KLA Corporation (KLAC) Earnings Call Transcript & Summary

Well, good morning, everyone. This is C.J. Muse with Evercore ISI. Thank you for joining us with our Automotive and AI virtual Forum. Very pleased to have KLA's Executive Vice President of the company's Electronics, Packaging and Components division, Oreste Donzella with us today. Oreste has been with the company for 23 years and has been EVP of the EPC division for over 3 years, where he's now responsible for KLA's ICOS, SPTS Technologies and Orbotech organizations. I've known Oreste for years, and he's a phenomenal speaker, know KLA's business inside and out and very excited to talk about automotive today. Oreste has a few prepared slides that he'll go through, and then we'll do the Q&A. And there's a chat box for you. And if you'd like to put a question in for me to ask, please do so. And if you can't figure that out, please e-mail me directly and I'll look at that periodically. And look forward to the next 45 minutes of really digging into KLA's automotive business, which I don't think many investors have spent much time focusing on and I think we'll be excited after learning more about it. So with that, I'll turn it over to you, Oreste, and welcome.

Thank you, C.J., for the kind introduction and for the opportunity to showcase KLA initiatives in automotive that as they -- I started these initiatives with my team in 2017, working with the broader automotive ecosystem, including carmakers, Tier 1, Tier 2, fabless. And eventually, we come up with very differentiated solutions that are now paying off in terms of revenue and the future prospect of growth. Good morning, everyone, and I'll dig into it. So before going to the automotive initiatives, just a summary page about KLA. These are 2022 numbers. We closed calendar year '22 at $10.5 billion. We are headquartered in Milpitas, California. We are globally located in 19 locations. We have almost 15,000 employees right now. And of course, we are proud of our innovative spirit that there's a lot of PhDs and master degree engineers in the organization to invent and innovate a new solution. So the presence of KLA is very broad in the electronics manufacturing ecosystem. Of course, the KLA, almost all the investors know KLA is a world-leading semiconductor process control company. And in particular, in the left side of this page, our, of course, histories of leadership in inspection and metrology solutions for wafer front-end fabrication. However, during the last few years, through acquisitions, through some organic development of some of the solutions that we are able to show right now, we also made some good inroads in other segments of the electronics ecosystem, in particular, wafer-level packaging, components, PCB and even flat panel display. But today, I'm going to talk about automotive in details because this is the core of the presentation, the core of the conference that Evercore ISI is organizing right now. So first of all, in an automotive electronics. So in 2017, when we started initiatives of automotive, we thought about, okay, so automotive is in front of a secular shift. So we were talking about, of course, race like the autonomous connectivity, electrification and the venture, ride-sharing. I remember at the time where they -- okay, these 4 pillars of the future of the automotive industry we used to call race. So we started to participate as KLA to participate in many, many associations. And most of the associations didn't have any presence of semiconductor [ capital ] keep my company. So we were the only one. And now 6, 7 years after the introduction in the world of automotive industry and with a pretty massive pandemic in between, so we learned quite a bit more of what the challenges of automotives are today and they will become moving forward. First of all, we went through a massive chip shortage that eventually changed the way how the automotive -- the automakers are approaching the semiconductor industry. So we see more and more involved in first persona, in all the discussion about qualification of the chips, qualification and design of the new products, introduction of new materials. I'm going to talk about silicon carbide a lot later. But again, the chip shortage really made everybody aware of the importance of semiconductor, not only to advance the technology in the cars but also to provide and to manage the supply chain. And then, of course, electrification increased quite a bit. As you know, there is a pretty big inflection point right now with the EV taking a very serious lead in terms of growth, of course, relative growth, not absolute growth, in particular, in everywhere in the world. So we see electrification moving forward, growing much faster than the normal ICE cars. And with electrification, comes a huge increase in semiconductor content, and also, somehow, complexity, in particular, the subs that are used for electrical vehicles. Then of course, we have driver assistance, and driver assistance, is also leading a new trend in automotive that in automotive semiconductor. That is the introduction of more and more advanced semiconductor devices. As you may know, automotive industry is historically a very conservative industry in adopting new stuff because you need to have maturity in the device that are going into the car for reliability reasons. And for that kind of reason, generally, automotive, in the past, automotive has been lagging 2 or maybe even 3 technology nodes, the mobile, the PC, the lot of the data center market. But now with the rise of more autonomous features and connectivity