Nova Ltd. (NVMI) Earnings Call Transcript & Summary

Very good afternoon. Welcome, everyone. I'm really delighted to have the team from Nova Measurement join us this afternoon, Dror David, the current CFO and really lucky to have Guy Kizner also corporate -- sorry Vice President of Corporate Finance and the incoming CFO as of July 1, so you have just a few more weeks to relax, and then you are in the hot seat, so to say. We're really happy that you can join as well.

So maybe, Dror, just to kind of bring people who are on the webcast and in the room up to speed, could you maybe just take a high-level perspective of Nova, describe what are your areas of strategic focus and then importantly, as the company is so well levered to AI, which specific areas are you playing in and how levered are you to that big growth driver?

Sure. So obviously, we are operating in the semiconductor process control segment, selling control equipment to the semiconductor industry. Actually, I think one of the elements of Nova story is its wide exposure across the different segments of the industry. The company is exposed to memory, leading edge, trailing edge, also to advanced packaging. So across the segments, practically, there is no single fab being built today without Nova equipment. In terms of the AI, obviously, I would say that in all elements related to semiconductors almost, we are exposed. We are exposed to HPM, to advanced packaging, hybrid bonding, and obviously, the high-end chips, which are manufactured in the front end to eventually get into the AI chiplets and applications in all these fabs, we are present.

Got it. Now many in the U.S. right now of the large competitor, right, KLA in this market, how does Nova differentiate itself from KLA?

Yes. So 25 years ago or 30 years ago when the company was established, it was established around the optical CD arena and technology. This is optical instrumentation, which is doing high-end measurements of the specific transistors. And we were competing with KLA since the beginning of this process. Actually, since then, the company has developed or acquired many different technologies. And today, actually between around 40% of our business is not competing with KLA. These are mainly the materials metrology tools. Within the segments that we do compete with KLA, which is the optical CD, it's around 40%, 50% of our business. We do have unique capabilities also in this arena related to technology developments that we have. The latest one is the Nova Prism, which has a unique capability that enables to measure deep Vias and deep structures. And we can discuss exactly what it means.

Exactly. Okay. So first, in terms of the broader spending environment, I think that at the start of the year, you described the wafer fab equipment, WFE, market this year kind of growing mid-single digit. That's what a lot of your peers right have suggested also. Now that we are roughly kind of halfway through the year, is that still a reasonable way to think about just the overall WFE environment? And if you could help unpack right, whether it's foundry logic or memory as part of that, how are you seeing it today?

So the answer is yes. We still see the same, I think, level of investments on an annual basis from where we are right now. In terms of the different segments and their impact on the market, obviously, trailing edge in China was very strong in the first half. Actually, we do believe that there will be some moderation from the high level of H1 into H2 of China investments, not a significant reduction, but some moderation. On the other hand, we have other vectors, which should positively impact the investments in the second half. One that I can mention is, for example, TSMC, which invested around $10 billion in the first half of the year of '24, and their plan for the year is around $30 billion. So we will see some additions in the second half. In terms of memory, we do see investments mainly in HPM applications, the VNAND and the DDR5 regular DRAM expansions have not yet started, some will start at the second half. And in terms of leading-edge gate-all-around investment, this is more a '25 plan. So overall, still mid-single digit in 2024. And obviously, Nova, with its exposure to the different elements and the new technologies, we expect to significantly outperform this mid-single-digit level.

Got it. That was actually my next question that Nova has been able to consistently outgrow the WFE right environment. And I assume you expect that to happen this year. How is typically your visibility? Does it extend 1 quarter? Does it extend 2 quarters? How is the visibility in the second half of the year?

So visibility is actually quite good right now in terms of backlog, bookings, also into the second half of the year. It's very important to note the numbers because when you look at the last 5 years, the industry grew on average wafer fab equipment around 10%, 12% a year. In the last 5 years, Nova grew approximately 20%. So we actually significantly outperformed the market around 50%. We expect to do the same also in 2024. And the reason for that is the fact that we are expanding our addressable market, entering new applications within the semiconductor that were not in our revenue stream before such as advanced packaging HPM and also, we have a very wide presence across the foundry and the memory segment. So we do expect to outperform in '24. Obviously, '25, '26, a new story of investments in the U.S., gate-all-around, which we have a very good position. Probably, we'll discuss that later on.

Right. Just one thing about the next few years. So there's a lot of interest in the different Chips Act, right, that are being implemented, whether it's in the U.S. or there are versions of that in Europe and China. And I think China this weekend said what 45 plus -- [ 47 ] plus. Have you already started to see some benefits of that reshoring? Or you think that is still to come as we look over the next 1 to 2 years?

