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

Thank you, and welcome, everyone, to FormFactor's Second Quarter 2023 Earnings Conference Call. On today's call are Chief Executive Officer, Mike Slessor; and Chief Financial Officer, Shai Shahar. Before we begin, Stan Finkelstein, the company's VP of Investor Relations, will remind you of some important information.

Thank you. Today, the company will be discussing GAAP P&L results and some important non-GAAP results intended to supplement your understanding of the company's financials. Reconciliations of GAAP to non-GAAP measures and other financial information are available in the press release issued today by the company and on the Investor Relations section of our website. Today's discussion contains forward-looking statements within the meaning of the federal securities laws. Examples of such forward-looking statements include those with respect to the projections of financial and business performance, future macroeconomic and geopolitical conditions; the benefits of acquisitions and investments in capacity and in new technologies; the impacts of global, regional and national health crisis including the COVID-19 pandemic; anticipated industry trends; potential disruptions in our supply chain; the impact of regulatory changes, including the recent U.S.-China trade restrictions; the anticipated demand for products; our ability to develop, produce and sell products; and the assumptions upon which the statements are based. These statements are subject to known and unknown risks and uncertainties that could cause actual results to differ materially from those expressed during this call. Information on risk factors and uncertainties is contained in our most recent filing on Form 10-K with the SEC for the fiscal year ended December 31, 2022, and in our other SEC filings, which are available on the SEC's website at www.sec.gov and in our press release issued today. Forward-looking statements are made as of today, August 2, 2023, and we assume no obligation to update them. With that, we will now turn the call over to FormFactor's CEO, Mike Slessor.

Thanks, everyone, for joining us today. FormFactor's second quarter revenue was below our outlook range due to shipment push-outs. However, higher-than-expected gross margin and good operating expense control produced non-GAAP earnings per share at the top end of our outlook range. We continue to operate in an overall demand environment that remains relatively stable. This is evident from our third quarter outlook. As we navigate the current cyclical downturn, we continue to benefit from FormFactor's diversification strategy, which differentiates us from our direct competitors with a broad lab-to-fab product portfolio across Foundry and Logic, DRAM and Flash probe cards, together with our System segment products. This unique portfolio enables us to compete for business across diverse demand pools at all major customers, producing the relatively stable top line results we've demonstrated for the past several quarters, and that we expect again in the third quarter, as we explore growing areas like high-bandwidth memory and DRAM probe cards and co-packaged silicon photonics and systems, helping to offset the impact of areas that are presently soft, like mobile handsets and CPUs in Foundry and Logic probe cards. While the demand softness that has persisted for several quarters continues in some of our most important high unit volume end markets, like client PC and mobile, we're carefully balancing short-term results and long-term investments with disciplined cost control that maximizes quarterly profitability and protects our strong balance sheet. We continue to invest in R&D for new product innovation and competitive differentiation as well as in the long lead time facilities and equipment components of our capacity increase plans. These investments are designed to produce market share gains and above industry revenue and profit growth, when we emerge from the current cyclical trough, enabling FormFactor to achieve and then surpass the levels of our current target financial model. Turning now to segment level details. In Foundry and Logic probe cards, our largest business, we experienced the expected moderation in demand in the second quarter, as customers digested first quarter shipments for initial production ramps of 5-nanometer and 3-nanometer mobile application processors, tile-based client microprocessors and high-performance compute chips. In the third quarter, we expect to deliver sequential growth in Foundry and Logic as we see modest strengthening in both mobile and client PC markets, primarily driven by adoption of advanced packaging processes like Foveros and 3D Fabric. Advanced packaging and chiplet innovations were prominently featured at the recent Semicon West industry trade show, where a variety of major front and back-end suppliers along with leading customers highlighted the increasing role advanced packaging and chiplets are playing in reaccelerating industry innovation in the face of the slowing of front-end-driven Moore's law. As we've noted in the past, advanced packaging integration schemes drive both higher test intensity, which expands the number of probe cards required for good die out and higher test complexity, which raises the performance requirements for each probe card. Probe card architectures like FormFactor's MEMS technology are essential to meeting these challenging performance requirements on short lead times at a compelling cost of ownership. In Memory, as expected, we experienced stronger second quarter demand for DRAM probe cards, while Flash remained weak. The sequential increase in DRAM was driven almost entirely by DDR5 and third-generation high-bandwidth memory designs that are enabling the rapid growth in generative artificial intelligence. DDR5 and HBM together more than offset the softness from each of the memory manufacturers' reductions in wafer starts that are helping consume the high inventory of older memory designs like DDR4. HBM is yet another example of how advanced packaging is enabling the industry's innovation and long-term growth since one HBM chip is a stack of 8, 12 or even 16 individual DRAM die. To ensure high yields of this stacked DRAM chip, customers probe and test each component DRAM die prior to stacking and probe and test a multi-die DRAM stack at various points during the stacking process, leading to an increase in the overall probe card intensity for good die out. In addition, the technical requirements for HBM tests are significantly more advanced than for standard unstacked DRAM products, involving higher test feeds and more challenging thermal scaling specifications. As one example of increased test complexity, the 100,000 MEMS probes on a single HBM probe card let's remain within a few microns of their target position across an entire 300-millimeter wafer, as the wafer temperature varies by almost 100 degrees. FormFactor's MEMS based SmartMatrix DRAM probe card architecture meets these advanced requirements, providing significant value to our customers and differentiation for FormFactor. We believe our superior performance capabilities will drive both market share and profitability gains as HBM grows from a current single-digit percentage of DRAM wafer starts to 20% by 2026, as projected in a recent Wall Street Journal article. Shifting now to Systems. Our Systems business continues to deliver strong results with revenue at near-record levels again in the second quarter. These levels were, however, less than we expected in our outlook due to some delivery pushouts caused by a combination of customer facilities' readiness delays, single-point supplier issues and internal first pass yield shortfalls. We expect these systems to be shipped in the current third quarter, and we expect to again deliver strong Systems' segment revenues in the quarter. The strength is driven by our market-leading test and measurement products for early development of applications like co-packaged silicon photonics, quantum computing and advanced packaging metrology. Co-packaged optics enabled by silicon photonics continues to be an important and exciting driver for FormFactor's current and future business. This technology is poised to revolutionize chip-to-chip communication in the data center, with significantly reduced power consumption at high speed, driving industry forecasts of 30-plus percent growth rates for silicon photonics devices over the next several years. In the transition from early R&D to low-volume production, customers are deploying FormFactor's turnkey electrooptical measurement systems built on our CM300 and Summit 200 engineering probers with subsystems from leading partners like Keysight Technologies and PI. As production volumes increase over the next several years, our product road map delivers both systems and consumable probe cards to test and improve yields of electrooptical silicon photonics chips, enabling our customers to seamlessly transition their co-packaged optics devices from the lab to the fab, ramping quickly to high volumes to meet the demand for these revolutionary and innovative devices. Finally, I want to share an important customer highlight from the second quarter. FormFactor was 1 of only 6 suppliers to receive the exclusive Intel Epic Program Outstanding Supplier Award for 2023. This award recognizes the absolute top performers in the Intel supply chain for their dedication to continuous quality improvement, performance, collaboration and inclusion over the past year. I'm extremely proud of the Global FormFactor team for the dedication and performance that resulted in this top-tier recognition from Intel, and I'd like to take this opportunity to thank and congratulate our world-class team. In closing, let me reiterate that in the short term, we continue to be encouraged by the stabilization of demand across our diversified product and technology portfolio as we progress through the cyclical challenges of 2023. Over the longer term, we remain excited and confident in the growth prospects for FormFactor and the industry overall, driven by the fundamental trends of semiconductor content growth and innovations, like chiplets and co-packaged silicon photonics. These are trends where FormFactor is well positioned as an industry and technology leader, and we're confident that our commitment to invest in R&D and capacity will position FormFactor to emerge from the current cyclical downturn a stronger and leaner competitor, enabling us to achieve our target model, deliver $2 of non-GAAP earnings per share on $850 million of revenue. Shai, over to you.

