Q4 2023 FormFactor Inc Earnings Call
Participants
Stan Finkelstein; Investor Relations; FormFactor Inc
Michael Slessor; President & CEO; FormFactor Inc
Shai Shahar; Chief Financial Officer; FormFactor Inc
Brian Chin; Analyst; Stifel Nicolaus and Company, Incorporated
David Duley; Analyst; Steelhead Securities
Krish Sankar; Analyst; TD Cowen
Tom Diffely; Analyst; D.A. Davidson & Company
Christian Schwab; Analyst; Craig Hallum
Craig Ellis; Analyst; B. Riley Securities
David Silver; Analyst; C.L. King & Associates
Presentation
Operator
Thank you, and welcome, everyone, to FormFactor's fourth-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.
Stan Finkelstein
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 impacts of regulatory changes, including the recent US-China trade restrictions; the anticipated demand for products; our ability to develop, produce, and sell products; and the assumptions upon which such 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 the 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, February 7, 2024, and we assume no obligation to update them.
With that, we will now turn the call over to FormFactor's CEO, Mike Slessor.
Michael Slessor
Thank you, everyone, for joining us today on FormFactor's fourth-quarter earnings call. Formfactor's fourth-quarter results were in line with the outlook we provided last November. Compared to that outlook, moderately higher revenue and gross margins were offset by a higher tax rate, producing non-GAAP EPS at the midpoint of the range.
As we enter 2024, we continue to operate in an overall demand environment that remains similar to the levels we experienced during the past year, and we expect first-quarter results to be similar to those achieved in recent quarters.
The relatively stable aggregate demand across our combined served markets is a benefit of FormFactor's diversification strategy and sets us apart from our direct competitors. FormFactor has 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 relatively consistent top-line results as we've demonstrated for the past four quarters and which we expect again in the current quarter.
Customer investments in growing areas, driven by generative AI like high-bandwidth memory and co-packaged silicon photonics, is helping offset the impact of areas that are presently at cyclically low levels like mobile handsets and PCs and those entering cyclical downturns like automotive and industrial. This stability also enables FormFactor to keep investing in R&D for new product innovation and competitive differentiation, especially in our product roadmap for advanced packaging applications like chiplets and tiles, high-bandwidth memory, and co-packaged optics.
It also enables us to invest in capacity and other strategic initiatives designed to produce market share gains in above-industry revenue and profit growth when the industry returns to growth, which will enable FormFactor to achieve and then surpass the levels of our current target financial model. Because of the strength and stability of our balanced product portfolio, we can make these investments while maximizing quarterly profitability and protecting our strong balance sheet throughout prolonged periods of industry softness.
Turning now to segment level details. DRAM probe cards are an area of current strength, and we expect first-quarter DRAM revenue to approach the peak levels experienced in 2021. DRAM's strength is being driven by two factors. First, and perhaps surprisingly, we're experiencing strong demand generated by new DDR5 DRAM designs, as customers prepare for high-volume production when the eventual DRAM upturn arrives.
This provides insight into the unique characteristics of probe card demand since probe cards are a device-specific consumable customized to each individual chip design. Early production activity for these new DDR5 designs is driving demand for new probe cards, albeit at lower overall levels and with somewhat greater short-term volatility than we might expect to see in a full-scale cyclical upturn.
The second factor behind our strong DRAM outlook is the continuing acceleration in demand from multiple customers for probe card to test high-bandwidth memory, or HBM. Together with the headline-grabbing GPUs, HBM is a key enabler for generative AI, and we are forecasting nearly 50% sequential growth in our HBM probe card business in the first quarter.
HBM offers a great example of how advanced packaging is driving FormFactor's results. As we mentioned in the past, advanced packaging applications like HBM produced both higher test intensity, which expands the number of probe cards required per good die out and higher test complexity, which raises the performance requirements for each probe card.
Each HBM chip is a stack of 8, 12 or even 16 individual DRAM die assembled with advanced packaging processes such as through silicon vias and hybrid bonding. To ensure high yields of the stack DRAM chip, customers probe and test each component DRAM die prior to stacking and probe and test the multi-die DRAM stack at various points during the assembly process, leading to a substantial increase in the overall probe card intensity for good die out.
In addition, the technical requirements for HBM test are significantly more advanced than for standard unstacked DRAM products, involving higher test speeds and more challenging thermal scaling specifications. FormFactor's MEMS-based SmartMatrix 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 continues to grow, driven by the accelerating adoption of generative AI.
Shifting to foundry and logic probe cards, our largest business, we experienced the expected sequential reduction in the fourth quarter, as key customers digested the significant third-quarter shipments of probe cards for a major tile-based client PC design and two major high-performance compute designs in the foundry space. We expect similar demand levels for foundry and logic probe cards in the current quarter as customers continue to closely manage their output now that inventories have stabilized at healthy levels in high unit volume end markets like client PC, and mobile.
