Q3 2023 Teradyne Inc Earnings Call
Participants
Andrew Blanchard
Gregory S. Smith; President, CEO & Director; Teradyne, Inc.
Sanjay Mehta; CFO, VP & Treasurer; Teradyne, Inc.
Brian Edward Chin; Associate; Stifel, Nicolaus & Company, Incorporated, Research Division
Joseph Lawrence Moore; Executive Director; Morgan Stanley, Research Division
Krish Sankar; MD & Senior Research Analyst; TD Cowen, Research Division
Mehdi Hosseini; Senior Analyst; Susquehanna Financial Group, LLLP, Research Division
Samik Chatterjee; Analyst; JPMorgan Chase & Co, Research Division
Steve Barger
Timothy Michael Arcuri; MD and Head of Semiconductors & Semiconductor Equipment; UBS Investment Bank, Research Division
Toshiya Hari; MD; Goldman Sachs Group, Inc., Research Division
Vivek Arya; MD in Equity Research & Senior Semiconductor Analyst; BofA Securities, Research Division
Presentation
Operator
Greetings. Welcome to the Teradyne Q3 2023 Earnings Call and Webcast. (Operator Instructions) Please note this conference is being recorded. I will now turn the conference over to your host, Andy Blanchard. You may begin.
Andrew Blanchard
Thank you, and good morning, everyone, and welcome to our discussion of Teradyne's most recent financial results. I'm joined this morning by our CEO, Greg Smith; and our CFO, Sanjay Mehta. Following our opening remarks, we'll provide details of our performance for 2023's third quarter along with our outlook for the fourth quarter.
The press release containing our third quarter results was issued last evening. We're providing slides on the Investor page of the website that may be helpful to you in following the discussion. Replays of this call will be available via the same page after the call ends. The matters that we discuss today will include forward-looking statements that involve risk factors that could cause Teradyne's results to differ materially from management's current expectations.
We encourage you to review the safe harbor statement contained in the earnings release as well as our most recent SEC filings. Additionally, those forward-looking statements are made as of today, and we take no obligation to update them as a result of developments occurring after this call.
During today's call, we'll make reference to non-GAAP financial measures. We've posted additional information concerning these non-GAAP financial measures, including reconciliation to the most directly comparable GAAP financial measure where available on the Investor page of our website. Looking ahead, between now and our next earnings call, Teradyne expects to participate in technology or industrial-focused investor conferences hosted by Baird, UBS and Wolfe Research.
Now let's get on with the rest of the agenda. First, Greg will comment on our recent results and the market conditions as we enter the fourth quarter. Sanjay will then offer more details on our quarterly results, along with our guidance for the fourth quarter. We'll then answer your questions, and this call is scheduled for 1 hour. Greg?
Gregory S. Smith
Thanks, Andy, and good morning, everyone. Today, I will summarize our Q3 results, describe the current business conditions and provide some insight on how we're thinking about 2024 and beyond. Sanjay will then provide the financial details on Q3, our outlook for Q4 and offer some comments on the modeling next year.
Third quarter sales and earnings were at the high end of our guidance range as robotics sales came in above plan, and we cleared some supply constraints and tests. The second half of 2023 is playing out as we described in July. We expect to close out the year with strong robotic shipments amplified by new product shipments at UR and seasonally softer test shipments.
In semiconductor test, the mobility correction cycle persists and shipments remain well below historic levels, while our automotive test shipments remain high in Q3. Memory test shipments in Q3 were down sequentially due to the timing of shipments, but demand remains strong. LPDDR5 and HBM, both of which require higher speed testers drove the results.
In Wireless, demand remained muted in the quarter given the weak smartphone market and lack of new wireless standards this year. In System Test, defense and aerospace and storage test groups were on plan, while production board test softened in the quarter. Robotics demand has stabilized. In the first half of 2023, demand was quite low down 21% versus the first half of 2022.
In Q3, demand strengthened with revenue nearly at 2022 levels and up 20% from Q2. Our shipments have stepped up as we execute an aggressive ramp of the new UR20 product and PMI seems to have stabilized a bit. Looking at the full year of 2023, our estimates of the SOC test market size are unchanged at $3.7 billion to $4.1 billion, down about 15% at the midpoint from last year.
The weakest segment of SOC is mobility, down about 40% from 2022 on slower complexity growth in smartphone semiconductors and year-on decline in units. The compute Automotive and Industrial analog segments of the market will finish the year at similar levels to 2022. In Memory Test, we expect the market will be at the low end of the $900 million to $1 billion range, also unchanged from our July view.
The demand for high-speed DRAM test remains high as we close out 2023, which will help us pick up a few points of memory share for the full year. Teradyne's System Test group will finish 2023 down more than 20% from 2022. Within this group, we expect Defense and Aerospace will grow 10% year-on-year as global defense spending ticked up. The other segments of the business were negatively impacted by oversupply in the HDD market and mobility weakness.