Thank you, Mike, and good afternoon. As you saw in our press release, and as Mike mentioned, Q2 revenues were slightly below the low end of our outlook range, non-GAAP gross margin was above the high end of the range and non-GAAP EPS was at the high end of the range. Second quarter revenues were $156 million, a 6.9% sequential decrease from our first quarter revenues and a year-over-year decrease of 23.5% from our Q2 '22 near-record quarterly revenues. Q2 revenues were $1 million below the low end of our outlook range due to pushouts of shipments in our Systems segment. Probe card segment revenues were $115.3 million in the second quarter, a decrease of $12 million or 9.4% from Q1. The decrease was driven by lower Foundry and Logic and Flash revenues, partially offset by an increase in DRAM revenues. The Systems segment continues to operate at near-record levels and revenues were $40.6 million in Q2, a $0.5 million increase from the first quarter, and comprised 26% of total company revenues, slightly up from 24% in Q1. Within the probe card segment, Q2 Foundry and Logic revenues were $81.9 million, a 19% decrease from Q1. Foundry and Logic revenues decreased to 53% of total company revenues compared to 61% in the first quarter. DRAM revenues were $30.5 million in Q2, $10.7 million or 54% higher than in the first quarter and increased to 20% of total quarterly revenues as compared to 12% in the first quarter. Flash revenues of $2.9 million in Q2 were $3 million lower than in the first quarter and over 2% of total revenue in Q2 as compared to 4% in Q1. GAAP gross margin for the second quarter was 38.7% as compared to 36.5% in Q1. Cost of revenues included $3 million of GAAP to non-GAAP reconciling items, which we outlined in our press release issued today and in the reconciliation table available in the Investor Relations section of our website. On a non-GAAP basis, gross margin for the second quarter was 40.6%, 2.2 percentage points higher than the 38.4% non-GAAP gross margin in Q1 and 110 basis points above the high end of our outlook range. Both segments gross margins were higher than Q1 gross margin and higher than forecasted with a more significant upside in the Probe card segment. Our Probe card segment gross margin was 36.5% in the second quarter, an increase of 2.2 percentage points compared to 34.3% in Q1. The increase from Q1 is due to the net effect of 3 main factors: First, 600 basis points of the increase relates to improved factory utilization and lower manufacturing spending. The improved factory utilization, even at lower revenue levels, relates to higher manufacturing activity near the end of the quarter. Second, 90 basis points of the increase relates mainly to reduced excess and obsolete inventory reserves. Partially offsetting these 2 positive factors was 400 basis points related to lower revenue and a less favorable product mix with lower Foundry and Logic revenues and higher Memory revenues. Our Q2 Systems segment gross margin was 52%, 30 basis points higher than the 51.7% gross margin in the first quarter, reflecting slightly higher revenue and the more favorable product mix. Our GAAP operating expenses were $61.6 million for the second quarter as compared to $61 million in the first quarter. Non-GAAP operating expenses for the second quarter were $52.1 million or 33.4% of revenues as compared with $51.2 million or 30.6% of revenues in Q1. The $0.9 million increase relates to G&A expenses specific to Q2, partially offset by lower performance-based compensation. Company noncash expenses for the second quarter included $9.2 million for stock-based compensation, $2.7 million for the amortization of acquisition-related intangibles and depreciation of $7.5 million, all at similar levels to the first quarter. GAAP operating loss was $1.3 million for Q2 compared with GAAP operating income of $0.1 million in Q1. Non-GAAP operating income for the second quarter was $11.2 million compared with $13.2 million in the first quarter. GAAP net income for the second quarter was $0.8 million, or $0.01 per fully diluted share compared with a GAAP net income of $1.3 million or $0.02 per fully diluted share in the previous quarter. The non-GAAP effective tax rate for the second quarter was 14.2%, similar to the 13.9% in the first quarter. We expect our annual non-GAAP effective tax rate to be between 13% and 17%. Second quarter non-GAAP net income was $11.2 million or $0.14 per fully diluted share compared to $12.5 million or $0.16 per fully diluted share in Q1. Moving to the balance sheet and cash flows. We generated free cash flow of $2.1 million in the second quarter compared to a negative free cash flow of $7.3 million in Q1. Cash provided by operating activities in Q2 was $22.5 million, $10.2 million higher than the $12.3 million in Q1. The main reason for the increase in cash provided by operating activities is improved working capital management. We invested $20.5 million in capital expenditures during the second quarter compared to $19.7 million in Q1. With the core drivers underpinning our strategy still in place, we continue to execute on increasing our long lead time facilities and equipment portions of our capacity increase plans, albeit replacing equipment in service at a slow rate to ensure capacity does not significantly outpace demand. Accordingly, we are updating our expected CapEx for 2023 to range between $55 million and $65 million, $5 million higher than the previously communicated range. Overall, at quarter end, total cash and investments were $240 million, similar to Q1. As of the end of the second quarter, we had 1 term loan remaining with a balance totaling $50 million. Regarding stock buyback. During the second quarter, we did not purchase additional shares under our $75 million 2-year buyback program. At Q2 quarter end, $18.6 million remained available for future repurchases. Turning to the third quarter non-GAAP outlook. We expect Q3 revenues to be higher than in Q2 with higher Systems and Foundry and Logic revenues. This results in a Q3 revenue outlook of $167 million, plus or minus $5 million. Third quarter non-GAAP gross margin is expected to be 40%, plus or minus 150 basis points. At the midpoint of these outlook ranges, we expect Q3 operating expenses to be $53 million, plus or minus $1 million. Non-GAAP earnings per fully diluted share for Q3 is expected to be $0.17, plus or minus $0.04. A reconciliation of our GAAP to non-GAAP Q3 outlook is available on the Investor Relations section of our website and in our press release issued today. With that, let's open the call for questions. Operator?