Looking further ahead, the adoption of advanced packaging processes like chiplets and tiles by major foundry and logic customers continues to accelerate as they seek to advance their roadmaps in the face of a slowing Moore's Law. As in high-bandwidth memory, advanced packaging in foundry and logic applications again demands both higher test intensity and higher test complexity. The positive impact of this trend, together with the potential for increased compute power in PCs and handsets to support AI at the edge, will drive significant growth in the foundry and logic probe card market.
As anticipated, we experienced a sequential reduction in systems segment revenue in the fourth quarter, mainly due to the sale of our FRT Metrology business. The systems business continues to be driven by strong demand for our market-leading test and measurement products for early development of applications like co-packaged silicon photonics and quantum computing. More broadly, systems is an important element of our diversification strategy, as spending in this segment is primarily driven by customer innovation and R&D budgets and does not directly correlate with semiconductor production activity.
Co-packaged optics enabled by silicon photonics remains an important and exciting driver for FormFactor's current systems and future probe card businesses. The transition of silicon photonics from early R&D to low-volume production continues, and we have now installed the CM300 silicon photonics system at the world's leading foundry to support low-volume production, as well as serve as a platform for the co-development of the enhancements needed for high volume electrical and optical test of co-package optics.
As production volumes increase over the next several years, our product road map delivers both systems and consumable probe cards to test these electro-optical devices and improve yields, enabling our customers to seamlessly transition from the lab to the fab.
Finally, earlier today, we announced an agreement to divest our subsidiaries in China to Grand Junction Semiconductor, together with a long-term exclusive distribution and partnership agreement. This transaction is designed to adjust our operational strategy in the region in light of export controls that caused our China revenues to decline over the past several quarters.
In closing, we continue to operate efficiently and prudently in what we see as a relatively stable demand environment across our diversified product and technology portfolio. 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 exciting innovations like HBM, 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 investments in R&D and capacity position FormFactor to emerge from the current cyclical downturn, a stronger and leaner competitor, enabling us to achieve our target model that delivers $2 of non-GAAP earnings per share on $850 million of revenue.
Shai, over to you.
Shai Shahar
Thank you, Mike, and good afternoon. As you saw in our press release, Q4 revenues and non-GAAP gross margin were at the high end of our outlook range and non-GAAP EPS was at the midpoint of the range. Fourth-quarter revenues were $168.2 million, a 2% sequential decrease from our third-quarter revenues and a year-over-year increase of 1.3% from our Q4 2022 revenues.
Probe card segment revenues were $126.8 million (sic - see Supplemental Financial Information, "$126.9 million") in the fourth quarter, a decrease of $1.5 million, or 1.3% from the third-quarter revenues, and a year-over-year increase of 1.9%. The decrease from Q3 was driven by lower foundry and logic revenues, partially offset by an increase in DRAM and flash revenues.
Systems segment revenues were $41.2 million in Q4, a $2 million decrease from third-quarter record revenues and essentially flat year over year. The decrease from the third quarter is mainly due to the sale of FRT, which closed on October 31. Systems segment revenues comprised 24.5% of total company revenues, slightly down from 25.2% in the third quarter.
Within the probe card segment, Q4 foundry and logic revenues were $83.6 million (sic - see Supplemental Financial Information, "$83.8 million"), a 13.3% decrease from the third-quarter revenues. foundry and logic revenues decreased to 49.8% of total company revenues compared to 56.2% in the third quarter.
DRAM revenues were $35.9 million in Q4, $8.5 million or 31% higher than in the third quarter, an increase to 21.4% of total quarterly revenues as compared to 16% in the third quarter. Flash revenues of $7.3 million in Q4 were $2.8 million higher than in the third quarter and were 4.3% of total revenues in Q4 as compared to 2.6% in Q3.
Gross margin for the fourth -- sorry, GAAP gross margin for the fourth quarter was 40.4%, same as in Q3. Cost of revenues included $2.8 million of GAAP to non-GAAP reconciling items, which we outlined in our press release issued today and in a reconciliation table available in the Investor Relations section of our website.
On a non-GAAP basis, gross margin for the fourth quarter was 42.1%, 0.3 percentage points higher than the 41.8% non-GAAP gross margin in Q3 and 1.1 percentage points above the midpoint of our outlook range. The increases compared to Q3 and the upside versus the midpoint of our outlook range were mostly a result of higher gross margin in the probe card segment, partially offset by lower systems segment gross margin.
Our probe card segment gross margin was 39.6% in the fourth quarter, an increase of 1.1 percentage points compared to 38.5% in Q3. The increase from Q3 is mainly due to improved factory utilization and lower excess and obsolete inventory expense, partially offset by a less favorable mix and lower volume.
Our Q4 systems segment gross margin was 49.6%, 220 basis points lower than the 51.8% gross margin in the third quarter. The decrease relates mainly to the sale of FRT during the quarter, which resulted in a less favorable mix and inventory adjustments related to the transaction close.
