FormFactor, Inc. (FORM) Earnings Call Transcript & Summary

Good afternoon, everyone. My name is [ Elizabeth Stan ]. I'm a senior associate here at Citi. I'm working with my analyst, Atif Malik, on semiconductor equipment stocks. And here now, I have FormFactor's CEO, Mike Slessor with me this afternoon for a fireside chat. And I guess I'll start with some questions first, and then I'll open up to the audience to see if there are any in the audience. So first, Mike, thanks for joining us today for our tech conference.

Thanks for having us.

Thanks. And maybe we can start with a very high-level introduction of what FormFactor do for those who are not that familiar with your company and what's new in the past like 6 to 12 months?

Sure. So let's -- I'll just -- high-level view of FormFactor from our investor presentation, which is on our website. FormFactor, a leader in semiconductor and broader optoelectronics test and measurement. Trailing 12-month revenues, a little under USD 700 million. And that business splits primarily into 2 reportable segments. The first, the business for which FormFactor is probably best known, something called a probe card. This is the consumable that serves as the interface built by automated test equipment built by people like Advantest and Teradyne and our customers' chips on the wafer. So large customers, 10% customers recently have been people you see on the lower right here, like Intel, TSMC, Micron, SK Hynix and Samsung, essentially leaders in the semiconductor industry. And for some of our foundry customers, they're fabless customers as well, our direct and indirect customers. Probe cards comprise about 70% to 80% of our business, really geared towards enabling customers to ramp new designs on advanced technology nodes and achieve high yield so that they can meet their shipment targets. The balance of the business is the other segment that we report called engineering systems, and these are mostly capital equipment used very early in our customers' product development cycles. Again, similar customer makeup, similar customer composition. But instead of being used in production, these are used in early development as customers characterize devices broadly from brand-new transistor structures to some of the qubit structures for quantum computing. We've recently seen some great growth in this business from silicon photonics and co-packaged optics. So really enabling our customers next generation of development in the systems business. We're a little over 2,000 people worldwide. And as most leaders in this industry, we do have a global footprint, a rate throughout the world in supporting these businesses and our customers. I think I went through -- touched on some of this. The leftmost picture you see there is a probe card, and a probe card to test DRAM devices. In that business, essentially, we make contact with every chip on the customer wafer. Approximately 150,000 individual MEMS fabricated probes, all making precise contact with various points on the wafer and testing all of those chips at once. Some of the drivers that you've seen in our business really work even here today during a pretty significant semiconductor industry downturn. Advanced packaging is a very exciting opportunity for us, and that may be a little bit confusing. And though we're a test provider, so why would advanced packaging be a driver for us? It turns out that changes and drives up both the complexity of test and the intensity of test. So customers, when they break a chip into chiplets using these advanced packaging processes, need to test each of those chiplets much more comprehensively than they do the composite chip. So that extends lead times -- sorry, extends test times leading to higher test intensity, higher spend on probe cards, but it also raises the technical requirements for the probe card, higher speeds, higher temperatures, different kind of voltage and current relationships, so really raises the bar on the technical requirements for the probe card. That's one of the reasons, as we see in this middle bar, that we really take a significant amount of revenue and reinvested in R&D. Again, paired with these leading customers, we're really pointing that R&D budget and our priorities at solving some of their next-generation challenges like higher speeds and higher temperatures associated with testing. As we've gone through the downturn, we've maintained non-GAAP and, in most cases, GAAP profitability. So got a structural model. Even though we're fairly heavily vertically integrated, we've managed to adjust our cost structure and continue to generate cash as we've gone through the downturn. And our product and -- want to continue to be active in mergers and acquisitions. Yes, we want to organically execute on the businesses we lead and serve today, but we also believe that M&A is a very important part of the broader diversification and growth of the company. Again, around the themes of test and measurement, metrology and inspection. But M&A has been a part of our history, and I think, hopefully, we'll continue to be a part of the company's future. So I mean, with that, we'll stop there and see your questions, Elizabeth.

Yes. Thanks for the introduction. And I really like you highlighted the opportunities in advanced packaging, so we can probably get back here a little bit later on, a little bit deeper. So let's start with some broad market questions. So how do you look at underlying market growth and how FormFactor is positioned right now? And also like if you could provide some sense of like how much of your business is unit-driven versus like CapEx driven.