Shifting now to Robotics. The macro environment for industry is incrementally better than last quarter with global PMI stable or improving slightly. The highlight of the quarter was a well-executed volume ramp of our UR20 collaborative robot. We delivered more than 300 units in Q3, and we expect to deliver a multiple of that in Q4. The UR20 extends UR's ease of use and quick ROI to higher payload and longer reach applications, expanding the market in many segments.
The strongest segments for UR20 so far are welding and palletizing. The distribution channel transformation that we described in past calls is also making steady progress. Complementing our existing distribution channel with direct coverage of large accounts and adding OEM partners is a long-term project, and we're beginning to see positive benefits.
For example, in the OEM space, we have added 48 new OEM partners so far in 2023, bringing the total to 144. And we have seen direct OEM orders grow nearly 20%, driven by the high demand from the palletizing market. At MiR, our account strategy continues to deliver with our top 10 customers expanding their collective installed base by over 15% year-to-date. A rate that's more than 50% greater than the overall installed base growth.
Shifting to the future, I'd like to describe our current thinking about 2024. Please bear in mind that it is still too early and visibility is too limited to be certain about what will happen next year. However, there are some longer-term trends that we expect to play out. As we've previously discussed, we expect the SOC market and our revenue to grow from 2023 on broader 3-nanometer adoption in the mobility space, driving market growth and continued strength in the compute market.
The real question is the magnitude of the mobility recovery, which depends on smartphone unit growth, complexity growth and how quickly the industry can consume idle test capacity. For reference, we estimate subcon tester utilization is still low, up only marginally from our July estimate and well below the typical Q2 to Q3 increase. The automotive test market has been sustaining at a higher level in 2023 than we originally expected. It appears that channel inventories in Automotive are stabilizing, and we have seen some spot weakness in the market.
We aren't expecting a significant change in the full year market size next year as unit forecasts and semiconductor attach rates driven by the crossover from internal combustion to EV remain bullish. The technology buys that have supported the memory TAM in 2023 should continue into next year, and we expect the memory market to grow as HBM, DDR5 and LPDDR5 penetration expands to support AI and computing growth.
In flash, as protocol interface speeds continue to increase, we expect flash package test demand to grow as well. Overall, we expect the total ATE TAM to be up modestly from this year, and the key factor is the strength of recovery in mobile. Growth in our wireless business, LitePoint will be strongly linked to handset growth, a recovery in the PC market and the start of the rollout of WiFi 7. The supply-demand imbalance in HDD is likely to persist through 2024, and we expect HDD test to remain [leap].
System-level test will depend largely on smartphone unit growth in the near term while we expect our defense and aerospace business to grow in 2024 on increased defense investments worldwide. In robotics, we're finishing 2023 on a positive note in a tough market with Q4 revenues up about 10% year-on-year on the strength of the new UR20 product introduction. That performance reinforces our optimism in robotics. We see robotics as a marathon, not a sprint. We are serving in an emerging market of $2 billion this year that we expect to grow to tens of billions of dollars per year in the future.
Our operating model for robotics is built for that marathon with a strategy that prioritizes product and support investments that deliver value to customers now. We are counting on building relationships with those paying customers to help guide our ongoing investments to meet their evolving needs for the future. The key to this strategy is driving towards our model of 5% to 15% profit from our robotics portfolio. While we will fall short of this objective in 2023, it remains a key operating metric for 2024.
We do this to ensure that we remain focused on our customers' most important automation priorities while we've grown the business. Rolling it all up, 2024 looks to be stronger than 2023. With all of the uncertainty around chip inventories, low utilization rates and macroeconomic worries. I'd call it incrementally stronger, but we'll get a better view over the next quarter or so.
We're also assuming a quarterly revenue profile in 2024, similar to 2023, with Q1 as the low point and then growth from there. While early, we're modeling Q1 sales to be similar to Q1 '23. As we finish the year, I'm encouraged by indications that our largest market, Semi Test, appears to have troughed in 2023 at a level that delivers an operating profit of 20% for the total company.
We are confident about the long-term growth outlook of the semiconductor market as the substantial fab equipment investments made in the recent past have not yet seen matching test investments. Also, we see consistent investment in tooling to enable continued process development, whether it is building a family of process nodes at 3-nanometer or enabling gate all around and 2-nanometer technologies. While the timing of test investments will be driven by end market chip demand and complexity growth, we are confident that this investment will happen.
To be clear, our customers are still cautious about their near-term demand, and we're reflecting that caution in our initial outlook for next year. But long term, there is significant upside potential. In Robotics, we have a pipeline of new products, new applications and distribution changes that are now beginning to yield. At the end of the day, the global population trends are inarguable.