Our GAAP operating expenses were $59.6 million for the fourth quarter as compared to $66.6 million in the third quarter. The main reasons for the decrease versus Q3 were non-recurring M&A expenses in the third quarter and only one month of FRT OpEx in the fourth quarter, partially offset by higher performance-based compensation in Q4.
Non-GAAP operating expenses for the fourth quarter were $51.6 million, or 30.7% of revenues, as compared with $54.5 million, or 31.8% of revenues in Q3. The decrease relates to similar reasons I mentioned regarding GAAP operating expenses. Company non-cash expenses for the fourth quarter included $9.3 million for stock-based compensation, $1.6 million lower than in the third quarter due to timing of grants; $0.8 million for the amortization of acquisition-related intangibles, $0.5 million lower than in Q3; and depreciation of $7.7 million similar to the third quarter.
GAAP operating income was $81.3 million for Q4 and includes a $73 million gain from the sale of FRT, compared with GAAP operating income of $2.7 million in Q3. The gain from the sale of FRT has been excluded from our fourth-quarter non-GAAP results.
Non-GAAP operating income for the fourth quarter was $19.1 million compared with $17.3 million in the third quarter, an increase of $1.9 million or 10.8%. GAAP net income for the fourth quarter was $75.8 million, or $0.97 per fully diluted share, compared with a GAAP net income of $4.4 million, or $0.06 per fully diluted share in the previous quarter. The main reason for the increase is the gain on the sale of FRT.
The non-GAAP effective tax rate for the fourth quarter was 21.2% as compared with the 12.2% in the third quarter, mainly due to usual year-end adjustments and changes in mix of foreign versus domestic taxable income. For the full fiscal 2023, the non-GAAP effective tax rate was 15.6%, within the range of 13% to 17% previously communicated and similar to the 15.4% in 2022. We estimate that our annual non-GAAP effective tax rate for 2024 will be between 14% and 18%.
Fourth-quarter non-GAAP net income was $15.7 million, or $0.20 per fully diluted share, compared to $17.3 million were $0.22 per fully diluted share in Q3.
Moving to the balance sheet and cash flows. We used $0.3 million in free cash flow in the fourth quarter compared to generating $16.9 million in Q3. The main reasons for the change were timing of shipments to and collection from customers during the fourth quarter and higher capital expenditures.
We invested $9.9 million in capital expenditures during the fourth quarter compared to $5.9 million in Q3. This brings our 2023 annual capital expenditures to $56 million, within the range previously communicated. The decrease in CapEx in the second half of 2023 as compared to the first half and as compared with the $65.3 million invested in 2022 is due to completing the majority of the long lead time facilities and equipment investments required to reach our target financial model. Accordingly, we expect CapEx in 2024 to be between $35 million and $45 million.
At quarter end, total cash and investments were [$332 million], an increase of $84 million from Q3. The increase was mainly a result of approximately $104 million received from the sale of FRT, partially offset by the $9.9 million of capital expenditure I just described, and by stock buyback in the amount of $20 million. As of the end of the fourth quarter, we had one term loan remaining with a balance totaling $14 million.
Regarding the stock buyback, as I mentioned, during the fourth quarter, we purchased $20 million worth of shares to fully utilize our $75 million two-year buyback program that was approved in May 2022 and started utilizing the new $75 million two-year plan that was approved in October 2023. As of Q4 quarter end, $73.8 million remain available for future repurchases under this plan. As a reminder, the main purpose of the share repurchase program is to offset dilution from stock-based compensation.
Turning to the first-quarter non-GAAP outlook. We expect Q1 revenues of $165 million, plus or minus $5 million. At the midpoint of our outlook range, Q1 revenues is expected to be approximately $3 million lower than in Q4. The expected decrease relates to lower systems segment revenues, mainly due to the sale of FRT in Q4.
And in the probe card segment, we expect flat foundry and logic and lower flash revenues in the fourth quarter, partially offset by an increase in DRAM revenues. The sale of our Chinese subsidiaries that we announced today is expected to close in the first half of 2024 and it is not expected to have a significant impact on our financial results.
First-quarter non-GAAP gross margin is expected to be 41%, plus or minus 150 basis points. The expected decrease at the midpoint of this range as compared to Q4 '23 gross margin is a result of a less favorable product mix, mainly the expected decrease in system segment revenues and the increasing DRAM revenues.
Although we expect HBM revenues to grow in the first quarter, non-HBM DRAM revenues, which has a lower gross margin profile, is also expected to grow. At the midpoint of this outlook ranges, we expect Q1 operating expenses to be $53 million, plus or minus $1 million. The increase as compared to Q4 is mainly a result of the annual benefits reset.
Non-GAAP earnings per fully diluted share for Q1 is expected to be $0.19, plus or minus $0.04. A reconciliation of our GAAP to non-GAAP Q1 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?
Question and Answer Session
Operator
(Operator Instructions) Brian Chin, Stifel.