Yes. Well, I think there's different demand drivers in our different businesses. If I start with the probe card business, one of the important things to understand about this business, is it really is driven by customers releasing and ramping new designs in volume? So there's an element of new design release, but there's a number of units they're going to produce on that new design as well because that drives the number of test cells and probe cards that they're going to need. So it's a bit of a hybrid demand driver of new designs and the number of units on that design. As a consequence, you can often see in our results and our competitors' results and the overall probe card industry decoupling from WFE, decoupling from industry capital expenditures. And in some sense, we've been seeing that now. We'll come back, I think, to some of the segments. But we saw, going from Q1 to Q2 of this year, an approximately 50% increase in DRAM probe card revenues amidst one of the grimmest DRAM overall market situations we've seen in a long time. And the reason for that is customers releasing and ramping new designs and volume for a memory technology called HBM or High Bandwidth Memory. This is enabling a lot of the compute infrastructure associated with generative artificial intelligence. And at the risk of jumping on that bandwagon, it is a significant driver of our DRAM business and gives you some sense of how we're a little bit different than CapEx trends in the industry. To see DRAM business increased by 50% in the middle of this year, I think, is pretty unusual. Different drivers across the systems business, that's more driven by customer R&D spending, as you might imagine. And again, we're seeing some nice growth at present associated with silicon photonics, co-packaged optics. Some of these things that are just coming into initial pilot production. So the demand drivers there, again, are going to be different than WFE and CapEx, more driven by customer innovation and R&D investment.

Okay. That's encouraging to hear even especially in this WFE downturn. So you just touched a bit on DRAM. You said DRAM in the second quarter was like 50% growth and driven by HBM and DDR5. So how should we think about DRAM growth going forward like in the second half of the year or like even going into next year?

Yes. We do expect, and this was -- if you parse through our third quarter outlook that we gave in the August earnings call, we do expect DRAM demand to stay relatively stable at the levels of the second quarter as we come through the third quarter here. We haven't set expectations or given an outlook for the fourth quarter, primarily because one of the interesting things about the probe card business is you operate on very short lead times. Typical lead times are well within a quarter, think of an average of being 6, 7 or 8 weeks, primarily because you're essentially providing this device-specific consumable just in time for customers to test the wafers. And so your lead times are indexed very closely to wafer lead times in the fab. That comes with the challenge, obviously, a fairly limited backlog visibility. So we tend to only provide an outlook for the current quarter. But as Elizabeth notes, we did indicate that we expected DRAM to remain at these surprisingly strong levels through the middle part of the year here into the third quarter, again, primarily driven by HBM and our strong competitive advantage in that area.

Okay. Sounds great. And you just -- you said lead time is very short in this? Is that the same -- about the same level for the corporate average or the long term as well?

Yes. I think probe cards in general is a business where customers really -- yes, they have plans on what wafers they're going to start and therefore what probe cards they're going to need. But they really don't trigger the buy signal and the purchase orders to us until either the wafer start or very close to the wafer starting in the fab. And for those of you who follow the industry closely, that ends up being not coincidentally 6, 7 or 8 weeks for the wafers to transit through the fab. So we and our competitors in the overall probe card industry have basically built operational strategies and architectures that will allow us to deliver on those lead times and give ourselves the customers, the flexibility to adapt their tooling buy plans so that, again, the probe cards are there just in time. It does create quite a burden on having very high incoming quality and making sure you're delivering on time and delivering something that works at the right time. But I think the industry, us and our competitors, have done a good job of improving and evolving over the years to drive on that very efficient just-in-time model. But think of probe cards very much as being indexed to the cycle time of wafers going through with that.

That makes sense. And also, we talk about DRAM, and then I want to move on to foundry and logic. I think you've guided next quarter going to be -- like September quarter going to be higher quarter-over-quarter. So is it mostly driven by -- I guess, it's mostly driven by trailing edge versus like leading edge because everyone's talking about leading edge like weakness? Or is it -- is that like a lot driven by advanced packaging?