The long-term demand for advanced automation must grow to deal with the increasing shortage of manufacturing workers, that coupled with market conditions that favor low-cost, short ROI automation investments and our team's growing execution skill. I expect renewed growth in robotics in 2024 as well. With that, I'll turn things over to Sanjay for the financial details. Sanjay?
Sanjay Mehta
Thank you, Greg. Good morning, everyone. Today, I'll cover the financial summary of Q3, provide our Q4 outlook and update you on our supply chain and resiliency progress. Now Q3. Third quarter sales were $704 million with non-GAAP EPS of $0.80, both at the high end of our guidance as robotics delivered above plan and some supply constraints eased in test.
Non-GAAP gross margins were 56.6%, in line with our guidance. Non-GAAP operating expenses were $243 million, down about 3% from the second quarter on spending controls and lower variable compensation. Non-GAAP operating profit rate was 22%. We had two 10% customers in the quarter. The tax rate, excluding discrete items for the quarter was 14.5% on a GAAP basis and 15.7% on a non-GAAP basis.
Semi Test revenue for the quarter was $498 million with SOC revenue contributing $404 million and memory $94 million. As noted earlier, we continue to see strength in SOC concentrated in auto end market and image sensor parts of our business in the quarter. Memory sales continue to be weighted towards technology retooling for higher-speed protocol flash for smartphones, DDR5 and HBM DRAM for server applications.
System Test Group revenue was $83 million with $38 million in storage test as SLT and HDD production demand remains [muted]. In Wireless Test, revenue was $37 million in Q3, with low demand from both PC and smartphone end markets. We expect this market trend to continue over the next several quarters. Now to Robotics. Revenue in Q3 was $86 million with UR contributing $71 million and MiR $15 million, which was above plan, as Greg noted.
Shifting back to the company level financials. Our free cash flow was $140 million in the quarter, and we returned 97% to shareholders. We repurchased $119 million of shares in the quarter, paid $17 million in dividends and settled $9 million of debt. We have the final $24 million of convertible debt, which will be repaid in the fourth quarter. We ended the quarter with $820 million in cash and marketable securities.
Now to our outlook for Q4. Q4 sales are expected to be between $640 million and $700 million, with non-GAAP EPS in a range of $0.61 to $0.81 on 162 million diluted shares. Fourth quarter guidance excludes the amortization of acquired intangibles, restructuring and other charges. This outlook is in line with our July view at the company level for the second half. Our revenue guide for Q4 has no material supply constraints. As supply has become more in line with demand, we are now back to including normal supply issues in the revenue range.
As a result of supply and demand coming into balance, our lead times continue to improve. This enables customers to place orders more in line with their incremental production requirements. In Q4, there is a component of this behavior. But we expect to see more book ship variability in Q1 as lead times continue to be reduced.
Fourth quarter gross margins are estimated at 56% to 57%, OpEx is expected to run at 35% to 38% of fourth quarter sales in line with Q3. Non-GAAP operating profit rate at the midpoint of our fourth quarter guidance is 20%. A little more color on gross margins and OpEx profile for the second half. Recall, our long-term model has gross margins at 59% to 60%. In 2021 and 2022, we were in our model range with margins of 59.6% and 59.2%, respectively.
In our July call, we noted the gross margin profile by quarter, which show lower gross margins in Q3 and Q4 due to the timing of spending to strengthen our supply chain, but that we expected our full year 2023 gross margins in the 57% to 58% range. That outlook is unchanged. Turning to resiliency spending. In operations, manufacturing spend will continue in Q4, but the spend associated with enabling our new factories to be qualified and producing testers is behind us.
Packing up inventory and some capital equipment from old locations is what is left to do for our manufacturing in Q4. In short, we have successfully completed our objectives of moving many product lines to new locations. I would like to thank our internal teams and our partners for their tireless effort in derisking our supply chain. Some component qualification will continue in 2024 but the majority of the test operational resiliency spending is behind us.
Regarding OpEx. As noted, our full year spend will be flat to slightly down versus 2022 spend levels. This is due to both spending controls and lower variable compensation. Recall, our operating model as a variable component for operating expenses, where the model flexes compensation expense with revenue and profits as both are lower than 2022 levels, we're spending less than 2023. As revenues are expected to grow in 2024 in future years, that variable compensation component will also grow.
For the full year 2023, at the midpoint of our guidance, revenue will be slightly below $2.7 billion with non-GAAP EPS of $2.85 and operating profit of 20%. Gross margin for the full year should be above 57.5%. Our GAAP and non-GAAP tax rates are forecasted to be 15.75% and 16.5%, respectively, in 2023. Looking at 2024 business levels, Greg noted we expected revenue growth in 2024. How much growth is tough to call at this point? Starting the year with Q1 '24 revenues, similar to Q1 '23 levels, Q1 is expected to have unfavorable product mix yielding lower gross margins in Q4 '23.