[Operator Instructions] Our first question comes from the line of Krish Sankar of Cowen.

First one for Shai. It seems like the mix is skewing more towards Foundry and Logic in September quarter, but the gross margin seems flattish. I would like to know what's going on there? Because I thought the Foundry and Logic has a better gross margin than DRAM. And I'm going to have a follow-up for Mike.

Sure. Thanks, Krish. I want to start with -- we've talked before about expecting high 30s gross margin at the current revenue level. So we are obviously encouraged by the Q2 results and by the Q3 gross margin outlook of around 40% at the midpoint of this outlook. So we expect gross margin to increase to low 40s once we achieve quarterly revenues of about $180 million on our way to our target model of 47% of revenue levels, around $200 million, $210 million, as we demonstrated a year ago, right, with a record revenue, and we expect the gross margin to fluctuate along this trend line as we make progress. Specifically for Q3, gross margin as compared to Q2, the mix impact is actually expected to be flat once you go down to 1 level below the segment revenue. The impact of the higher revenue versus the second quarter is expected to be offset by a return to normal factory utilization after strong manufacturing activity during the second quarter of Q2 for Q3 shipments. Also, we expect some higher material costs in Q3 as compared to Q2.

Got it. And then a bigger picture question for Mike. Thanks for the color on HBM. When I look at GPUs, historically, it looks like they use legacy probe card technology, but with AI and CoWoS, it seems like they need to use MEMS. So I'm just kind of curious, A, is that true? If so, kind of when are you expecting to see the benefit and how to kind of like quantify the revenue opportunity set for FormFactor as we go into MEMS-based probe card technology for CoWoS?