So most of the investments in areas outside China are yet to happen in the coming years. The best example is, for example, TSMC investments in Arizona. Most of that will be '25 and forward. Some of that already started. Intel, definitely mainly '25, '26. Samsung Taylor fab in the U.S., the same '25 probably and not this year. So most of these investments are going to happen in the coming years and probably drive another step-up in WFE.

I see. Do you think that increases WFE intensity in that -- now that you have a lot more new fab construction, does that increase -- does it step up WFE intensity versus what the trend line would have been otherwise?

Yes, that's a good point because we all know there is an end in terms of the intensity, right? If intensity is too high, it becomes unbearable for the manufacturers, and it becomes a problem. I think that what has changed in the recent years is the profit, I would say, profile of the semiconductor manufacturers. And if you take WFE relative to their profitability and not necessarily relative to the revenues, you would see that the trend has not really changed significantly, and it's still on a reasonable level relative to profitability. Although relative to revenues, it might increase in the coming years, which is a good thing.

I see. That's right. Success-based spending rather than just building it for its own sake. How much is trailing edge exposure for Nova right now? And where do you see it heading this year and next year?

So trailing edge currently is mainly China. We are not disclosing our allocation in these areas on a quarterly basis. Last year, trailing edge was around 50% on an annual basis. I think overall, trailing edge at the beginning of the year was slightly higher because of the major investments in China which most of them are trailing edge. However, on an annual basis, with the expectation for more investments in leading edge, the second half should be probably, hopefully, the same as last year, maybe a little bit higher.

Okay. And your broader assumption about China is that spending is flattish this year, plus/minus?

Actually, dollar-wise, China level would probably increase, yes, even up of high level of last year. Obviously, they are spending a lot. They have this new fab, $50 billion, which was just announced, and we see the orders, we see the projects. So I think dollar-wise, it will be a little bit higher than last year, also because they have started to invest also in DRAM capacity in parallel to foundry. So that's another function of increase. But as a percent of revenues, this year, it will be probably the same as last year. And next year, it will start to reduce. We do have visibility to many 2025 projects. So we do expect China to remain stable, at least in the foreseeable future from where we are right now.

All right. One thing we hear from your kind of end customers, right, is just this supply shortages that they are seeing in packaging, right? Especially CoWoS, right? And that's out there. I think for Nova, last year, Advanced packaging was about 10% or so of sales. This year, I think the outlook is what -- for it to grow 50% right or so, but is that a conservative view because when we look at what's happening to accelerator demand, right, that's like more than doubling this year. Obviously, there are timing differences between when they sell something versus when you sell something. But do you think there is upside potential to that market outlook?

For advanced packaging?

Yes.

There is some upside because a lot is happening in advanced packaging. And I think the market is eager for new developments and new capabilities. But anyway, 50% growth relative to WFE growing 5%, it's still significant. So anything at that range is probably where we will be by the end of the year.

Okay. And how does the visibility look in that business going into next year? Because we are all making assumptions about TSMCs packaging, right? There's other CoWoS players, right, coming up. So are you seeing, number one, the [indiscernible] broaden? And number two, how is the visibility looking like as you look into next year?

So two elements are happening for Nova. First of all, the visibility for investments in advanced packaging is quite good, I would say. The projects are clear. All the customers are doing actions to take their market share. We just heard the Micron announcement regarding HPM. By the way, they did not really start to invest, so they are just doing whatever they can in order to move some capacity to this direction, but they have not really started to invest and this will happen in the coming quarters. So on the, I think, visibility side, we are okay. And I think one of the major elements for us is that we are doing a lot of qualifications right now for advanced packaging applications with the big customers, which once they will qualify and we'll move to the next steps, we are talking about HPM core and others in many applications by Intel, by CoWoS. These qualifications will translate to more business and not just a sustainable business in the coming years from advanced packaging for Nova.

Okay. And how do you see this mix between chemical and optical as you look at how customers are making these investments?

For advanced packaging?

Yes.

So today, I would say that 10% approximately of the 15 is chemical and the rest is optical. And I think that given the qualifications that we are at and the new applications that we see for optical CD, the opportunity there might be even higher than growth in the coming years than for this chemical metrology.

On memory, let's go through the DRAM side first. What we hear from the memory companies is that they are transitioning more of their capacity towards HPM, they're not really adding a lot more wafers per se. Is that what you are seeing also? So how does it benefit you this transition versus them adding more wafers for DRAM?

So first of all, this is the case that we see. We just discussed Micron as well. So you would expect that they would invest heavily right now in HPM, but this is not necessarily the case. They are just transitioning capacity. The issue is that through this transition, in HPM related DRAM chip relative to a regular DRAM chip, there are other applications and new applications and new structures that you need to do in order for it to be ready for HPM stacking. This includes TSV, this includes flatness of the wafer which needs to be at a different level. And these applications are the ones that we are selling the optical CD tools into. Therefore, despite the fact that it's only migration of capacity, we do see additional business in these areas.