Brian Chin
Hi there. Good afternoon. Thanks for letting us ask a few questions. Maybe first, and I know, Shai, you just said that upon closure in first half, the China investor kind of has no impact. But I mean, I mentioned there are some financial implications, either in terms of cost and margins, some expectations as it relates to revenue, even maybe some cash on the balance sheet.
Can you maybe break that down a little bit more and also whatever regulatory hurdles or other hurdles need to be cleared to consummate the transaction?
Shai Shahar
Sure. So as I said, we don't expect it to have a significant impact on our financial results. When it comes to the balance sheet, once the deal closes, we will make the appropriate disclosure with more details. But since the details of the transaction were not disclosed, you can imagine that we concluded it's immaterial.
When it comes to the P&L, again, not a material impact. We maintain -- we continue to maintain the direct relationships with the multi-national customers in the region. And with respect to the local China customers, we have the new distribution agreement, and we expect and hope that this will continue to grow our China business as we move forward. And that's the conclusion; we don't expect it to have material impact.
And Mike, do you want to refer to the regulatory?
Michael Slessor
Yeah. Brian, this has been structured in a way where we don't believe there are significant regulatory approvals required. Basically, what we're doing is divesting what's in China already. And so there's no technology transfer, no significant technology licensing. It really is a focus and a shift to our operational and customer channel strategy in the region, given some of the challenges we've had with the export control.
You've seen our domestic China revenue contract over the past several quarters, and this is a response to that. We are hopeful that Grand Junction can bring some more resources and focus to those business, and through the distribution channel, help grow our domestic China business for us.
Brian Chin
Okay. Yeah, it's helpful. And maybe for my follow-up, can you quantify -- I guess it's been operating maybe around $10 million a quarter, but what the HBM contribution was to the $36 million year end in Q4?
And then also kind of more broadly, do you see wafer probe card test intensity to continue to increase alongside your customers' roadmaps based on more die stacking, hybrid bonding, et cetera? Or do you see test intensity leveling off at some point as customers' yields and manufacturing efficiency improves?
Shai Shahar
Yeah, I think it's a question of timing on the second one. I think for the medium term, certainly through 2024 and into 2025 with HBM, customers are still riding up a pretty significant yield learning curve. And so the increase in test intensity that we've talked about associated with die stacking in advanced packaging of 20% to 30%, I think we're pretty confident is going to hold here as we go through the early part of the HBM3 and into the HBM4 ramp.
Over time, as with anything, we see this in new nodes as well. Customers accomplish the yield learning using products like ours, reducing the overall test intensity, improving their yields. And so we would expect it to go down over time. But I think we're pretty confident in increased test intensity as all the different yield reduction modes associated with die stacking are worked through in HBM and in the logic space as well.
On the numbers, HBM in the fourth quarter was approximately $10 million, as you said, maybe a little bit more of the $36 million of DRAM revenue. And as we said in the prepared remarks, we expect that to increase around 50% sequentially. So up into the, call it mid-teens in the Q1 outlook.
Brian Chin
Okay. Great. Thank you.
Shai Shahar
Thanks, Brian.
Operator
David Duley, Steelhead. David Duley, you might have your phone on mute.
David Duley
Yeah, you're right. Thanks. I noticed the foundry and logic revenue was down and it looks it's similar decline that we saw in the revenue out of geographic region of Taiwan. I was just wondering if you can make some more elaborate comments about that? That's my first question.
And my second question is could you remind us what percentage of the overall DRAM market uses MEMS-based probe cards? And I think you mentioned that all the high-bandwidth memory stuff is going to be on MEMS. Does the growth in high-bandwidth memory really benefit you guys since you're a MEMS-based provider?
Michael Slessor
Yeah. I'll take the second one first. The DRAM wafer probe market is almost entirely MEMS-based. Us and our primary Japanese competitor in high-end DRAM, really the only way to build these probe cards, which tests the entire wafer at once and have approximately 100,000, 150,000 individual contacts on them. The only way to do that in a cost-effective and high-quality way and deliver the speed performance, the high-power performance is with immense technology.
And so I don't view HBM in and of itself as a headwind or a tailwind for MEMS adoption, considering we're already largely adopted with MEMS and DRAM. But the requirements, as we talked about in the prepared remarks, are more stringent. And so it increases, one, the value to our customers. It also increases our competitive differentiation on things like speed -- high-speed performance and thermal scaling.
And so it's not purely a MEMS adoption story, but really a raising of the bar of the existing MEMS probe card for HBM, which as Shai commented, helps our gross margins in the segment a little bit.
On the foundry and logic observation, if you recall back on our previous call, we had a very strong third quarter with some of our major foundry and logic customers, and in particular in the foundry space, delivered high volumes of two high performance -- probe cards for high-performance compute projects. And those are in a digestion mode right now.
So the inference you've drawn is correct. It also ties to the expectations we said on the previous call about some digestion in the foundry space as we go through Q4 and now Q1.
David Duley
And just as my follow on, you talked about in your prepared comments, I think someplace about you saw unusually strong demand in DDR5. Is that associated with a specific end market? Or could you just maybe just elaborate further on that particular comment? Thank you.