Krish, good question, and I agree with the underlying assumptions in it. Historically, the leading GPU manufacturer has used legacy non-advanced probe card technologies and hasn't needed a MEMS probe card from us or our major competitor. As you know, that's changing with the adoption of CoWoS and advanced packaging. And both of us are in the middle of qualifying MEMS-based technologies for these advanced packages for these CoWoS technologies. It's a relatively modest number of wafer starts at this stage, which I think you know. So hopefully, this trend continues. This growth in AI certainly is expected to continue, and we'll be able to see that as a more significant part of our 2024 revenue as the MEMS adoption associated with the CoWoS technology for GPUs really takes hold, and we see that ramp in more significant volume.

Our next question comes from the line of Vedvati Shrotre of Jefferies.

My first question, I guess, last earnings call, you talked about seeing very strong design activity and the volumes were muted. And then can you just compare and contrast how this has changed this quarter? In terms of design activity, has that increased significantly from last time what you saw in either volumes starting to come back now that AMD and Intel are noting recovery in their PC business?

Yes. So -- go ahead.

Sorry, I have a follow-up after that.

All right. So design starts and design activity, we're seeing similar levels as we have for the past couple of quarters. Remember that a probe card is specific to each customer chip design. And so as customers release their new designs, they work closely with our design teams to customize the probe card architecture to meet the specific test requirements they have for that chip design. And as we've noted in the past couple of quarters, we've got a very robust design activity where -- but not shipping very many probe cards units on each new NPI. In an upturn, we might expect to shift tens or even hundreds to meet the production ramp, now we're shipping a few. And this is really a dynamic that most of our customers are adapting to as they try and consume the inventory that they have on their balance sheets, that the channel has of their parts. Obviously, they don't want to obsolete this inventory by overwhelming it with new designs. And so we're seeing a nice setup as that inventory does get consumed, where these new designs that we've already qualified and are ready to ship in volume, will then refill the channel. But as you know, there's a few green shoots in some areas like client PC. I don't expect a rapid turn there. I don't expect a rapid turn in mobile, but I do think we're well positioned, and we are seeing a little bit of increased activity. I wouldn't say the situation on design versus units has substantially changed from previous quarters.

Got it. And then on the HBM side. So can you help me understand how -- maybe your largest customer on HBM has talked about their revenues doubling in 2 quarters. How does that translate to your business? Like going forward in the second half of 2023, do you think the DRAM portion of your business would be much bigger than where it is now? Or is that kind of coming in?

The second quarter is a pretty good example of the strength in HBM. As we said in the prepared remarks, most of the increase we saw was really attributable to HBM, some to DDR5 as well. If you look at from Q2 to Q3 -- sorry, from Q1 to Q2, we grew the DRAM revenue by about 50%. Almost all of that growth comes from HBM. So you're seeing that already. We do expect continued strength there as we go through 2023, but there will be some digestion. I wouldn't expect it to double again in the third quarter, for example, from an HBM perspective. But it's great -- considering how commodity DRAM or non-HBM DRAM is still relatively weak in working through some significant inventories that are still on our customers' balance sheets and then the channel, HBM is an awfully nice growth vector to be exposed to right now.

Our next question comes from the line of Charles Shi of Needham & Company.

Thank you allowing me to ask a couple of questions. I want to ask about the gross margin. I'm glad to see margins back to the 40s range -- has been for Q2 and going to be for Q3 after -- I know we have a few quarters of margins in the 30s range. Are you comfortable to really call maybe that low-margin period like in the 30s, it's kind of behind us and going forward maybe we will be able to maintain at above 40s kind of gross margin?

Yes. So the 2 main factors that influence gross margin at this point are revenue levels and mix. I answered the previous question by repeating what we said before that in order to gross margin to increase to the low 40s, call it, 42%, 43%, we need quarterly revenue to go up to around $180 million. And in order to get back to the 47%, at our target model gross margin like we did a year ago, we need revenues to grow to $200 million and $210 million. And we also made this growth to come from our higher-margin markets, which is mainly Foundry and Logic. Now along this trend line that I just described where we are now with, let's call it, 40% to 47%, if revenue stays at these levels and mix remains similar, we should be able to maintain this trend line that I was talking about, but we do expect to continue and experience fluctuations along that trend line.

Maybe the other question, I'm glad to see some strength on DRAM probe card. Well, certainly, I was not expecting it to bounce back to $30-plus million for the quarter this soon. Maybe just a question for Mike. Originally, we were expecting maybe it's going to take a let down in the middle part of 2023, but now we're not quite seeing that, although you did provide some more cautionary statement around, okay, maybe it's not going to go 50% higher next quarter. But really just wanted to understand, this $30-plus million for the quarter, do you think this is going to be a sustainable quarterly run rate for your DRAM business going forward or maybe some fluctuation will get us back to the 20s range?

It's an interesting question, Charles. And maybe to come at it from a different direction. Our DRAM probe card business right now is really the tale of 2 submarkets. One, which is very strong that we talked about in the prepared remarks and have touched on here in Q&A is HBM driven by generative AI, right? I think everybody understands the connection where HBM is one of the key enabling semiconductor components along with the GPU for generative AI. And so we're seeing some very strong growth there. The other part of the DRAM business, sort of legacy DDR4, low-power DDR4 is really very weak, right? And if I were to estimate it for you, the second quarter revenue was composed for DRAM almost 1/3 in HBM. We've never seen that kind of contribution level from HBM before. And an initial ramp like that, HBM is still only probably 5% of industry-wide DRAM wafer starts, projected to grow significantly over the next couple of years. It's going to be a little bit lumpy on the way there. If we can get both components of DRAM healthy and active, if you had HBM on top of the, if you like, commodity or legacy DRAM, then I think you've got the opportunity for DRAM probe card revenues and DRAM probe card spending intensity to drive up significantly. But I'd expect some lumpiness on the way there.