All right. What -- obviously, with HPM, right, it requires like almost 3x the wafers versus DRAM. So do you think that is helping to kind of help them eat up into the existing supply fast enough? When do you -- let me ask a question this way. When do you think the memory industry will be ready to add new capacity to DRAM? Is it this year? Is it next year?

This year. In the second half of the year, we should see starting to build regular DRAM and expand regular DRAM investments. How fast, how much we fall into this year is yet to be seen. The challenge with the memory industry that it's highly sensitive to what's going on in the market. Therefore, the cycle of investment is actually very fast. The lead time of the equipment and everything related to it is very, very fast in terms of turnaround, but we are ready. I can tell that we are ready.

Okay. And what is the -- is there a simple way to look at how your content and opportunity changes as the mix moves more to HPM over time?

In terms of quantification of dollar value, I'm not sure we are there yet. But again, this 5% of business, which is going into HPM is -- most of it is related to HPM.

Okay. Have you seen a pickup on NAND yet?

No. Actually, no activity, special activity in NAND. We actually do not see also clear plans from the customers in terms of this -- of the next step of investment in NAND. It seems like this will be more of a '25, '26 play. And on one hand, we are waiting for this. On the other hand, it's good because '24 will be a great year for Nova. And then we had all these new factors coming in. So it will help us continue to grow.

Okay. Some of the NAND players have made this linkage to AI that AI servers have a lot more NAND also and that as enterprises start adopting AI, right, the demand for SSDs is going to grow. Do you think it's too early to start [ eating ] comfortable with those drivers that they are not big enough to matter quite yet.

I think it's too early. Yes, I agree.

Okay. Next thing in terms of the different metrology products ELIPSON, METRION and PRISM, right? Can you talk about the evaluation pipeline? Where are you? And is that, again, more a '25 or '26 growth driver?

Yes. So actually, what is happening right now is that the industry is very aggressive in terms of concluding the preparations for gate-all-around manufacturing. So we do expect that by the end of this year, most of the qualifications of the production tools and also the process control tools will to conclude. Obviously, we are in the midst of this qualification processes with all the customers, which are planned to do gate-all-around. So assuming we conclude this evaluation successfully, assuming the gate-all-around major investments are starting to happen next year, we will see a significant adoption cycle of these products that you mentioned. By the way, Nova PRISM is also being qualified for additional advanced packaging applications. So we are quite optimistic that the usage of these technologies Nova PRISM, ELIPSON and METRION will significantly increase next year -- starting next year.

Again, kind of a similar question to what I asked before. Is there a content lift for Nova? Is there greater opportunity right? Like if you go back to the transition from planar to FinFET right, as an example, how would you contrast what we might see in the move from FinFET to gate-all-around versus what you saw in the prior transition?

So actually, we are very excited about the next transition to gate-all-around. And we can look at it from several, I would say, directions or vectors. First of all, in the move to gate-all-around from FinFET, we expect to see 30% more metrology steps which is a larger -- significantly larger opportunity as a market for us in process control. On top of that, some of the products that we have technologies, which include Nova PRISM, METRION, ELIPSON, are going through the technology adoption cycle within this move. And these products are with high ASP relative to our existing product portfolio. On top of that, when you analyze the customers that are going to go through the gate-all-around transition, and there are four customers right now, the usual three suspects, plus one in Japan. In each one of these customers, we already know that we have a better position in terms of the number of technologies that each of these customers is adopting is higher than in FinFET. In some cases, it's a new customer, like the Japanese one. So the ripple effect of 30% more business, adoption of technologies and higher position is something we are excited about looking forward.

Got it. So you have already seen those gate-all-around orders this year? What percentage of the business is it today?

So first of all, yes, we just issued a press release, I think, 1 month ago, saying that we received, in parallel, orders from all these four customers. Obviously, these are not the capacity expansion volume orders that we should see in the coming years. But they are initial orders for pilot lines and obviously, R&D processes. In terms of how much is gate-all-around of our revenue, I can't disclose, obviously, because it's customer information. There is only one who is really investing in '24, but all of them are starting. We received orders that will be shipped in the next 12 months, over the course of the next 12 months. And again, it's a significant opportunity for Nova.

I see. The reason I asked that, Dror, is that we have had Lam say, they will have almost $1 billion, right, of sales. And I think applied -- I think, had actually more than twice of that. So is it that your opportunity comes a little later? Is it coincident? Like if you were to -- I know you haven't disclosed the absolute number, but if you look at and work backwards from what those companies are seeing, are you seeing that kind of similar proportional benefit for Nova also? Or is that still to come next year?