For the full year '24, we expect gross margins to be better than '23 on higher volume and lower resiliency spending. Summing up, our second half is playing out as expected with the highlight being strong execution by our UR team as they ramp UR20 to meet high customer demand. For the full year, while the end markets have softened in 2023, our company operating model has flexed costs down to support profitability while enabling our R&D and go-to-market investments to support our long-term growth objectives. We've transformed our supply chain to reduce geographic risk and strengthened our operations capacity.
From a shareholder return perspective, year-to-date, we've returned 181% of our free cash flow to owners. Our cash position and strength in our balance sheet enables us to continue to invest in strategic organic initiatives and has the firepower to support a wide range of M&A options for inorganic growth in the future.
With that, I'll turn the call back to Andy. Andy?
Question and Answer Session
Andrew Blanchard
Thanks, Sanjay. Operator, we'd now like to take some questions. And as a reminder, please limit yourself to one question and a follow-up.
Operator
Thank you. At this time, we will be conducting a question and answer session. (Operator Instructions) Our first question comes from the line of Tim Arcuri with UBS.
Timothy Michael Arcuri
Greg, you had talked about there being some critical points where utilization has to get to, to then see mobility start to grow again. So how do you think about where we are sort of in aggregate? I know your large customer has their own dynamics. But how do you think about where we are right now in terms of utilization versus where you think we have to get until -- before the nonlarge customers would come back and start to buy again?
Gregory S. Smith
Yes. That's -- thanks, Tim. So the -- right now, our -- I'm hesitant to give you exact numbers because our measurements of utilization are indirect. So we tend to look more at the changes from quarter-to-quarter than the absolute level. But they are significantly lower than what we've seen for people to want to buy.
So -- and there's also a big difference between our IDM customers and our OSAT customers. So our IDM customer utilization by our measurements are up in the low 80s, and that's usually at a high enough level that trigger buys. Our OSAT customers are like 20 points lower, there's usually a few points of inflection, like an increase in utilization from Q2 to Q3, we saw very little in our data this time. And so it's still on the order of 20 points lower than what we're seeing inside of IDMs.
So I think there's a -- there's a pretty big hill to climb in terms of utilization before that would trigger buys from OSATs to support the mobile space.
Timothy Michael Arcuri
And then as you think about your -- I mean I think a lot depends on your large customer next year. But as you think about you read some of the plans in terms of what they're planning to do and having a new chips throughout the entire product portfolio going from N3B to N3E, through the entire product portfolio, you do get 30% more transistor density.
So it depends on the die size, obviously, but it does seem like there's going to be a fairly solid increase in transistors across the entire portfolio next year. So maybe can you just -- I'm not asking you to predict what happens with that large customer, but can you talk about sort of the things that you're watching to sort of determine whether you think that next year could be a good year for that customer or not?
Gregory S. Smith
Sure. So yes, I'm -- I think we're both sort of working off of the same source data when it comes to like reading the tea leaves about this large customer. We don't really know what their product plans are. The things that we are watching for are how quickly 3-nanometer goes into their high-performance computing process -- products that the more of that product line that's on 3-nanometer, the more complexity there will be, and that's a tailwind towards loading.
The next is what's the relationship between -- like how do they use the next-generation 3-nanometer process? Do they use that to try to decrease die size and attack cost? Or do they use that to add features to sort of keep die size the same and a lot of complexity. So we're watching very closely to see how that plays out. And then the last is if there's any change to the strategy of using the [N-1] processor in the lower-end phone product line. If that changes, that compresses the period of time that they have to build the up chips, and that helps to drive peak loading.
So we don't have information about that. We are -- so we're being pretty cautious in terms of how we model that going into the future. But there are -- like as you point out, there are a number of things that could drive some upside.
Operator
And our next question comes from the line of Krish Sankar with TD Cowen.
Krish Sankar
I turn then to -- first one is on next year outlook to follow up on an earlier question. Historically, your big customers come around April or May time frame to confirm kind of like a test outlook for the year. Was it a similar -- and it seems like that was a different pattern this year. Do you expect to go back to regular patterns next year? Or is it still too early to commit this call? And I have a follow-up.
Gregory S. Smith
So I think that they actually followed a familiar pattern this year. But what ended up happening in April, May was that they ended up meeting not a lot of additional capacity. So I think we've pointed out in prior calls that we expect that major customer to be less than 10% customer in 2023. But we haven't seen any significant change in the timing.
The thing that we said in the prior call about this was we didn't see -- like we didn't have demand confirmed in that April, May time frame, but there was still some uncertainty about what their peak loading might be and whether they need additional capacity. And that played out the way it played out. So I think they're still on basically the same schedule.