Yes. So one of the other important things to understand about FormFactor in the advanced probe card business in general is it's very heavily leveraged towards leading edge because, essentially, what you're doing is helping customers screen out, die that are essentially bad at the end of the wafer fab and are not going to end up as good at the end of the packaging process. You're helping customers screen those out. And you can imagine that yields at advanced nodes are not that high, and so the value of doing wafer test with a probe card is much, much higher for an advanced node and advanced packaging where the packaging costs are relatively high than it is for trailing edge nodes. The flip side of that is where the capital equipment provider suppliers have had some very strong results from trailing edge nodes, primarily out of China. We really haven't seen that. FormFactor, again, you can think of as being leveraged primarily to the leading edge, primarily to advanced nodes without a lot of trailing edge exposure. So as we come through the middle part of the year and we did sequentially guide the foundry and logic business up, a lot of that's just puts and takes between the quarter. I wouldn't read anything sort of structural. We're certainly not calling the next upturn here at this point. I think we've been relatively consistent as we've gone through 2023 that we're pleased to see demand levels stabilizing at these reduced levels for us, somewhere around $160 million, $165 million a quarter, but we're certainly not calling the next upturn here with a slight sequential increase in guidance.

Yes, that's fair. And speaking of the advanced tech packaging, how much share gains or like additional content is there for probe cards or in general for FormFactor?

Yes. So advanced packaging is interesting for us. And because -- go back to why fundamentally customers do wafer test, they're trying to screen out bad die from the downstream processes. And in advanced packaging, this is an even more important step to do because you can imagine that if you're assembling an advanced package with 4 or 5 different chiplets in it at relatively low yields, any 1 of those 4 or 5 chiplets could end up killing the other 4 or 5 -- the other 3 or 4 chiplets in the package. HBM is a great example of this. HBM is these high-speed memories built by stacking, a set of DRAM die on top of each other. Sometimes 8-high, sometimes 12 highs, even 16 die high. And so you can imagine in assembling that stack, our customers want to make sure that each of those die is good, especially as they get to the top levels of that stack before committing the next die to that stack. And so that's raising what we call test intensity, essentially the amount of time that test each of these chiplets and therefore, the number of testers and probe cards required to test each of them. A general rule of thumb, although there are differences with any rule of thumb, specific differences. Advanced packaging on a like-for-like basis, you tend to see about a 20% higher test intensity. One of the proof points of that is if you look at industry spending on advanced probe cards over the past decade or so, 10 years ago, probe cards were about 0.35% of overall semiconductor revenues. Over the past several years, that's moved up to as high as 0.45% of semiconductor revenues. And that's a good indicator of the increase in test intensity associated with trends like advanced packaging in the industry.

Got it. And another key technology transition is gate-all-around. So I want to touch base on that, make sure we touch base on that. And so what are the opportunities there for probe cards in the gate-all-around transition? And will FormFactor gain share through this transition? When do you think that's going to like start to ramp up?

Yes. Although gate-all-around, a very exciting front-end innovation for the industry. Honestly, I don't think it's going to have a really big impact on FormFactor. As excited as we are about advanced packaging, I think we're probably -- new nodes are nice. So the transition from 5-nanometer to 3-nanometer is good just because it drives new designs. But in and of itself, the transition from FinFET to gate-all-around, I would not expect to have a big impact primarily because you'll see a little bit of impact associated with customers trying to resolve the new yield loss mechanisms associated with gate-all-around. So that will drive a short-term bump up in test intensity, but I don't think there's really anything structural there that they won't be able to figure out over a relatively short amount of time as they drive up a yield learning curve.

Just like at the beginning of this transaction.

Yes, just like any node transition, I think. So where we're really excited about advanced packaging, I -- and that's one of the areas where we think we have a lot of exposure to the industry and are continuing to drive additional exposure for FormFactor, I think gate-all-around is probably business as usual.

Okay. All right. That makes sense. And I want to ask about the end market exposure. So could you give us some sense of your end market exposure among PCs, smartphones like all those servers. I know you guys talk about 1/3 memory and 2/3 non-memory. And about within that 2/3, like how is that...