May I ask a quick follow-up. So the standard DRAM sounds like roughly $20 million each -- was from the standard DRAM, it was kind of like in line with the Q1 numbers. Do you think that's really probably the cyclical bottom level run rate there? Because I did notice that was probably the case historically for FormFactor's DRAM business over there.

I think as long as DRAM continues to get healthy from an end market perspective -- you've heard all our customers talk about inventory starting to be consumed. They've taken down wafer starts. They're managing their way through this. As long as that continues, I think it's a reasonable assumption that the $20 million-ish level is the cyclical low for DRAM. We should see it build from there. And hopefully, we'll see continued growth associated with HBM and AI that will put another leg up on the overall DRAM market with our strong market share. I think you can extrapolate those into FormFactor's growth and results.

Our next question comes from the line of Craig Ellis of B. Riley.

As long as there's attention on high-bandwidth memory, given that surge, I'll go in with a few more. First, Mike, can you talk about your customer shipment diversity within high-bandwidth memory? Clearly, there's been one player that's led that market historically, another aspires to gain share, another is interested in getting product out. But where are we now with respect to customer shipments? And how should we think about the coming 2 to 4 quarters? And what will happen there from an OEM standpoint? And related to high-bandwidth memory, given how tight we are on supply, where is Form on being able to deliver for probe cards, when and where customers need them right now?

So on the customer HBM side of the equation, I think for those of you who follow this market, you do understand that there's one pretty significant leader in terms of current market share of HBM. And it turns out that, that's a customer that has a very strong historical relationship with FormFactor and one where we have relatively high market share. So that's helping us right now. If you think about our engagements with the other 2 major DRAM manufacturers, I think they line up reasonably well with the share dynamics. We're active with each of them. We're engaged with each of them. But if I think about the current shipment profile, it's is heavily biased towards the leading provider of HBM memories right now. If you think about constraints, as far as we sit, we're in pretty good shape. I don't feel like we're a constraint on the overall HBM ramp. There's probably other areas of process tools, some yields, which secondarily help us, right? A lower yield part needs more probe cards to get a certain number of die out, and test intensity is going to go up in that situation. We're currently enjoying that and helping resolve those yield problem. So I don't think we're the current constraint on HBM production and the ramp associated with those things. And you can imagine that given the capacity investments we've made and the CapEx we've deployed, if we need to turn on a few elements of that to meet this HBM demand, we're at the ready to do that.

That would be a nice problem to have. And then following up, and I'll just spin it over to Foundry and Logic. When I look at supplement that shows revenue mix by customer, we do see Intel drop into the lowest percent of sales, and I believe it would be lowest dollar sales in the last 4 quarters. Charles asked about a trough in DRAM. Do you feel like your largest customer is at trough levels? Or how do we think about the gives and takes as we look into the back half of this year and 2024 there?

I think your observation is a good one. I think there's 2 components to this. One is, our largest customer and generally Foundry and Logic, there are some important leading-edge designs that are beginning to ramp on advanced packaging technologies, 3D packaging chiplets, those kind of enabling technologies. And we're just in the middle of that transition and ramp with that customer. I think layered on top of that, you've seen some pretty significant weakness in the client PC business, analogous to what's happened in Memory. It looks to be now rationalizing and recovering, but it's also going to take end market growth, consumer PC refresh cycles to really start to pull the inventory in some of these new products through where customers are going to ramp wafer starts significantly, buy probe cards for those wafer starts. So I think as long as things stay in this relatively stable situation that we currently have and start to consume inventory, new products coming into the channel that need new probe cards to test and get to the back end, I think we'll probably see things turn up from here. And I made a comment in the prepared remarks that one of the areas of strength sequentially in Foundry and Logic we do see is client PC. And I think that's consistent with our customers and their narratives over the past couple of weeks associated with that market.

That's really helpful. And then if I could just sneak in one more. You and Shai noted that Systems was an area where there was a pushout in the business. But if we adjust for that pushout, if not for that, Systems would have been at significant record levels in the quarter. And so the question is, given where natural demand is, if not for that shipped, can you talk about how you're feeling about the trajectory of the Systems business? And are we starting to move into new higher levels of revenue generative capacity, given some of the trends that you talked about earlier in your remarks?

We're very excited about the Systems business. We're disappointed in the pushouts. If you extrapolate $5 million to $6 million gap to our midpoint in the outlook, again, reiterating that we are planning to ship those systems in the year in Q3. So it's not perishable. But relatively high ASPs, as you might infer, and tools that are really enabling the next generation of semiconductor innovation. I highlighted co-packaged optics and silicon photonics in the prepared remarks. That's a real driver. But there are some other ones where we have unique capabilities, whether it be cryogenic systems for test and measurements in quantum computing, metrology and inspection for advanced packaging. It's an area where customers continue to invest. Semiconductor R&D spending by our customers continues to be at record levels as they continue to innovate in all these exciting areas. And our tools in the Systems segment are really enabling that and customers are continuing to spend aggressively on that, and so a good example of FormFactor's diversified strategy where we've got some strength in the Systems business, good momentum in the Systems business that's helping at least partially offset some of the softness overall in the production probe card business because of the cyclical downturn.