and networking, we see like 5-nanometer semiconductor front-end technology getting in the car. And of course, this advanced technology is not as mature, as for example, a 65- or 90-nanometer [indiscernible] that poses new questions about reliability and process control. So this is an extremely, extremely important point. And the fourth thing that now I hear more and more in particular from the carmakers is, and we heard that many of those in the last few earning calls, capitalization from software, even Tesla. So I mean, there is a possibility that more and more, the car will become really an AI computer. And the real monetization in the car industry is coming from software and the applications are going to generate in the car while you are driving and collecting data. So really, the automotive industry is being transformed, and as I said, driving a very fast growth for semiconductor. You can see strong growth, assuming forecast, considering that there is a possibility that the other semi revenue will cross the $100 billion in 2028. Also another important point is the pie chart in the middle because you can see the variety of device types that are getting in the cars. So we have everything from power, of course, that is the #1 product that is getting in the electrical vehicles. From semiconductor point of view, you see microcontroller but also now an increasing percentage of memory, increasing percentage of sensor, MEMS, analog, RF, processors that are becoming more and more critical, in particular, going into more connectivity, autonomous features. And also, there are, as I said, many, many design rules now. And the design rules of the semiconductor devices that are getting into the cars will increment the capacity that is needed by the automotive industry, in particular in terms of new fabs. And there you see multiple new fabs have been planned by 2027 across the board, across a very wide spectrum of technology nodes. And as I said also, the automotive fab profile is different. In the past, we had automotive semiconductor only focused on the left side of this page, that is mature design rule fabs. So we are talking about, in the past, 150 millimeter, 200 millimeter. And now we see more and more adoption of 200 millimeter, also starting 300 millimeter, in particular, microcontroller, silicon power. And in the middle part of this slide, you can see the other 1 device in the vehicles that is power semiconductor. And also power semiconductor fabs are going through massive changes because on IGBT that is based on silicon, we are seeing the transition from 200 millimeter to 300 millimeter. And of course, the big inflection point for electrical vehicles is the introduction of silicon carbide for EV engines and gallium nitride, in particular, for battery chargers and so on. And finally, as I already said in the previous slide, we have the introduction of the advanced logic and memory inside cars, and of course, putting more and more stress on reliability and build due to the maturity of the technology node that it is way more advanced than the ones that have been used in the automotive industry for a long time. So really, everything is changing in automotive. So we have the fab profile changing, the introduction of advanced technology, the power semiconductor that is now really going through a secular shift of wafer size or new substrate with a wide band compound semi. And this is a huge opportunity for KLA that, again, we preannounced like 6, 7 years ago, and finally it is materializing right now. Now let me talk a little bit at a glance of automotive KLA. So first of all, this is the KLA revenue in automotive. This is, of course, relatively scale. You can see 2017, 2018, '19, '20. Since 2019, of course, we had the special semiconductor process, that is the dark blue in the bar chart on the left side of this slide. So you see like the light blue is really what KLA process control in the front-end wafer fabrication is, like inspection metrology. In dark blue, you see the incremental revenue that has been generated by SPTS division within my organization, within my EPC in the U.K. So overall, 37% out of system revenue CAGR. I don't have service revenue in this chart. You see 11% out of service revenue CAGR, not only system revenue actually, also the service revenue associated to our tools in automotive semiconductor fabs is increasing. And the reason is, first of all, because of the incredible shortage of chips that put all these automotive fabs really under incredible pressure to have the fab operating 24/7, all the tools extremely reliable, available and so on. So service is getting more and more demand of service contract, maintenance and parts because of the capacity needs in automotive industry in automotive semiconductor fabs. The second reason is, of course, with the advancement of technology nodes inside the automotive fabs, we see more and more advanced technology, advanced inspection and metrology products to be used. That means also the service revenues increasing as a percentage of the system price of the tool. You can see we were able to triple -- more than triple the CAGR of the auto semiconductor revenue. In terms of power, silicon versus silicon carbide, and this is an extremely prouder number from my side, so we were able to achieve $300 million, more than $300 million in silicon carbide revenue in 2022. And 7 years ago, when we started -- 6 years ago when we started this initially, it was pretty much 0. So we were able to go like from near $0 to $300 million-plus in a few years. And this is thanks to all the