Michael Slessor
Sure. The overall DRAM guidance, as we said, our outlook is that we're going to approach the levels that we approached in the middle part of 2021. So somewhere in the low-$40 millions of quarterly DRAM revenue. And that's composed -- you can do the math from the previous question-and-answer session. You can do the math on how much of that is HBM, which leaves you with some pretty solid growth associated with non-HBM memory.
But yeah, we're seeing DDR5 for both mobile and server/PC applications be pretty strong in the first quarter. I want to be clear that we're not calling a DRAM upturn at this point. It's pretty concentrated on a couple of designs and mostly at one customer. But it is an interesting indication of our customers beginning to invest in new designs because we know DRAM is cyclical and we know it's going to turn at some point. So this represents some investments in them getting ready for the ramp when it does turn.
David Duley
Thank you.
Operator
(Operator Instructions) Krish Sankar, TD Cowen.
Krish Sankar
Yeah, hi. Thanks for taking my question. I have two of them. First one, if I look at your guidance, I'm just curious about -- Mike or Shai, you said foundry and logic revenue is flat, HBM grows 50%, and overall DRAM revenue is approaching peak levels in Q2 of '21, which is $42 million. And DDR5 is growing too.
But if I do the math, it looks like HBM as a percentage of DRAM revenues is going to be higher in March versus in the December quarter. So why is gross margin still lower? Shouldn't gross margin be better than December then?
Shai Shahar
Well, let's not forget the systems segment, right? I did talk about the call that systems segment is expected to go down in Q4. And as you recall, our systems segment gross margin is historically higher. So we're going to have less favorable mix between systems and probes. And that's another driver that pushes gross margin a little lower than Q4.
Krish Sankar
Got it. Got it. And then I just had two other quick questions. Mike, if we look at the HBM share, you've clearly done a great job in HBM3 3E versus HBM1 and 2. I'm just curious, do you expect the momentum to continue next year or so when we get into HBM4 and beyond?
Michael Slessor
Yeah, Krish. I think we see the same progression of test complexity. That really shifts competitive advantage towards us as we move from HBM3 to 4. I think one of the obvious ones is it's a higher -- more die being stacked. And that clearly, as we look at customers test strategies, as I said in the prepared remarks, they're not just probing each individual component die, but they're probing the stack as they build it up. And so higher stacks probably mean less yield, but also probably mean higher test complexity, higher test intensity as we move to HBM4.
In addition, obviously, the clock speeds go up, the densities go up. And so those are all good trends. And I think it's one of the things where we're spending a good fraction of our R&D budget on making sure we can continue to drive that capability roadmap forward and serve the increased complexity of HBM4 and derive some competitive advantage from it.
If I go back to your initial comment that Shai answered, actually the mix between regular DRAM and HBM DRAM in the low-40s is nominally the same as it was in Q4. So that's one of the reasons why you're not seeing a margin uplift associated with DRAM.
Krish Sankar
Got it. Got it. Another quick clarification, I think Brian asked this on the China, of the divestment. To the multi-nationals in China, is that still going to go through the Grand Junction stemming, or is it going to be directly from FORM?
Michael Slessor
No. Yeah, no. Because the multi-nationals that operate in the region are global customers for FormFactor, they are not part of this distribution agreement. FormFactor will continue to support those customers in a global way; and where we need local resources, we'll contract those things out. But you can think of the multi-nationals, which historically has been the vast majority of our China revenue, continue to be managed by FormFactor and are not part of the distribution arrangement.
Krish Sankar
Got it. Thank you very much, Mike. Thank you.
Michael Slessor
Thanks, Krish.
Operator
Tom Diffely, D.A. Davidson.
Tom Diffely
Yeah. Thank you and good afternoon. I guess first of all, when you look at the memory thing, you talked about some next-generation DRAM chips. Are you seeing anything on the flash side, any next-generation flash chips that require new designs?
Michael Slessor
Tom, as I think you know, our flash market share and consequently our flash revenue is pretty minimal compared to DRAM. It's a market where the requirements are not that challenging from a speed or density perspective, which is really where we derive competitive advantage and why we're strong in DRAM.
Flash, because of die sizes, because of test speeds, because of things like built-in self-test of the chips, really hard to drive a competitive advantage. So we do have some share there. But as you can tell from our historical financial results, it's not significant. We do see some new designs, but given my statement about our market share, we don't have the broad view of the flash market that would allow us to make the same kind of observations we can in DRAM.
Having said that, I think the general view when we talk to customers and other suppliers is that flash is probably going to trail DRAM in terms of recovery, that inventories are still high and pricing is not where it should be to drive an upturn. But we are continuing to be opportunistic in the flash market, and where we can differentiate with speed and/or capability, we're taking advantage of them.
Tom Diffely
Okay. And would you expect your flash share to go up over time, given the acquisition you made a couple of years ago of some technology there? Or is it kind of in this segment or in this range for a bit?