Our next question comes from the line of David Duley of Steelhead Securities.

I have a couple. Mike, I've asked this question in the past, but I'd just like a little bit of further elaboration. You've talked about how high-bandwidth memory is more probe and test intensive for two reasons. One is it's packaged and there's more steps in the stack of packages. And then also because the chips and cells are more complex. So we add these 2 things up, how much more probe intensive do you think high-bandwidth memory is versus standard memory?

On a like-for-like basis, let's call it bits out the door, HBM, as with most advanced packaging technologies, somewhere between 20% and 30% more test intensive as we see it. Now early in any of these technology ramps, you're going to get an intensity bump because yields are low. Customers are working through all of the yield improvement challenges associated with things like stacking die. And that's one of the areas where FormFactor provides unique value, right? So the high fidelity tests, the app speed tests we offer really allows customers to detect and then fix these yield problems associated with new technologies so that they can then ramp them aggressively in volume at higher yields. So rule of thumb, I think, we see it in HBM, we see it in most chiplet technologies somewhere around 20% to 30% more test intensive on a like-for-like out basis.

Okay. Then as my second question. The Foundry and Logic business declined, I think you mentioned 19% or 20% or so sequentially. Can you just talk -- elaborate a little bit about what the key weakness was behind that drop? And then just kind of a similar question, what percentage do you think of 3-, 5- and 7-nanometer nodes are using MEMS technology now for probe? Are there other big segments that don't use MEMS, such as the GPU segment that's switching over at this point?

So Foundry and Logic was down in the quarter. If you go back to what we said on the Q1 call, that was an expected decline. We had quite a strong Q1 in Foundry and Logic. And we had an expectation for some digestion there and that's exactly what we saw. We now, in our third quarter outlook, see Foundry and Logic rebounding some, some strength driven by both modest increases in mobile and client PC. But as our largest business, this is an important one that we continue to compete hard in and continue to make sure we're engaged with all the top customers. If we think about areas of that market that are still unserved or underserved by MEMS probes, we've talked already on this call about the GPUs, and that's really because they are very large footprint and don't need the density that a MEMS probe has. So customers can still utilize the older technologies with probably quite a bit lower cost. And so that's a rational decision for them. As they move to some of these advanced technologies like CoWoS at 3-nanometer and 5-nanometer, you really have the MEMS probe. And so that's an opportunity that's in front of us and in front of the 2 primary MEMS probe suppliers in Foundry and Logic. That's probably the last real hole in the advanced node adoption of MEMS probes. We see widespread adoption across most of MPUs, application processors, any kind of high-performance compute, GPUs is really the hole at this point.

Our next question comes from the line of Christian Schwab of Craig-Hallum.

Following up on the GPU MEMS movement in the last advanced node adoption, I mean how big of a business could that be in 2 to 3 years, bigger than it is currently?

I think the one thing to be somewhat cautious about it is, and I think most people have understood the GPU dynamics and generative AI adoption is the units are relatively low, right? These are very large and complex products that they're extremely high ASPs for that customer. And so when you think about it in terms of test intensity, dollars spend on probe cards more specifically, there's probably not a huge needle mover there, at least in the initial adoption. The foundries are talking about adding a few thousand wafer starts associated with CoWoS. If you contrast that to mobile and some of the big projects of the industry, the high unit volumes associated with mobile completely dwarf, right, the increases we're talking about with CoWoS. And so while it's an exciting opportunity, and we're seeing its impact in DRAM with HBM, it's an area where we want to be somewhat cautious in creating expectations and adding capacity and planning around it. I think we'll see exactly the same dynamics as we've seen in other parts of the business with the adoption of MEMS probes where you see increased value, increased test intensity associated with both the probe card sale, but also the test intensity as these customers try and adapt their integration strategies to the advanced packaging techniques like CoWoS.

Okay. Great. And then my last question have to -- two-pronged, have to do with Foundry and Logic. First of all, how do you guys view your market share position on Meteor Lake by your largest customer? Our checks suggest that will be greater than the previous generation Sapphire Rapids. Just want to confirm that. And this is a chip that Intel is massively excited about, thinks it's the biggest chip introduction since bringing WiFi to laptops in 2000 as they're going to bring AI to the Edge and to the personal computer. So if they're remotely directionally accurate about that, that should be a big deal if you are positioned well there from a market share perspective. That's part one of the question. Part two of the question is, it's my understanding that you guys are close to potentially selling to, call it, Intel's largest CPU competitor, a person you previously haven't had material market share with that could ramp to be a material business. So if you could give us a quick update on both of those, that would be great.