So the answer is yes. But most of our deliveries are in the first half of '25, which is probably one quarter after the big guys, the equipment is coming in. Normally, the process control tools are coming a little bit in a lag after the equipment, the process equipment itself. In terms of the booking itself, again, these bookings that we received were initial orders for the first phases. These are not the capacity expansions high volume we expect looking forward.

Not to go off on a tangent, but is there always a lag that process control sees it a little later? And where I'm going with that question is that all these Chips Act and all these reshoring and the new fabs that are being built, is process control an early beneficiary or is it like one step after some of these other front-end tools are...

It depends from which angle you look. So if you look from the angle of how much equipment they are buying in the first 6 months of running the fab, I think process control is beneficiary because you buy a little bit more process control to ramp up the capability and so forth. In terms of the timing, it's always a little bit afterwards. The way to think about it is on ASML, right? ASML is the first tool in the fab. You put it in, you make it work and only then you take all the rest of the equipment. So it's more or less in that context.

I see. Do you have any leverage to their adoption of the high NA EUV tools? I know it comes over a period of time. But now that they are saying that they will start to shift some to all of the leading edge, so is there some benefit to Nova also?

In the high NA, not necessarily because we are not that strong in the lithography process control. This is an area where more Applied and KLA are playing and obviously, ASML as well with their process control capability.

Okay. One of the trends we have seen is front-end tools being converted for back end. What is Nova doing in that business? And is that accretive overall to your business model?

Yes. So actually, we have been waiting for a long time for this process to happen because the technologies that we are using for front end or the customers are using for front end are very advanced, relatively high ASP relative to the normal metrology which you use in the packaging and advanced packaging. So we have been waiting for a long time for the evolution of advanced packaging to be in a way that will require some front-end equipment. And that's a good point because 3 years ago, Nova had almost 0 revenues in advanced packaging. And today, it's 15% also because of the adoption of front-end optical CD tools, mainly Nova PRISM integrated into the advanced packaging arena. So we are already enjoying that. We expect that to increase with more applications and more business. And at the end of the day, we're using more or less the same equipment that we sell for front-end or these back-end processes, meaning the pricing is good, gross margins are good. Normally, it's the same customers buying practically the same equipment for front-end and back-end. So definitely, we enjoy that, including pricing and good gross margins.

Okay. I had one other question on HPM. We have seen a number of these GPU companies now spell out that they're going to accelerate their product cadence, right, to 1 year from what used to be 2 years and a part of that seems to be the move towards better and faster HPM, right, every year from 8 high to 12 high stacks, HPM 4 after that. How does your opportunity change as they move to these denser HPM cubes and then they move to HPM 4?

So the way it works -- and we have learned that in the VNAND arena, right, where they are doing stacks of not chips but structures, is that as much as you add more stacks, you need the accuracy of the lower-level stack and the whole structure to be better. If you don't improve the accuracy and so forth, it just doesn't work. In that context, what we expect is as the industry is stacking more chips on top of the other as part of the HPM evolution, the requirements and the accuracy requirements from each chip will become higher. And this is obviously a good environment for our dimensional and material measurement capabilities.

Okay. Two questions on financials. One is last year, revenues were a little bit over $0.5 billion or so, but Nova has this plan laid out for the March to $1 billion or so. How confident are you about that plan? When do you think you'll have line of sight to those kind of revenue levels? And how much of that is organic versus inorganic?

So you are correct that this plan is aggressive. Our plan is to move for $0.5 billion in '23 to $1 billion in '27. This is 4 years doubling the revenues in the semiconductor arena. So that's one thing. It is an aggressive plan. We believe we have the building blocks to do it, but we need to execute in that context. I think the good thing is that when we look at the different building blocks that we need in order to execute this plan, we have two elements. One is the technologies and the product portfolio and -- sorry, three and second is the market and when will the market meet these technologies. And the third is obviously infrastructure, manufacturing facilities. Still, this is doubling the revenues of the company. And I think in all of these elements, things are already organized in the right manner. The ELIPSON, the METRION are in their final phases of entering the market. In terms of market meeting technology, we have these transitions of gate-all-around, which will add with the adoption of these technologies. And in terms of infrastructure by the end of this year, we will already have all the manufacturing capacity in order to meet this strategic plan. Our goal is to cross the $800 million revenue mark organically. And on top of that, add between $150 million to $200 million of M&A and execute the strategic plan by '27.

All right. On that optimistic note, thank you so much, Dror. Really appreciate it, and good luck in your next adventure. And now I think Guy has to execute on $1 billion plan.

Yes, he was part of planning it and strategizing it.

Excellent, now executing it.

Exactly. Thank you.

Thank you.