differential solutions that we have developed in these years, targeting this particular market of power devices. And in particular, on SPTS, that again is the dark blue in the bar chart on the left side of this chart, you can see that more than 35% of the revenue of SPTS came from automotive sector. So it's very exposed to automotive. And when I say what were the drivers, M&A, of course, because SPTS was an M&A that increased quite a bit our presence in automotive, increasing process control intensity because of the more and more focus on reliability, quality and yield in these fabs, and of course, introduction of silicon carbide. So there are 2 drivers that I would like to dig a little bit more in detail. So one is the enablement of the zero defect policy that has been known in the automotive industries for a long time but now it's really paramount, is really the zero defect or whatever we call, minimizing the escape or at the functional and electrical test because, of course, the biggest driver behind the zero defect policies, the automakers want to absolutely minimize down to 0 the possibility that the defect coming from a wafer is going to escape every single functional test, electrical test, final test, burning that the fabs are doing either before or after simulation. And then eventually, this defect becomes a so-called latent failure. It gets into automotive supply chain, eventually in a car, driving huge and expensive recall that the automakers don't really want to handle. So that's the reason why they are pushing more and more this philosophy of zero defect to make sure they don't get really exposed to these very, very expensive recall and eventually reliability issue. So this is what [indiscernible] reliability test escapes. So as you know, the semiconductor manufacturing is a fully automated process that has multiple checks in terms of in-line inspection and metrology, generally are done on sampling base. And then every single wafer, everything single chip is tested electrically. And sometimes, there are escapes at test because the test is not infallible. The test can have some escapes, and this is what the automakers are really worried about. The second 1 is even trickier. So this is what we call later reliability defect. So a defect that eventually can pass all the tests can get in the car and because of the harsh environmental conditions that a car is subject to in terms of temperature, humidity, pressure, this defect becomes killer or it's a latent because it was almost hidden, dormant and then eventually because of the stress that the chip is submitted in a car, started pretty much putting -- making this defect like a killer defect, a reliability defect and 1 is already inside a vehicle. So these are the 2 big nightmares of the carmakers. This is what they want to avoid. And when they called KLA in 2016, '17, we started thinking about talk about what do we have to do with automotive? I mean, we sell tools. Then we figured that and say, well, first of all, we can make our tools more customized for the automotive semiconductor industry. So we got some tools like, for example, what we call Broadband Plasma imaging tool and unpatterned wafer inspection and the macro with inspection tools, and we customized, we tailored to the need for automotive, semiconductor, in general, they want to have less escape, be less immune to noise, find defects that are more critical in the certain part of the die or the wafer. And then we also developed, invented, developed, patented and released the so-called I-PAT. That really is the core of this reliability screening strategy that I'm going to focus in the next couple of slides. So what is I-PAT? When we responded to this automotive requests about, okay, I really need to limit, to neutralize, minimize all these potential latent reliability defects or test escapes. And then we thought about and say, okay, so we have inspection in the wafer fabs. We have metrology. Unfortunately, we don't do inspection 100% of the wafer, 100% of process steps. It would be extremely expensive for the wafer fabs. So there is a set of sampling. And but we have data. Now we don't have access to test the data. We don't have access to reliability data because, of course, this is [indiscernible] our customers. So we had an idea to aggregate all this information from KLA, from other inspection tools, other metrology tools, from tester, from prober, from reliability, from failure analysis, and really develop a neural network model based on AI that give the perception, give pretty much the guidance to the wafer fabs, which particular process step, which particular defect, which particular die is more and part of the die circling the die is more exposed to potential latent defects because sometimes, the defects, the same kind of defects that are generating loss can be latent until they get into the car. So what we may know based on AI, which circuit or which part of the die or which part of the process step or process flow is more exposed to this? So we went to the automakers, went to the Tier 1, and they pretty much, of course, signed an incredibly restricted non-disclosure agreement, a firewall in the fab. We get access to this data. So we aggregated all the data and eventually, we come up with some neural network model that is predicting which kind of defect is more exposed to the possibility to become a reliability latent failures when it enters the automotive supply chain. And this is what we have done with I-PAT. And just to give you an idea about I-PAT, I-PAT is now written in the spec of the Automotive