Sure, sure. I think a couple of things. With our largest customer, a design like Meteor Lake has us all excited for a couple of reasons. One, I think the value it delivers to their customers is significant, and there's some fantastic innovation in what it does to the client PC experience. The other reason that's exciting is it's the first major high unit volume consumer product that's going to utilize advanced packaging, this die stacking. And so it's exciting from a couple of different perspectives. Compared to some of the server chips like Sapphire Rapids, we do have a better share position in the client products like Meteor Lake. And so it's an area where we're excited about that ramp. I think you still need to be a little bit circumspect, given the weakness in the client PC market. It does appear to be stabilizing and improving, but we're definitely not off to the races like we were in the midst of the pandemic when everybody was refreshing their PCs. The other point I would make before shifting to the second part of the question is, we were 1 of 6 suppliers who received the highest supplier award from Intel, and we were the only test and measurement supplier. So I think that's an indicator of the strength of that relationship, the robustness and health of that relationship, which presumably translates into market share as well. On qualifying at the other MPU customer, we -- it continues to be an extremely high priority for FormFactor. Clearly, to be only qualified at one of them, you are then exposed to whatever share shifts occur in that market. And those share shifts, I think everybody understands have not been friendly to our largest customer. Having said that, we're still working with that customer to modify our technology, to adapt our technology to their specific test requirements, which like many customers are unique. Their test engineering teams have built up a strategy that they believe offers them unique value, and we continue to modify our technology, develop our technology, so that we can be qualified and grow our business there. So it continues to represent a growth opportunity for us, continues to represent a high priority for us. We understand the strategic imperative of getting qualified there.

And just a follow-up on that, if I can. When will we verify or validate the fact that you would get there? And then how long, if you did, would it take you to get a meaningful market leadership position at that customer? I assume it would take 2 to 3 years, but it could be faster depending on how your competitor is doing. But my guess is they may spend at least $100 million on these products a year. So it's kind of a substantial opportunity for you and -- in the ballpark or is that not right?

I think you're a little high, but let's call you are close to the ballpark, maybe in the parking lot. They are a significant spender. When you think about the timing that you just articulated, that's about what it takes, right, to get qualified and then win a few designs and ramp. So this is an opportunity that is very much still in front of us. I want to reiterate that it is a strategic imperative for FormFactor to get ourselves qualified and realize that a significant amount of share of that significant spend on probe cards in the industry.

[Operator Instructions] Our next question come from the line of Linda Umwali of D.A. Davidson.

Tom Diffely. So previously, earlier, you referenced some inventory issues at your clients that was still kind of gating your ability to ship some new product. What do you think the industry is or your clients are in the burn of their excess inventory levels at this point? Are there still a few quarters left of that?

I think it's different in different segments and end markets. If you look at something like Flash, the Flash has a long way to go, at least NAND flash has a long way to go. DRAM probably in a little bit better shape, but you heard narrative from some of the big DRAM manufacturers in the last several weeks where they feel like that's turning, but there's still a significant amount of inventory to get through. Mobile, probably a few quarters to go as well. Some of the areas where we don't have as much exposure, things like automotive, trailing edge nodes, analog, things like that, where there's still a pretty healthy supply-demand balance and relatively little inventory, those are pretty cleaned up. So I think you've got the whole gamut across the semiconductor industry, but the areas that drive significant volumes of new probe cards for us and new unit volumes, higher unit volumes of probe cards on new designs, there's still some inventory to get through here.

Okay. That's helpful. And Mike, I was curious, so I was wondering if you could give us just a little more information on the silicon photonics opportunity for you? Maybe how big it is today? How you think the ramp up is in that space over the next few years and how big it could be?

Silicon photonics and co-packaged optics, we talked about in the prepared remarks, is a really exciting area for us. One, because we've got a strong position with our tools and the partnerships we have with companies like Keysight and PI and providing a turnkey electrooptical measurement system. And what customers need to do is, test these chips both optically and electrically. And we're one of a handful of suppliers, maybe the only supplier in the world, who can bring those technologies together to provide that kind of hybrid electrooptical test system. This is still in late R&D, early pilot production. So one of the things we're now focused on is how do we take those hybrid electrooptical technologies and really capitalize on them as silicon photonics grows. It's got 30-odd percent growth rates from most external third-party market research firms. Co-packaged optics in the data center, a wonderfully compelling trend just because of the power efficiency improvement. So what we're trying to do right now is position a business that's maybe 10% to 20% of Systems segment revenues, and really position ourselves to take advantage of the growth of that in our production business, in the probe card business, and probably this is one of the places where we extend the Systems business into a more production-like environment over the next couple of years. So I'd expect, it's one of the reasons why we keep talking about it. It's one of the areas where investing a lot of in R&D, we're coupled with the leading customers in the industry, and it's an area where we would expect to track this 30-odd percent growth rate of the industry when we're successful in executing.

So if it's in late R&D at this point, are we a couple of years away from meaningful production? Or could it be faster than that?

It could be faster now. A couple of customers have time lines that are -- they're still into late 2024, but I think that's a reasonable expectation for some of these production ramps to start.

Our next question comes from the line of Gus Richard of Northland Capital Markets.

I have my first question, this is a second a 7-part question, just kidding. If you think about silicon photonics, can you give us an idea how many people you are engaged with?

So we've sold tools to many, many customers numbering in the dozens. But I think this is one area where you have new technology where you really want to partner with the people who you feel are going to be the leaders. And so we focus a lot of our R&D engagements on a handful of customers, making sure we're working closely with them and developing the systems and modifying the systems and making sure we understand their road maps to go into production. So while it's a fairly broad-based -- broad installed base of silicon photonics tools that we have, again, numbering in the dozens, really focused on a couple of driver customers. And that's a strategy that's consistent with the way we drive most of our R&D across all of our businesses.