Michael Slessor
Yeah. I think it's in this range for a bit. So the acquisition you're talking about, we acquired Advantest probe card businesses a few years ago or business a few years ago. And if you remember, we primarily did that, although it is a flash test platform, a flash probe card platform. We primarily did it because it feeds elements of our DRAM roadmap that we're now in the final qualification stages at with a key customer. That will help us from an overall market share perspective, at least at that customer. So we, again, acquired it more for the DRAM capability, although being engaged in flash does help you work out the bugs at a lower complexity level.
Tom Diffely
Great. Thanks. And the follow-up, Shai, when you look at the target model going from where we were in 2023 to the target model, like 700 basis points of margin improvement, is most of that just coming from overhead absorption or there are some cost reduction programs that need to take place?
Shai Shahar
I would say three things. One is volume. Yes, as you pointed out, we have been adding capacity. So we need volume to go up to fully utilize this capacity and improve absorption and utilization.
The second element is mix. We expect the growth from where we are today to our target model to come mostly from foundry and logic, which is an area that has a higher gross margin.
And the third component is the cost reductions. We have few initiatives across the company and some task forces to improve our gross margins, focusing on costs, things like automation and vertical integration, more efficient production. And these things also expect to contribute on our way to reach the target model of 47%.
Tom Diffely
Okay. Would you say that overhead absorption is the biggest factor?
Shai Shahar
I would say it's distributed equally between the three components.
Tom Diffely
Okay. That's helpful. Appreciate your time today. Thank you.
Shai Shahar
Thanks, Tom.
Operator
Christian Schwab, Craig Hallum Capital.
Christian Schwab
Great. Mike, on the DRAM business and the 50% sequential revenue growth, is that driven by your current lead customer or is that the expansion starting to ramp with the second customer?
Michael Slessor
Yeah. When you break down the increase, it really is driven by our lead HBM customer, who also leads the overall HBM market in share. So that's the primary driver.
I did -- it's probably a subtlety, but I did in the prepared remarks say that we expect revenue from all three major DRAM manufacturers in HBM. So there's engagements for those. But when we think about the Q1 step up, it's very much driven by our lead customer. It's a key 2024 strategic objective for us to make sure that we capitalize on HBM growth that all major DRAM manufacturers. And I think we're making some pretty good progress with that.
Christian Schwab
Once you have all three, how big could the high-bandwidth memory portion of the business be?
Michael Slessor
Yeah, it's a good question. So I think we'll stick with the idea that our HBM revenue in 2024 on a quarterly basis could double from the quarterly run rate that we had in 2023. And the reason for that is it's a little difficult right now to sort out how much double-counting is going on in terms of forecasting between the different suppliers of HBM DRAM, our customers. I think if you back out the forecast and add them all together, the growth assumptions are probably there's some double-counting in there.
So I think the expectation that we communicated on the last call and through some of the investor conferences as we've gone through the back part of the year and to start 2024 of the HBM run rate doubling on a quarterly basis from where it was in 2023, the right expectation. And then you can see in Q1, we've made some good steps towards that.
Christian Schwab
Okay. That's great. And the foundry and logic business, leading manufacturers of chips, whether it's Intel or AMD or TSM, are kind of pretty optimistic regarding aggregate unit growth and the huge volume markets of PCs and smartphones. They're just not all that enthusiastic about it in March, but they are as we go throughout the course of the year. And given some of the new design ramps and new chips coming to market throughout the course of the year, is it logical that that business exits the year at a $100 million-plus run rate again?
Michael Slessor
Yeah, I think you've identified the key drivers for it. Certainly, in the first part of the year, whether you look at our Q1 guidance, extrapolate what some of those customers in the high-unit volume markets like PCs and handsets are conveying, there's some optimism associated with the second half, given our visibility, right? Remember that probe cards have lead times well within a quarter, much less than a quarter. And so our direct visibility doesn't extend anywhere close to that.
But I think the combination of inventories being at much healthier levels and probably now a balance between output at these key customers and the actual sell through, that's now been resolved. So that puts things in a better place. And I think if there's some in-demand pick up in these large unit volume markets like PCs and handsets in the second half, we should see some growth there, but it's beyond our visibility, beyond the headlines. What I would say to close that topic is when I compare things to a year ago and people were optimistic about a second-half recovery in 2023, I feel like the industry is in much better shape from an inventory and production and capacity standpoint a year later here in the first part of 2024.
Christian Schwab
Great. And then I'll just slip in one more question if I can. You've been working on the qualification with a potential large customer for some time, I think a few quarters anyway, I guess. Is there any update on where we sit with that? And do you feel better about that? Do you feel better about potential timing of that? Any update?
Michael Slessor
Yeah. We don't have a huge update for you, but we have talked about this in the past. Obviously, the large fabless microprocessor and GPU customer is a place where we do not have significant share. And given our strategy of diversification, different customers, different markets, that's an important place for us to qualify and hold a significant share position, especially given their leadership in some of the end markets.
