s not much to migrate. You just need more — so — and I think that’s on the back of a cyclical recovery because inventories are being consumed.
So that migration, having the space to do migration or the — that — I don’t see that. It might be available in the non-EUV space but in the EUV space, that is actually scarce tool which you don’t have available in abundance and that tool will be the first to be fully utilized. So I don’t know whether it’s answering your question but it’s just — but I’m just trying to think of what you meant or what the other question meant. But I think the recovery in memory is cyclical. And on top of that, I think we have the technology transition into DDR and into HVM. And that drives memory. And I don’t think it’s going to be something that is very short, i.e., one or two quarters. This will continue.
Operator: And the next question comes from the line of Tammy Qiu from Berenberg.
Tammy Qiu: First one is on High-NA. So there has been some concerns relating to High-NA EUV being too expensive to be used for things like 1.4 nanometer. So do you say if you have any feedback on customers’ current development process? I know we’re still at early stage that is in going to be the best option when you go to 1.4 nanometer?
Peter Wennink: Yes, I think it’s a good question because we always need to answer or ask that question to ourselves and to the customers, whether we feel that the next-generation lithography tool is economic for our customers as you know, more loss than a big economics. Now having said that, there is no doubt in my mind now that the high of all our customers that are using EUV is the right choice from an economic point of view. We kept that corroborated now very clearly through our customer contacts and that used to be a question some time ago. But I think everything that we’re currently seeing and also looking at alternative patterns of multi-patterning low-AUV, High-NA is very clearly in vertically the most cost-effective solution.
I think that is also driven by the asset that Roger gave that is as we’ve had — he said last year, we have double-digit orders in the order book and I said in every quarter, we comfortably add a couple of those. Those customers give us those orders because they do those calculations and they see this. So yes, I can understand the question. I think our confidence that it’s the most cost-effective solution, both in memory and Logic, Logic and Memory has only gone up.
Roger Dassen: I think the commitment of customers is not just visible in the orders but for instance, also, as you saw, 1 of our customers has entered into a joint research center which is really focused on this and on the utilization of High-NA. So I think that’s another very strong underpinning of the fact that customers do believe that this is an important way forward.
Peter Wennink: Absolutely. And I think the fact that our launching customer of High-NA was very happy that we were shipping on time and the pressure that they rightfully put on us was also felt throughout our entire reorganization. We have to be on time because this is the tool that they need.
Tammy Qiu: Amazing. And also another follow-up regarding memory adoption of High-NA. So my understanding is, so memory will be on Low-NA EUV for a few years. How would you view this High-NA adoption time line for memory versus 3D DRAM?
Peter Wennink: Versus 3D DRAM?
Tammy Qiu: 5 Yes, I mean which one — which will come first.
Peter Wennink: I don’t think that’s a question. I think high NA will be introduced way before 3D DRAM. So it will just be introduced and I think will be introduced in about the same time frame as logic. So it’s not a competing technology in that sense from a timing point of view.
Roger Dassen: And that’s also clear on the order intake. So in the order intake that we see for High-NA, we equally see memory orders for High-NA as we see logic orders. So I think that is — it is clear that customers are looking at the same time frame for the introduction of High-NA, both into Logic and into Memory.
Peter Wennink: And not considering any trade-off against the 3 DRAM introduction. That’s not — that’s not on the road map.
Operator: And your next question comes from the line of Didier Scemama from Bank of America.
Didier Scemama: It’s Didier from Bank of America. My first question, Peter, when we see this big wave of AI orders coming from the memory vendors and coming from foundries. I just wonder, related to your point earlier that HBM and DDR5 could drive an upside to that 30% contribution to volumes in 2025. Do you think there is upside to your potential demand capacity limit. You’ve got like 75, 80 units from HBM first or from others?