Great. That's helpful. And then thinking about your Foundry and Logic business, if you kind of cut that segment by, say, how much of it is mobile and PC clients and how much of it is chiplets. I know there are 2 different categories, but if you can kind of give us a sense of how big chiplets are and sort of how big the client piece of it is, that would be helpful.

Well, I think it also depends on what time frame you're talking about. If you go back to 2021 when client PCs were on fire, it was a very large piece of the business. And we also had mobile and 5G layered on top of that. So you can probably back into what the percentages are based on some of the historicals. I don't have that off the top of my head, and would not really want to speculate on the segmentation. It can often be difficult to sell, especially in mobile, exactly where a chip is destined for. So we want to be careful with some of those. But definitely, they are very large components of the Foundry and Logic business, well north of 50%. If you then think about the cut-through chiplets, it's relatively smaller in Foundry and Logic because we're just in the early innings of adoption. We talked about Meteor Lake in somebody else's question, that's an example of really the first high-volume client or consumer-driven part, whether it's client PC or mobile, that's adopting these chiplet architectures. And so if I were to hazard a guess, we're still in Foundry and Logic, probably mid-single digits, high single digits, relatively low percentage. It's one of the reasons why it's such an exciting opportunity. The increases in test intensity and test complexity associated with advanced packaging and chiplets, as chiplets become a bigger part of the overall Foundry and Logic industry, that's an exciting growth opportunity for us.

Our next question comes from the line of Brian Chin of Stifel.

I have a few questions. Maybe just to clarify something. So the pushout, it sounds like it's kind of $5 million, $6 million, and you expect to recapture that in 3Q. Was that entirely the Systems, the sort of on-demand related reasons that you laid out? Or was there any delay in terms of the probe card revenue as well?

Well, in the probe card segment, because it operates on such short lead times, we often have differences from the assumptions we've made in our outlook. Right now, we've still got essentially 2 months to go in the quarter. And so we do see changes in the probe card business as the quarter evolves. And as we book and ship things, we still have a good chunk of turnables to go here in the third quarter. And the same situation existed in the second quarter. So a little more difficult to make those comparisons in the probe card business. But if you look at, what we said on the call, the Systems business, the lead times are longer. And so we do have backlog level of visibility for what we should be shipping in the quarter. And in the second quarter, you can look at the delta to the midpoint of guidance that we gave. It's about a $6 million delta, and we had about a $6 million delta of shipments that we expected, the Systems that we expected to ship that we didn't.

Got it. And maybe, Mike, kind of a bigger picture question, kind of tying the things that have been discussed already, but just cyclically, things are clearly a bit more depressed currently. And I think everyone expects in time that there will be some better improvement off the bottom. But that being said, over the midterm, say, 2 to 3 years, are you upbeat on IC unit growth similar to the past 5 to 10 years? And moving forward, do you think the complexity might have to do a little bit more of the heavy lifting to drive strong growth in the industry?

Yes, it's a really interesting question. I think over a 2- to 3-year time frame, we are optimistic about overall IC growth, IC revenue growth. If you look at the industry historically, it's always gone through these cyclical ups and downs, but imposed on that, a very strong long-term growth trajectory. And so over a 2- to 3-year time frame, we're confident that the industry is going to emerge from this cyclical downturn and that we'll begin to see IC unit growth and dollar growth. Layered on top of that, though, the degree of complexity and value in each of these IC units is pretty significant, right? If you think about some of the projects we talked about on the call, whether they're HBM, whether they're projects like Meteor Lake, the complexity in those and the value they provide is certainly going to be part of the revenue growth superimposed on top of the IC unit growth.

Our next question comes from Vedvati Shrotre of Jefferies.

I guess just from a long-term perspective. So on your prepared remarks, you talked about HBM adding complexity to your probe cards, the number of steps increase, but is there a point, as you know, like 3 or 4 years down the line as HBM stacking continues and you see more and more stacks of DRAM, do you think that could drive your margins on memory similar -- to similar levels as [Technical Difficulty], given that the complexity -- these probe cards are complex compared to what your DRAM probe cards will be?

It's a good question, and it drives to some of the underlying gross margin fluctuations that Shai talked about, whether it's increasing the value of what we do in a market like DRAM by making sure that we have a compelling differentiated products for new things like HBM, that certainly helps gross margins out in a segment that has been gross margin challenged over the past several years. We're also investing in R&D to improve our competitiveness and profitability with some of these platforms. So over that kind of time frame -- I think you said 3 or 4 years, over that kind of time frame, you can see both of those things at work. It's one of the reasons that we're confident in the progression back towards the 47% of the target model. And if you've seen FormFactor's long-term gross margin, gross margin improvement and gross margin evolution, you can see that that's a very important metric for us. Tough to overcome the volume reductions we're having in the current cyclical downturn, but in the area of both increasing value because of things like HBM and increased complexity, but also us driving our internal efficiency and gross margin improvement to deliver a bit of COGS structure in the gross margin line.

I would now like to turn the conference back to Mike Slessor for closing remarks. Sir?

Thanks, everybody, for joining us today. We've got, as we usually do at this time of the year, several investor conferences lined up here as we go through the late part of the summer, and hope to see you at some of those. Until then, take care.

This concludes today's conference call. Thank you for participating. You may now disconnect.