Krish Sankar
Got it. Very helpful. And then a quick follow-up on the auto market. You said that it's been strong. We've seen some spot weakness, but still expected to grow or be strong in calendar '24. Is that purely because you see increase in semi content that is going to offset unit weakness? Is there anything else you're seeing in autos?
Gregory S. Smith
No. So I think that what we're seeing in automotive is most of the players in that space, they're large IDMs and they all have a significant capacity expansion projects in process. And they've -- and with the lead times that they have for front-end equipment and where our lead times were running, say, even 6 months ago, they needed to place orders way in advance.
And as they're building out that capacity, they need to phase their deliveries so that it lines up with their commissioning. And -- so we've seen some rescheduling, but we haven't seen anything that we would consider to be a significant signal of demand change.
Operator
Our next question comes from the line of Mehdi Hosseini with SIG.
Mehdi Hosseini
Yes, I have 2 follow-up. And I'm not going to ask you about how to forecast iPhone 16 builds for next year. But what I wanted to learn is, actually, I want to get an update on the compute end market. We all hear of more of an ASIC design ramping hyperscalers ramping their own ASIC solution. By now, everyone has seen, grab it on and keeping you in the headline. And in that context, what's the update on Paradigm's content market share? And I have a follow-up.
Gregory S. Smith
Okay. So right now, the computing market, if you like, just to sort of break it down, there are a couple of components of the compute market. There's end equipment PCs. That part is significantly weak and continues to be. Then there's cloud computing, and the thing that's driving cloud computing is really AI acceleration. The beneficiaries of that are GPUs and then hyperscalers doing their own silicon.
There's definitely increase in the amount of bespoke silicon that's going in, but it's still dwarfed by sort of traditional GPU-driven accelerators at this point. The TAM, sort of the amount of testers that are being sold to support the hyperscalers is very -- it's not very consistent at this point. We had a big hit last year from a revenue perspective, it's quieter this year, but we think that this is something that's going to play out over probably the next 3 or 4 years. So we're really looking at it in terms of socket wins.
And right now, I can tell you that we are on track with socket wins. We have low share in traditional compute. We were aiming to try and win half and half, like half the battles we get into on hyperscalers, and that's about what we're doing this year, that we're winning half of the sockets that we are fighting for and the other guys winning the other half.
Mehdi Hosseini
Okay. And my follow-up has to do with your color on Q1, and I appreciate (inaudible) way to better help us with the modeling. The midpoint implies an 8% sequential decline. And SOCs has historically been the weakest in Q1. So should we assume that Semi Test SOC will be down by more than 8% and everything else would be down less than 8%?
Sanjay Mehta
Mehdi, it's Sanjay. Yes. So the seasonality decline is really coming through, I would say, in robotics. Obviously, seasonality comes down. And then we're ramping UR20 Q3 and really Q4 that goes down to run rate. The other component of the decline in Q -- or the forecasted decline in Q1 is tied to storage, really the mobility of the end market and SLT. And from an SOC perspective, we're seeing that as roughly flattish at this point, Q4 to Q1. So those are really the drivers of what we see now.
And as I said in my prepared remarks, I'll just remind you that we're seeing, as lead times come down, a lot more kind of book ship and customers coming to us with incremental orders within lead time. So we're expecting to see kind of more bookings kind of at the tail end of for Q1 shipments. So there's a little bit more volatility there.
Operator
Our next question comes from the line of Samik Chatterjee with JPMorgan.
Samik Chatterjee
I guess if I can follow up on the last question Mehdi had -- actually, your response that you're getting the 50% of wins in the accelerated compute or sort of custom ASICs that you are targeting. Can you share a bit more in terms of then when we start to think about that translating the wind translating into revenue, when does that start to happen? And particularly, like when you look at a pipeline building here, how does that revenue progression look like? And I have a follow-up.
Gregory S. Smith
Yes. So right now, the noted characteristic of it is volatility year-on-year. Our model of it, though, is by the time you get out to sort of the 2026, 2027 time frame, we expect that these vertically integrated producers are going to represent about $400 million to $500 million of the compute test TAM. And for reference, that's probably going to be on the order of 1/4 to 1/3 of the total compute TAM out in that time frame.
Samik Chatterjee
And then for my follow-up, I know you're talking about 2024 being a year of incremental growth, and there are lots of puts and takes here, but you're starting 1Q in line with 1Q '23 and when I look at consensus, it has you growing 20% for the year. So obviously, as a magnitude sort of overall difference in sort of where you're starting 1Q versus what consensus is expecting. Any sort of thoughts around whether that's a realistic assumption? Or if that 20% growth were to be realized, what do you need to see in terms of sort of uplift, which quarter that needs to come through for you to be able to realize that kind of strong growth through the year?