Yes. So if you think about those end markets, in the probe card business, the exposure is quite heavy towards high unit volume consumer-type applications. So client PC, mobile handsets are definite big demand drivers for us. Now right now, those markets are fairly weak. They are -- you hear from our customers. Anybody exposed to those is working their way through inventory corrections, pretty tepid demand compared to where it was certainly in 2021 and into 2022, and that has a consequence on our business as well. If I back you up to the first half of 2022, we were operating at levels near our target model, $850 million of annualized revenue, 47% gross margin and $2 of non-GAAP earnings per share. We're well off of that as we go through the current downturn. And primarily, that's due to the outsized exposure that a company like us that's unit-driven, design-driven is going to have to things like mobile handsets and client PCs. Automotive, some exposure. But again, go back to our comment about advanced node, leading edge nodes versus trailing edge nodes, a lot of automotive chips are built on trailing edge nodes. And so we're not going to have a lot of inherent exposure there. There are counter examples for high-performance microcontrollers where we do have some nice recurring business. But really, the leverage at present is primarily geared towards advanced nodes and high unit volume consumer-driven applications like client PC and...

Okay. Yes. That's very helpful. And I want to switch gear a little bit to systems. It has been calling up very well in this year. It's actually during this down cycle. And I think right now, it's about 25% of our total revenue. So I think that was kind of your target model range of 25% system. So I was -- I want to ask about, is that going -- do you expect this mix to hold up maybe -- or maybe like ship a little bit as market recovers a little bit and like also like how does -- how should we think about your trajectory moving towards our target model with new [indiscernible] target revenue?

Right, right. So yes, the systems business, we spent a lot of time here on probe cards, but the systems business is a very important part of the overall strategy, both from a technology engagement standpoint with our customers, as I said, working with them in early development on next-generation devices like co-packaged optics and quantum computers. But it's also a nice financial contributor. As Elizabeth notes, we've been operating essentially at record levels at about $40 million a quarter. Good gross margins in this business in the neighborhood of 50%, so above what we've been delivering as a corporate average. I think as we continue to move forward, we're going to look for organic growth in that business. But as we progress back towards the $200-plus million quarterly levels revenue associated with the target model, the growth is really going to come from the probe card business, and I think primarily from foundry and logic. If you think about where the growth in advanced packaging is, where a lot of the innovation is in the industry, it is in the foundry and logic space. And we expect to be able to participate that -- participate in that and have that drive the primary delta in growth from where we are today back up to target model levels.

Okay. All right. That makes sense. And what about the U.S.-China restrictions? Is that -- does it have any impact on Form? And like how do you compete with like local, like domestic or like non-U.S. competitors?

Yes. So the export controls associated with semiconductor tooling and consumables and equipment into China for advanced nodes have had a relatively large impact on FormFactor. Again, not too surprising given the advanced node exposure that we have and the bias of the export controls restricting advanced node tooling and technologies. If you look at backing up 1 year, 1.5 years ago, we were doing something like $50 million into the region. And you can go back and look at our disclosures, we do report overall revenue into China. The important thing to recognize about that is it really is split into 2 very distinct components. One is revenue shipped to domestic China customers. The other is revenue shipped to the multinationals that operate in the region, some of our 10% customers. And those have very different risk profiles. The domestic China revenue, as you might imagine, both as a direct result of the export controls and the domestic China customers legitimate and rational response to those export controls, they've tried to work FormFactor out of their supply chain. Interestingly, they've tried to work most foreign suppliers out of their supply chain and try to work mostly local options. Now that comes at a cost, right? Yields aren't as high. The technology is not as good, the cost of test is higher. But nonetheless, they are able to build and test it. And we've been, I think, fairly transparent, but probably relatively pessimistic compared to the rest of the industry that we expect our domestic China revenue to essentially go to 0 in some time frame here over the next year, 1.5 years, maybe 2 years. There doesn't appear to be any reversal in these export controls. Present day headlines may be accepted. But I think there's going to be continued headwinds in an advanced node technology provider like FormFactor supplying into the region and the demand those customers will have for a U.S. supplier like us.

Got it. Yes, go ahead.

Okay. So over the last -- ever since you report the second quarter early in August, right? So there seems to be a lot of changes specific for HBM. There's [indiscernible] by Samsung [indiscernible], and there's tons of investment out there. And I'm just wondering your third quarter guidance will be based on the -- both from the [indiscernible] based on the CFO commentary. But is there any like upside offside financial complement that they've got really achieving [indiscernible]?