Electronic Council. KLA is a member of the AEC that aggregates pretty much all the most relevant players in the automotive semiconductor ecosystem. And AEC mentioned suggesting the automotive semiconductor fabs to adopt I-PAT methodology. And we also have 1 automaker that put in their electronic specs, I-PAT is the recommended methodology for screening chips in the wafer fabs. So this is what I-PAT is. It is a deep learning, machine learning artificial intelligence methodology applied to the massive quantity of data that is generated in the fab by using, in particular, inspection results from KLA, we are adopting right now the I-PAT on the 89xx platform, that is our macro inspection tool but we are extending to other platforms of inspection in KLA. And also in the software that we have is analytical hub in every single wafer fabs that is called Klarity. So the combination of the [indiscernible] from our inspection tool and the software that is already in every single -- almost every single wafer fab, together with the artificial intelligence algorithms, are giving now the possibility of screening potential reliability failure in fab when the defect occurs instead of waiting for the defect to enter the automotive supply chain. So this is one. The second trend that I would like to talk is silicon carbide. And as you know, there is a huge shift right now for the electrical vehicles from silicon power to a wide-bandgap wide compound semiconductor like silicon carbide. And the reason is, of course, power efficiency, is the possibility to control thermal conductivity better. So these substrates, either silicon carbide or gallium nitride, although silicon carbide is the perfect substrate for electrical vehicle engines is -- are really pretty much overtaking silicon in the future. However, there is an issue. The silicon carbide has significant cost issues. And the cost issues are becoming -- are coming from poor yield, and poor yield that starts from the subset. So what we call better silicon wafer that is an incredible yield is very, very clean. When you want to build power devices built on silicon carbide substrate, you just started with like 60% yield because 40% is completely lost on the substrate without even putting your device on. And this is a huge limiting factor right now. For example, to scale beyond the 6-inch, 150 millimeter. So you can see there are also other areas of yield losses, not only substrate, but also wafer yield fab, probe yield and so on, packaging as well. So at the end of the day, we have a great opportunity in silicon carbide, either in process and inspection and metrology process with SPTS division, of course, inspection and metrology with the same EPC organization to be a relevant player, to help our customers to identify and eventually solve the most critical defects in inspection and metrology and with process to design more robust, more stable deposition and etching process that can stand this yield in a much better way. So that's the reason why the combination of SPTS in process and semi EPC, inspection and metrology tools are making an incredible inroads in the silicon carbide power fabrication right now. It gave us a nice boost of revenue, as I said, in 2022 and moving forward. As I said, this really the solution are central to the silicon carbide cost reduction. And everybody now, the biggest focus for silicon carbide manufacturers is cost. That cost, again, is derived from a low yield and we can now, across the board, improving the yield, migrate the technology, innovating in the methodology for substrate fabrication, and of course, also driving what is going to be believed in the next couple of years that is the transition from 150-millimeter to 200-millimeter silicon carbide power devices. Now in summary, I would like to, first of all, to explain also another big milestone of our automotive initiative. A couple of weeks ago, we announced the so-called STAR, that is Semiconductor Talent Automotive Research is a new consortium that is based in Ann Arbor in Michigan, a few miles from Detroit, from Motor City. And this is going to be a collaboration between KLA, IMEC and few other entities that I'm going to present in the next slide, just to advance 2 things. First of all, to make sure that we develop a tight line of engineering talent in automotive and semiconductor in the region, in all Midwest of the United States. And second thing is advancing the technology roadmap for automotive and not only stocking wafer fab but also including eventually packaging integration. So there was an MOU sign-off with Michigan Governor, our CEO, Rick Wallace, the CEO of IMEC, Luc, and then there was participation, of course, from University of Michigan, from Washtenaw Community College, the city of Michigan was very well present here. And you can see on the right side, also General Motors is part of the consortium as well. So MEDC is also the Michigan Economic Development Center. So again, we are getting more and more inside the automotive ecosystem also with real initiatives that the automotive industry can benefit in the future. In summary, of course, a huge, huge secular shift in the automotive industry, in particular, relative to electronics and semiconductor. We worked very, very closely with automotive ecosystem. We developed absolutely customized product and software solution. And the future is bright. We have a possibility to make even a better impact with silicon carbide introduction and, of course, the advancement of technology nodes inside the vehicles. With that, I'll stop here and open for Q&A.