Sanjay Mehta
Yes. I think that -- it's Sanjay. I think that Greg really outlined kind of the potential headwinds and tailwinds in 2024. So right now, the first half and given the limited visibility, especially in mobility, it still continues to look a little dampened or weaker. But from a model perspective, where we see '24 tailwinds are in compute, mobility and robotics for 2024. And so that's where we do see the tailwinds.
The question is, to what degree. And we've provided some transparency of what we know about in Q1. And we think that, that is the low point. We think we'll grow from there. It's tough to call right now how it appears to be for the rest of the year. And we'll provide an update in January with what we know.
Operator
Our next question comes from the line of Vivek Arya with Bank of America.
Vivek Arya
First one on gross margins. Sanjay, could you remind us what drove second half gross margin is lower than the first half? And what will drive overall '24 gross margins higher than '23?
Sanjay Mehta
Yes. Good question. So it's a similar story to what I noted in July, Vivek. Fundamentally, it's product mix in the second half, and we deferred resiliency spending in operations really from first half to second half and in the first half, we were focused on chasing and meeting customer demand. So some of our projects were deferred into the second half. So it's mainly mix as well as deferred resiliency spending from first half to second half.
And 2024, really, I think if -- as we've noted in our prepared remarks, incremental volume. Incremental volume, product mix is a main driver of product mix. And then there's a bunch of other puts and takes. And as I've noted in my prepared remarks, we expect to see much of the operational resiliency costs behind us in '24.
Vivek Arya
Got it. And for my follow-up, Greg, I was hoping you could help us kind of square some of the more muted commentary from customers such as the Texas Instruments about industrial right demand, not as much automotive but a lot more industrial demand. I think they said the industrial weakness is broadening. How correlated is that demand to what you see and what your UR business sees? Because you're noticing strength in UR, but when we look at folks such as TI or analog devices, they were talking about weakness on the industrial side. So how correlated have these 2 trends been historically? And what is the kind of the right read across?
Gregory S. Smith
So I think the answer to your question is not very. The business that we're in with UR and MiR, tends to be -- it's a short lead time business. We tend to run with about 4 weeks of lead time and customers will use our stuff for smaller projects. Much of the rest of the industrial segment that TI and Infineon and ST are serving is towards much longer lead time projects. So some of our industrial robot competitors are working off of year-long backlogs and for process control equipment, it can be even longer than that.
So what we tend to see is that the cycles of investment for advanced automation lag we do are actually a little bit out of phase with the cycles of investment for, say, factory building. And right now, we're seeing that there is significant weakness in orders for our peers in Industrial Automation compared to where they were a year ago. And we're actually stabilizing relative to them. And that kind of makes sense. They will build a factory, they'll start running in the factory, and then they will look at different tasks in the factory that could be automated using AMRs and cobots and then they'll make a subsequent investment to do that.
So I think the answer to your question is that you can't really infer what's going to happen with our robotics business based on the industrial semi trends.
Operator
Our next question comes from the line of Toshiya Hari with Goldman Sachs.
Toshiya Hari
My first question, I have two. But my first question on Semi Test lead times, Greg, where are they today? How much have they come in over the past couple of quarters? And over the next couple of quarters, where do you see them going? And when you talk about Q1 of '24 revenue being flat, year-over-year. I guess that's for the overall company. But within your SOC test business or Semi Test business, what percentage of Q1 revenue do you think will be turns business?
Gregory S. Smith
So I'll take the lead time, and then I'll pass it off for the percent of turns to Sanjay. The lead times, if you go back about a year, our lead times were running out past 26 weeks out to 39 weeks, we were way oversubscribed. And right now, we've managed to be able to tighten up the supply chain and get us to the point where our lead times are running in the 16-week time frame, and we're aiming in the short term to get those down to something that's closer to 13 weeks. That's probably where we're going to be sitting.
But again, that's sort of an average lead time. We will be maintaining some an ability to work within lead time for particular high-priority orders. And so going into a quarter, we will probably have a very good idea about the majority of our revenue but we will have some ability to do book ship. And I don't know, we're scrambling through the papers to make sure we give you a good answer for the turns question here. It's actually kind of fun to watch. So have you gotten to the right page yet?
Sanjay Mehta
Sure. Yes. A little bit of context. Obviously, with very long lead times, looking out a quarter we'd be mainly booked. And then there's always the rescheduling and stuff drops in and just the tactical kind of shifts back and forth. But overall, for the business, we see it at like 25%, 30% and then dropping down into Semi Test, that is in the 15% to 20% range, which we're now starting to see that, that kind of book ship is starting to become a little bit more increasing and increasing over time. As you'd expect, lead times coming down, people are going to place the orders when they believe they need them. And that's when we get the POs.
Toshiya Hari
That makes sense. And then as my follow-up, a question on China. Your peers in WFE or on the front end side of things, they're seeing 40% to 50% of their revenue come from customers in China. I think your business peaked at about 20% China a couple of years ago, and I think the most recent quarter, you were in the low teens. With new customers and existing customers, expanding capacity, is China a potentially a growth area for you guys with the time lag over the next, call it, 24 months? Or is there a reason to believe that your exposure there could stay low in given competition or what have you?