We have made some -- as you know, we've been at this for a while. We have made some changes in how we're resourcing that. The good news is the customer has a clear two-suppliers strategy. And there's really only two suppliers that can do this, and FormFactor is one of them.
So it's really up to us to improve our execution and get there. I'm confident that we can qualify at least in 2024. That will take a while to build into a significant share and revenue position, but we're setting the groundwork now, and it's a very important objective for some key people at FormFactor in 2024.
Christian Schwab
Great. No other questions. Thank you.
Michael Slessor
Thanks, Christian.
Operator
(Operator Instructions) Craig Ellis, B. Riley Securities.
Craig Ellis
Yeah, thanks for taking the questions. So, Mike, hard to believe there could be even more DRAM questions, but there was some new ones I wanted to better understand. As we look at high bandwidth memory here in the first quarter, understandable that it's ramping given what's going on more broadly. But the question is this, as you look at the mix of your high-bandwidth memory exposure in 1Q from high-bandwidth memory 3 to potentially 3E, which I think really launches commercially on a system in 2Q, and high-bandwidth memory 4, which I think everybody expects to come out in early 2025. How much of the mix in 1Q is 3 versus 3E and 4 and to the extent we're not really generating revenue on the other two, when would you expect that to hit?
Michael Slessor
Yeah, there actually is a little bit of revenue on 3. So most of the production volume, as you'd note, is on third generation or HBM3. But 3E, the mini step, we are shipping probe cards in some significant volume. I don't have that mix off the top of my head, but I guess it's something like two-thirds HBM3, one-third HBM3E, and then a sprinkling of HBM4 in early R&D shipments.
Craig Ellis
Yeah. And then part of that, just so I don't offend the moderator with too many questions, there's growing chatter that in PCs, we could see a format change from [SO-DINS to LP-CANS]. And I don't know if that has any impact on probe card intensity, but are you seeing anything on the personal computer side that would drive up probe card intensity that's AI related?
Michael Slessor
I don't think so, Craig. I think those format changes, especially that's two steps downstream from us in the integration when they get integrated into the package and then into the module. I don't anticipate that having -- and we, of course, interact with the die at the wafer level. I don't anticipate that would have a significant impact. What we really need is a PC refresh cycle to drive unit volume increases of PC.
Craig Ellis
Yeah, I hear you on that one. And then for the second question, typically, I think in the first half of the year, the business within foundry and logic would start to benefit and have a sine wave relative to an APU program at a large foundry. Is everything on track there? And do you still expect to hold the same share that you've historically had there?
Michael Slessor
Yeah, I think so. It's going to be an interesting year with -- in the foundry business, a lot more high-performance compute than just the application processor, although I think the application processor projects in the foundry space generally continue to be big middle-of-the-year drivers of revenue. Those are key customers and very important focus areas for us as we move through 2024.
They're one element of the overall diversification strategy we talked about, right? The DRAM's running pretty hard right now; HBM, not so surprisingly; but DDR5 being, as I said in the prepared remarks, a bit surprising given where we are in the DRAM cycle. But it's exactly a case study of why we've put these different businesses together to try and produce a relatively stable top-line demand profile, so that we can weather this sub cyclicality and continue to invest in things like R&D for competitive differentiation and capacity.
Craig Ellis
Got it. Thanks, Mike. Appreciate the help.
Michael Slessor
Thanks, Craig.
Operator
David Silver, C.L. King & Associates.
David Silver
Yeah. Hi, thank you. I'd like to ask maybe a strategic question regarding capital deployment at this point in time. So, Mike, with the sale of the metrology business, I mean, you kind of had to make a strategic decision not to invest to grow that business. And I don't believe the Livermore capacity -- you have some extra capability at Livermore now. So when I look at your balance sheet, I mean it's certainly highly liquid.
Could you just maybe discuss some of your thoughts about the highest and best use for that, what you would call may be incremental or excess portion of your cash? In other words, does vertical integration either forward or back make sense? Are there parts of your global footprint that you could profitably expand? Or would there be something from the equity investment that your large competitor received recently that maybe requires a countermove or some kind step to retain or maintain your competitive balance?
But sitting here with a highly liquid balance sheet, what goes through your mind in terms of optimizing the use of at least a portion of that?
Michael Slessor
Yeah, I'll let Shai talk through the details. But I think in an environment where we still expect consolidation in the space, we don't think holding onto some cash, especially with decent returns, is a bad strategic move. We're going to need dry powder for M&A. We continue to be believers in M&A and consolidation, not necessarily inside our served segments, but to expand our served available market. And cash is a very useful asset to have.
So with that, I'll turn it over to Shai to talk a little bit about capital allocation priorities.
Shai Shahar
Right. Yeah. So as Mike mentioned, our priorities really didn't change compared to what we communicated before. We're focusing on reinvesting in R&D. We're still investing in capacity, although we slowed it down recently, things like vertical integration that you mentioned and automation.
