Nova Ltd. Ordinary Shares Citi Global TMT Conference

Yeah. I think, as you mentioned, we're relatively a small player in this industry of giants. We don't have really the ability to look on the WFE or have internal forecast. When we are talking about WFE, we're usually referring to the consensus in the market. We're relying on external sources. I think what is unique in 2026 is that every two months, we saw revisions of these numbers, and it's up as we go. It's driven mainly by AI, right? Even today that we're already just four months until the end of this year, there's still a wide range of estimates for 2026, right? You have Gartner talking about 20% growth, TechInsights in the high 20s. You have other analysts talking about 30, 35, 38, even though it's four months left, right?

When you have the big players, the backlog is out there, the shipment plans are out there, and there is still a wide range of WFE estimates. For us, definitely the numbers moved up when we're looking on the WFE, it will be above 20%. The exact number, it's very tough for us to call for a specific one.

When we're looking at 2026, definitely there are many drivers that working. First is the advanced node, as you mentioned, both on the Gate-All-Around, but also a lot of spending happening on the 3 nm as well. DRAM, definitely we see it's a significant additional capacity expansions that are happening there, and advanced packaging. I think for us, the main driver that is not working yet is NAND. As you know, even though NAND is spending, it's mainly skewed towards upgrades. Unfortunately, Nova is not benefit from the upgrades, and we need greenfield capacity additions. It doesn't happening yet. Hopefully, we will see some of it happening in 2027. This is, I would say, the longest down cycle on NAND side. Except NAND, I would say all the engines other than that working very strongly.

I think historically, if you look on over performance, you see that we are capable of outperforming WFE significantly. Just recent years, 2024, we grew by 30%, WFE didn't grow so much. 2025, we grew by 31%, WFE didn't grow so much. We are looking at 2026, and we talked about how significant this year is in terms of WFE spending, and we see a different, let's say, dynamic, and we see something change, and the question is why? I would say, the uniqueness of 2026 is definitely strong year, and Nova is growing nicely this year as well, but we don't see the same outperformance that we usually used to see from Nova, and the question is, why? I would say it's mainly driven by the investment mix. First, NAND, as I said before.

Even though NAND contributing to the WFE spending, Nova is not contributing for that because there is no capacity expansion. Second, a lot of investment happening in DRAM, and even though we are selling to the major players in the DRAM, the metrology intensity in DRAM is lower compared to the logic side. We talked about the fact that we are having our long-term model suggest 60/40 skewed towards logic. It's mainly because on the logic side, the metrology intensity is higher. Third, and this is mainly the most important one on the advanced nodes on the logic side, usually the investment pattern is happening when the customer is moving to new tech node, and after that, you have a high volume manufacturing. The customer transition to the newer tech node, and then there are a lot of capacity expansions there.

What we're seeing this year, given the fact that the AI customers didn't transition to the Gate-All-Around to the newer tech node, and they are still using the N-1, 3 nm and 5 nm, and obviously the huge demand that we're seeing there, the customers spending a lot of capacity expansions in the 3 nm. Both in U.S. and in Taiwan. So in a way, when you're looking on 2026 in terms of spending weight, you see that the metrology intensity went down because of the spending, I would say, mix.

Mix. Now the question is what's happening going forward.

How 2027 is going to look like. I think first, let's look on NAND. I think hopefully we will see NAND increasing the capacity. So practically for us, it's coming from zero to business because we are positioned with all the key customers there. Second, I would say on the DRAM side, we are not expecting significant change. DRAM will be a significant driver and we're going to enjoy from capacity expansions, but we are not expecting any significant metrology intensity increase. It will happen when the industry will move to the 4F², but it will happen in 2028 and going forward. I would say on the logic side, we do expect much more mixed queue towards the 2 nm and the Gate-All-Around, and less on the 3 nm.

I would say when we are looking on 2027, we will see much more metrology intensity going up in terms of the mix. Given the fact that Nova has its significant growth drivers to continue to outperform, I think we are well-positioned. We have the right drivers continue to outperform the industry going forward.