Sanjay Mehta
Sure. It's Sanjay. I'll take kind of the first half and provide some context about our current China kind of revenue. And then maybe, Greg, if you want to take the second part about the potential growth competition. So just to provide in context, right, in 2022, 15% of our revenue is in China in '23, year-to-date, we've got about 12% of our revenue from China. And that 12% -- about 1/4 of that 12% is really from robotics, production board test and LitePoint.
75% is from our Semi Test group and about 60% of that Semi Test, I know a lot of percentages, but 60% of that 70% is from indigenous Chinese customers and 40% is from multinationals. So the indigenous component is roughly about 5% of our overall business really tied to Semi Test.
Gregory S. Smith
And so I'll talk a little bit about the growth trends. If you decompose the business that we have in China right now, probably the part of the segment that's heaviest investment for us is really in memory. And we expect that to continue to grow. There's significant capital investment, WFE investment going into memory in China, and we have very good exposure to that. And there's also a significant growth in analog and power for automotive and industrial in China, and we serve that segment very well with our Eagle Test platform.
The biggest headwind that we have in China is we are unable to sell test equipment into Huawei. That's by U.S. regulations. And frankly, that takes the biggest single chunk of the test TAM in China out of our -- off of the table for us. So we're in there. We have a great team in China. We're competing for all the business that we can go after. And so we're hoping to kind of hold and potentially grow from where we are.
Operator
Our next question comes from the line of Brian Chin with Stifel.
Brian Edward Chin
I'll just ask a few questions. Greg, maybe taking your starting point, which is modest growth, in the test TAM -- semiconductor test TAM next year in calendar '24. A couple of questions. What kind of market share gain might you expect given maybe the initial expected profile of test spending next year? And also what kind of semiconductor IC unit growth or demand underpins your initial forecast for next year?
Gregory S. Smith
So right now, when we roll it up, we think that our overall ATE share is probably going to be flat to slightly up next year. No dramatic shifts from where we are. And I think there may be a bit of a tailwind because if the -- and that will depend on sort of the strength of the mobility recovery. So that's probably the biggest x factor in terms of where share goes. Now in terms of IC unit growth, I don't have the data at my fingertips in terms of what our current expectations are. I think it's typically -- I'm almost there.
Yes. So we're looking at -- probably -- yes. So unit growth was actually down this year, and it's coming back slightly, but it will still be below the units back in 2022 by our best guests. So the thing that's really going on is even though the units will be sort of at or below the peak that they had in 2022, the complexity growth that's happened since 2022 will drive test capacity requirements.
Brian Edward Chin
Got it. And is there any particular end markets where you see that sort of that unit gap feel better in terms of the complexity increase?
Gregory S. Smith
I'm sorry, could you repeat that?
Brian Edward Chin
Thinking about sort of the high watermark semiconductor shipments in calendar '22, given some markets maybe won't retest that unit level. Where is the complexity increases or quality control increases big enough to sort of fill that gap?
Gregory S. Smith
Yes. So the biggest lever there is the degree to which advanced processors move to 3-nanometer technology. So it's really in the digital space. There are some tailwinds as more of the compute segment moves towards chipset-based design. That's about a 10% to 20% tailwind on the amount of test that's required. And another area that is -- that complexity really helps is the more complex devices that are going into an automotive environment, it has a much higher test threshold than other markets. So high-performance processors for ADAS applications, those have very high test intensity for the same number of transistors is the same thing going into a fall. So that's where we're really looking for some sort of a complexity tailwind, I guess.
Brian Edward Chin
Okay. Got it. Maybe for my second question then just -- and maybe for Sanjay. I know maybe January is when you guys like to sort of refresh your thinking the most in terms of forward financial targets, but given sort of the more subdued initial outlook for calendar '24, what is sort of your initial thinking in terms of the calendar '26 financial model?
Sanjay Mehta
Yes. Great question. Every year, we go through a process where we go through a strategic planning process in Q4. It culminates with a review with our Board in January. And then we share an update to our earnings model. And we're just starting to go through that process now. So I think the first quick answer is [Sanjay] will share that with you in January.
But let me provide a little bit of color to your question. We still have conviction in the overall key drivers and fundamentals of the model in test and in robotics. I think Greg talked a little bit about them, we've narrated them. So over the mid and long term, we have conviction. Obviously, in the short term, the visibility is very muted and very cautious. And so we're going through that now. We're looking for key indicators of when the inflections will occur.
But fundamentally, we believe, over the mid and long term, we're going to get to those goals. The timing, we're going to work through and provide an update in January.
Operator
Our next question comes from the line of Joe Moore with Morgan Stanley.