Yes, I think so. I don't know whether people on the webcast could hear, so I'll basically paraphrase the question. And it was essentially, is there any upside associated with HBM here in the third quarter compared to the guidance we gave back at the start of August? We have not revised guidance upwards, and primarily the incremental growth going from Q2 to Q3, Elizabeth noted earlier, was primarily associated with foundry and logic. Anytime you have the sort of rapid ramp, you can see investment in advance of our customers producing chips. So a lot of the headlines you're seeing now associated with HBM, FormFactor delivered some of those probe cards and some of that tooling in the second quarter. And that was what we're responsible for the 50% step-up and then continuing at those levels here in the third quarter for -- by and large, for DRAM probe cards. There's a lot of -- in any of these new technology, new product ramps, there's a lot of different constraints throughout our customers' production. I don't think we're one of the constraints right now, but there's various places where they're having to make some large investments even to use the probe cards they already have. So we're not revising guidance upwards. Maybe there's a little bit of noise or upside or downside around the third quarter guidance. But by and large, I think the themes we conveyed a month ago on the earnings call remain intact.

Can you share with us in terms of the market share dynamics right now that either good customer say like, [indiscernible]?

For DRAM? For DRAM. I think historically -- so the question about market share -- probe card market share DRAM. DRAM probe cards, by and large, there's 2 suppliers, FormFactor and a Japanese company called Micronics Japan. Historically, FormFactor has had stronger share at SK Hynix. MJC has had stronger share at Samsung, although we've made some pretty good gains at Samsung in various areas as we've gone through some technology progression. I think anytime you see a 2-supplier market like this, I think both suppliers understand that they're going to be competing with each other, but there's also going to be a fairly narrow market share range that customers want to keep you in. Our customers need 2 suppliers. They need 2 strong viable suppliers, and they're going to continue to cultivate us and give us both business where we offer them some value from the technology to thrive. By and large, that's where the market share shifts occur. And because we have some competitive advantage in the detailed requirements associated with HBM, we have seen some nice share gains through the middle part of the year.

Is it fair to say like 60/40 in terms of [indiscernible]?

That's a good general rule of thumb window to think about. It's a 60/40, 2/3, 1/3 window.

All right. Thanks. Is there any other questions from the audience?

Its' [indiscernible]. The question is regarding operating expenses, which have held kind of flattish even as the business has turned down. How you think about managing that? What flexibility there is or isn't as you build expectations for future growth?

Yes. A couple of different elements to OpEx. One that is an important part of our operational philosophy is quite a bit of incentive-based compensation is geared towards company profitability. And as we went through 2022, you saw that step down pretty significantly as profitability decreased in the second half of the year. Other elements of OpEx that are maybe more structural in nature. I view, and I think we as a management team view R&D, sales and marketing and G&A is 3 very different and distinct buckets. R&D, we're investing today to drive revenue, competitive advantage and gross margin expansion a year, 2 years, 3 years from now, right? That's the innovation that fuels your road map. That's working closely with key customers that you see on our 10% list to make sure we're ready to solve some of the problems that they're going to face as they embark on things like advanced packaging in volume production. Sales and marketing, you've got a broad-based footprint around the world supporting all these global customers, and that sales and service team is an integral part, especially in a business, both the probe card business and the systems business are very high customer touch, high customer intimacy businesses because of the degree of interaction, because of the degree of customization you have inherent to these products. And so those are investments as well. I think G&A is an area where we continue to strive to drive efficiency. We've recently gone through -- not so recently, but during the pandemic went through a major ERP transition, which gave us some degree of efficiency gains across all of the different operations, really consolidating a lot of disparate systems from FormFactor's acquisition history and getting most of the company on to the same platform. G&A is an area where we're going to continue to try and squeeze efficiency and make sure that those costs don't scale with revenue. R&D costs, honestly, I would expect. And if you look at the target model, I'd expect R&D to continue to be a significant investment. I don't remember exactly what the target model is, but think about it as 15% of revenue feels about right.

Share with us in terms of ASP profile for HBM, probe cards [indiscernible].

Yes. ASPs in probe cards are a bit of a dangerous notion because, again, you're customizing to each individual customer chip design. And so the configurations they choose in DRAM, for example, the picture you see on the left here, 1 touchdown DRAM probe card, a probe card that tests the entire DRAM wafer once bears a much higher ASP than a 2 touchdown probe card. Yet sometimes customers will choose the 2 touch down probe card because they have extra test capacity. So we try and steer away from ASPs because it really is a notion that's not so useful in this business, given the degree of mix and customization.