Perfect. Thank you, Oreste. Great overview. I'm sure everyone is appreciative of that. So I was hoping to hit on a few things. So the first 1 was to really help investors understand kind of the size today and growth going forward. So I think roughly $700 million in total auto revenues in 2022 for KLA, so that's about 7% of revenues. Is there a way to kind of unpack that in terms of silicon carbide, process control? However you want to kind of slice and dice it, what's the best way to kind of think about that?

Yes. I showed the revenue portion in one of my slides in the left side, and I split the process control and process in that slide. I would say process control right now is maybe process is 30% and process control is 70% of the entire automotive revenue that we generated in 2022. And we are generating in 2023, I think the percentage will said the same, like 70-30, 75-25. So mostly is, of course, in the semiconductor process control, where we have a huge presence in the market with many, many products. But SPTS growth is incredible as well, as you saw from the previous slide. When you split the silicon carbide and silicon, $300 million was in silicon carbide and the $400 million was in silicon. So you see pretty fast growth in silicon carbide. I think as I said, a few years ago, silicon carbide was 0. So and we see silicon carbide getting the fastest potential growth in the market. So this is why I would like to split. I would say 70-30 process control, process, process means SPTS, and I would say, 60-40 right now silicone versus silicon carbide.

Perfect. And from a growth perspective, I think over the last 7 years, including your outlook for 2023, you're talking about a growth CAGR for you of about 30% versus auto semis, I don't know. Hard to call precisely but definitely, your number far better than that. How are you thinking about kind of the growth rate going forward? Obviously not asking for a specific guide, but are you expecting that kind of continued outperformance? And what would be the key drivers underlying that outperformance?

Yes. So good question. First of all, I hear as well 9% automotive semi revenue increase. I believe this estimate is a little bit conservative. I'll tell you why. Because it doesn't account that automotive industry is getting also the leading-edge technology. And whenever you have a leading edge technology, let's assume it's 5-nanometer and 7-nanometer, maybe 3-nanometer not now, but 5-nanometer is already in the car. Sometimes, you don't qualify this technology, this output as automotive. So sometimes, it gets into the 5-nanometer pile that goes somewhere else. But I believe that 9% is a slightly conservative estimate for the semiconductor revenue associated to automotive. Now as regards KLA, I believe we will outperform this number and we will outperform the broad equipment spending in the automotive semiconductor. And the reason is -- and there are multiple reasons. First of all, as I said, the silicon carbide is requiring completely different set of process and process control tools. And we are very much in the silicon carbide process with SPTS. And as I mentioned in silicon carbide, we are retooling real products to make sure that we can satisfy the needs of this market. The second thing, as I said, the advanced semiconductor. I mean, again, it's impossible to think about having a 5-nanometer, 7-nanometer technology in the car and the carmakers not pushing for more I-PAT, for example. So because they have been used to N minus 2, N minus 3 for a long time, now they get the end product, end technology node, and they will be extremely, extremely anxious with the potential reliability failures. So there will be a process control intensity for these nodes in automotive industry as well. So that is why I believe that again, the plus 9% for semiconductor automotive revenue can be slightly underestimated. I believe it's going to be more than that because of the end drivers, and KLA will outperform the broad market because of these reasons I said.

Fantastic. So relative, I guess, to process control, you have a very kind of broad portfolio, pattern inspection for R&D, separate tool for HVM. You've got the unpatterned Surfscan. You also talked on the call around I-PAT. And so I guess kind of as you think of that portfolio and as you think about your access to utilize kind of internal OEM fab data, how, I guess, are you thinking about process control intensity over time for the automotive world?

Yes. Let me specify. The OEM or fab data are used to develop the machine learning model that is the base of I-PAT. We don't have access to this data outside the firewall. So these are extremely protected data that the automaker says, and they'd be used only for the purpose of developing neural network model for I-PAT. So we don't have full access that we know all this data. This is not true. It's only firewalled for a specific reason to very, very few people. But the collaboration with our OEM and the fact that we have I-PAT now have written in the [ AC ] specs, give us the opportunity to push more and more for process control. And -- and the process control, again, is twofold benefits for KLA. On one hand, there is a possibility to increase the process control intensity because people are adopting more and more inspection and metrology. On the -- and the second benefit is with, as I said, the introduction of more advanced technology or exotic [indiscernible] silicon carbide, the people are using more advanced process control tools. So that, of course, gives us a boost in the ASP because these are newer tools and of course, higher ASP, more complex technology. So we have really -- I would say, it's not the fact that we have access to the data from auto OEM or Tier 1 of data because, again, we don't have full access to this data. Is in the collaboration and the solution that we provide to the industry, that we see the possibility to increase the process control intensity and move to next-generation inspection metrology tools in automotive semiconductor fabs.

The quality is obviously a priority, in particular, in the automotive industry. And you talked about working closely with OEMs. Curious if that's something where your work there is causing the OEMs to push on the manufacturers to adopt more process control or your conversations with your direct customers that they understand that they needed as well. How do you kind of think about that process?

That's very good. I would say both C.J. On one hand, of course, by directly working with some of the auto makers of Tier 1, of course, we let them understand the importance of process control and what we can offer in this field. And then they push their suppliers, of course, to adopt more and more process control. And at the same time, we go directly to our customers, their suppliers to promote the value of solutions like I-PAT or solutions like C205, BPP patterned inspection or the new wafer for automotive or 89xx. So we do both. We do both. We talk to the OEM and Tier 1 to promote the value, and they understand they push their supply chain in the same way we partner and we are direct customers to make sure we can provide the value they need.