And we have the share repurchase program that I talked about. We still have $75 million to invest in share repurchase, which is to offset dilution, and it's a two-year plan and it's been one quarter since we announced. So we still have one year and three quarters to do that, and M&A.
And with the current interest rates and cost of capital, accumulating cash is aligned with our M&A strategy. If you look at our historical acquisitions, the MicroProbe acquisition in 2012, the Cascade Microtech acquisition back in 2016, we've been pretty good in creating value through our M&A activity and has been good steward of capital.
David Silver
Okay, that's great. And I'll stop there. Thank you very much.
Michael Slessor
Thanks, David.
Operator
(Operator Instructions) David Duley, Steelhead.
David Duley
Yeah. I was just curious, are you comfortable with unit volume -- IC unit volume growth in the 6% to 8% range for 2024 for the foundry and logic segments? And if that's the case, do you guys think you can grow faster than that or slower than that as far as the market growth?
Michael Slessor
Yeah. I think as we look, there's a lot of variance in different assumptions, specifically the second half of the year to drive growth. I think where we're more focused in our planning and forecasting and resourcing activities is on the first half, because I think there's enough different variables where, sure, you could see 6% to 8% growth, you could see more, you could see less.
And as I think most of you know, our visibility really doesn't extend into the second half. If you were to have that scenario, I do think there is a very good likelihood that we can outgrow the unit growth of the market, especially if the bulk of it is associated with advanced nodes and some of the high-volume markets like PCs and mobile. And honestly, I don't see how you get to foundry and logic high-double digit unit growth without PC and handset growth in the second half of the year.
So in that scenario, yeah, I think it's reasonable given the adoption of advanced packaging, given the increase of test intensity, that we can outgrow the market. But I'm not sure I'd adopt that yet as a baseline scenario as we work our way through the first half.
David Duley
Fair enough. One final question. I'm just curious from your point of view where you sit, you've talked a lot about increased intensity from advanced packaging, all sorts of FormFactors and whatnot. And I'm wondering -- and that's generally been with Intel and TSMC, frankly.
I'm wondering if you're seeing that start to spread out onto the OSATs, because there's certainly a lot more chatter on their conference calls. And I think ASC's CapEx is up more than 50% year over year this year focused on advanced packaging. And so I'm just wondering, are you starting to see a ramp up and middle part of the market for advanced packaging? And are you well positioned to capture OSAT business? Thanks.
Michael Slessor
Yeah. So typically, the way the probe card market works is we have a very large installed base at the OSATs. But the OSATs often are not the purchaser of the probe cards. It's the fabless semiconductor manufacturers or the foundries themselves. And so there's a lot of interaction with the OSATs, but they're often not issuing purchase orders to us.
So I think for us, advanced packaging, if it happens at the OSATs, if it happens at the foundries, if it happens at the IDMs, all a good trend because it's going to drive up test intensity. And if a fabless customer has adopted advanced packaging and they're going to do it at foundry or at the OSAT, that's going to drive higher test intensity at either of those places no matter what. It's just a slightly different business model for us.
David Duley
Okay. Thanks.
Stan Finkelstein
All right. Thank you.
Operator
Krish Sankar, TD Cowen.
Krish Sankar
Thank for taking my follow-up. Mike, just a quick one. It looks like the high-density GPU customers are finally evaluating MEMS-based probe cards. How to think about when that eval will be done? When to expect first revenues from them? And if you can prognosticate, what do you think the market share between you and Technoprobe would be in that specific category?
Michael Slessor
Yeah. I think as the leading GPU manufacturer adopts advanced packaging, that's really the discontinuity where they adopt MEMS probes. And it's an interesting case because of their large die sizes, they haven't really needed to adopt MEMS probes. But as they move to the [COA's] advanced packaging technology, MEMS probes are an absolute must. And so that's an opportunity, as you know, for us and our primary competitor in foundry and logic.
Those qualifications are happening right now. I think the -- and pilot production right now. I'd expect over time that to be an application and a customer where shares in this two-supplier window. We typically talk about it as a two-third, one-third. Most customers like to balance at 50-50, but there's often places where either us or our competitor have some degree of differentiation. And so it never quite works out to be 50-50.
So I think you're going to start to see that in a significant way here in 2024. But of course, still the majority of silicon from that customer is not going to be on advanced packaging. That's a small, but very lucrative part of their overall product portfolio. I still expect monolithic GPUs, at least for 2024, to continue to stay on the legacy probe card technologies.
Krish Sankar
Got it. Got it. Thanks a lot, Mike. Thank you.
Michael Slessor
All right. Thanks, Krish.
Operator
Thank you. This does conclude the question-and-answer session of today's program. I'd like to hand the program back to Mike Slessor for any further remarks.
Michael Slessor
Thanks, everybody, for joining us today, and we'll talk to you again at some point in the spring, if not at our next earnings call. Take care.
Operator
Thank you, ladies and gentlemen, for your participation in today's conference. This does conclude the program. You may now disconnect. Good day.