Just in terms of like you mentioned this testing has been capacity [indiscernible] that gives you much higher profitability for HBM probe card versus the [indiscernible]?

Yes. For advanced packaging, in general, we've said -- and this is consistent with the narrative. I think you've heard from the ATE vendors as well, Teradyne and Advantest. On a like-for-like basis, so same number of DRAM bits out. Advanced packaging drives something like a 20% increase in test spending, in test intensity. That's turned out, sure, there's areas where customers spend more. There's also areas where customers spend less. But that's not a bad rule of thumb. And if you go back to the comment I made about overall industry spending on probe cards, you can see that the probe card industry has outgrown the semiconductor in part because of that increase in test intensity.

Any other questions from the audience?

Talk about the margin intensity or the margin profile of the advanced packaging business versus the probe card business.

So yes the...

The margin profile.

Across the probe card business?

Versus the advanced packaging, yes.

Yes. Well, there's different -- and again, this goes back a little bit to ASP, the comment I made about ASPs. I would say, in general, a more complex probe card, a probe card with a higher degree of capability, maybe that's higher speed, maybe that carries more power to the chip, will carry a higher price, right? Customers will pay for performance, will pay for capability. But then there's other trade-offs with things like, well, maybe they don't test as many die at the same time. So a lower parallel in card. So in the same way that ASPs are a somewhat dangerous notion in probe cards, so our general rules of thumb associated with advanced packaging versus that. What I will say is as test complexity increases, so if I'm one of our customers testing individual chiplets before they go into the advanced package and they need to test at higher speeds, that's a degree of capability that does drive a higher price. RF, for example, high-speed digital does drive a higher price, again, on a like-for-like basis. Same number of die, everything else being equal. So I think it's reasonable to expect that if we're able to execute when we're able to execute on our technology road map associated with the capabilities in advanced packaging that you will see us move higher towards the gross margins, again, associated with our target model, something like 47%.

Any other key drivers that goes to the target for the 7% gross margin?

Yes. I think one of the big ones, and you've seen it work in reverse as we've gone from the $200-ish million of the quarter's -- first quarter 2022, second quarter 2022 down to the $160-ish million levels we're at now, volume has a big, big effect on our gross margins. We're pretty vertically integrated manufacturer. The technologies, like our competitors, required to produce probe cards are fairly specialized. And so you end up developing those technologies and then building fairly vertically integrated factory structure to deliver those technologies in volume. The consequence of then a downturn is you're underabsorbing a lot of the fixed cost. And although in the third quarter of 2022, when it was clear to us that there was a downturn on the way, we took our cost structure down significantly. That was primarily associated with labor and things that were variable in the short term. We've continued to improve our factory footprint. We've continued to invest capital in making sure we have world-class facilities that can deliver the products we're developing on the lead times and quality levels that we need to, to remain competitive. And I think those are costs that you're just going to carry through the COGS line during a downturn like this. So volume, right, becomes one of the really big drivers on the way back up to the high 40 gross margin of the target.

Okay. That makes sense. We are almost run out of time. So before we wrap up, I just want to ask if you have -- there's anything that you think investors are not fully appreciative of FormFactor?

Yes. So I think one of the interesting things that's happened is we've come through the middle part of 2023, not just FormFactor, but the industry. So I think there's been a rapid increase in people's understanding of the importance of advanced packaging for the industry. If I back up a year, I think it was largely underappreciated how important advanced packaging and chiplets were going to be for the industry. I think one of the things that's happened against the backdrop of a pretty significant industry downturn here in 2023 is the strength associated with some of the advanced packaging projects like High Bandwidth Memory, like [indiscernible] to deliver the capability for the logic side of AI. And I think ironically, the downturns help provide some degree of contrast for how important advanced packaging is going to be for driving industry innovation and growth going forward. That's one of the reasons, one of the key themes for FormFactor. We've made organic investments there. We've made M&A investments there and continue to try and position ourselves to be heavily levered towards the -- what's one of the most exciting trends in the industry, certainly in the 25 years I've been in.

Okay. Great. Thanks, Mike, for joining us today. We really appreciate your time, and thanks for the audience.

All right. Thanks for having us.