Joseph Lawrence Moore
Great. You talked about a memory business that's being driven by speeds and kind of new standards, but we are seeing some volume pickup there. So I'm wondering, are you seeing test utilizations in memory that could drive business there at some point? And then some of your memory customers have underutilized their fabs, any indication that, that supply comes back online?
Gregory S. Smith
This is Greg. So we definitely are seeing that the market is being driven this year by technology-driven retool. So HBM, DDR5, LPDDR5 and next-generation protocols. The fundamentals in the memory market are getting better. So inventory levels are coming down a bit and production rates are going up.
Right now, we haven't seen that reach a trigger to drive a large amount of capacity buys. But that is one of the tailwinds that we're expecting to help a little bit in 2024, that we don't expect it to be a dramatic increase, and we would expect it to be somewhat back-end loaded because there is a fair amount of capacity that needs to be loaded before they trigger more buys, but it's definitely a tailwind.
Joseph Lawrence Moore
Great. And do you think that you can continue to gain share within memory? Given the transitions that are happening?
Gregory S. Smith
Well, we would love to. The thing that I can tell you is that our share in memory is highest in final test and we also tend to be a first mover that we have the capability to test next-generation standards and protocols, and that allows us to sort of capture more share in that part of the market. In the wafer sort part of the market that's driven really by capacity needs, there's less differentiation, less profit and more competitors.
So what we tend to see is when we're in a technology-driven retooling cycle, our share tends to be high. And when we're in a much broader capacity add that our share would tend to go down a little bit. But we're always fighting and we do have share in the wafer sort space. It's just not as high as our share in the final test space.
Operator
And our next question comes from the line of Steve Barger with KeyBanc Capital Markets.
Steve Barger
Just going back to the robotics demand stabilizing, we are seeing some plateauing in some industrial markets but the long-term model appears unchanged for robotics. So as you target bigger customers, do they have more durable robust plans relative to smaller customers? Or what gives you confidence that you can get back on track to that mid-20% CAGR by 2026?
Gregory S. Smith
Steve, so thanks for your question. Definitely, large customers are a very different sales process and application process than small customers and robotics. So in small customers, the sales cycle can be quite short but the level of repeat purchases tends to be low.
What we're seeing is we increase our direct coverage of large accounts from robotics is that they work against annual planning cycles and they come up with sort of multi project plans that they will put you into or not. And so since we started assigning more salespeople into this space in 2023, we now have a pretty rich opportunity funnel but it takes longer to get through that funnel than it does with smaller customers.
So we expect to see a much greater impact from large customers in '24 than we do in '23. At the same time, there's also a lag time associated with the build-out of our OEM channel. So just to remind you, when I say OEM, what I'm talking about is we'll sell a robot to someone who has developed a repeatable solution, whether it's for adhesive application or welding or palletizing, they basically have a product that has our robot inside of it. And then they have their own sales and marketing and service to take care of distribution and customer service.
So we signed up 48 new partners in 2023. Those partners have to go through that development process and build out their own distribution. So what we see is that, not all of the OEMs succeed. There's a certain percentage that do, and they tend to have an inflection about 18 to 24 months after the initial sign-up. So we have sign-ups that have come from 2021 and 2022 that have inflected and are about to inflect.
The ones that we signed up in 2023 are going to become a factor towards the end of '24 and into '25.
Steve Barger
So if I can summarize that, that longer selling cycle could help mitigate the cyclicality of the underlying markets just because of that length? Is that fair?
Gregory S. Smith
It's fair. There's still cyclicality because big companies have lean times and investment times as well. So they sort of follow these PMI cycles a bit, but they work on longer lead times. The thing that I think we're looking at long term as a key way to reduce revenue volatility is to try to increase the amount of software and service revenue as part of our robotics business.
So we're working to try and make sure that we can develop some recurring revenue streams in that space. And we think that, that will have a good effect. Not immediately, but by the end of this -- by sort of the '26, '27 time frame.
Steve Barger
And so that's a good segue to a quick follow-up. You mentioned being ready to execute M&A if something makes sense, which I don't recall being a big topic the past few quarters, are you thinking more about that? And where would you focus on that across the portfolio?
Gregory S. Smith
So we're always thinking about M&A. So the -- just to remind you, our capital allocation strategy is that accretive M&A is the highest priority identified use for capital. When we do not find suitable M&A targets, we return that cash flow to our investors, primarily through buybacks. And of course, we have the dividend as well. We always have a pipeline running. We're always looking at things, and we don't comment on what's in our pipeline.
Andrew Blanchard
And operator, we're about out of time. So I'd like to just thank everybody for joining. And if you have questions, please reach out directly to me, Andy Blanchard. Thank you and look forward to talking to you over the weeks ahead. Bye-bye.
Operator
And this concludes today's conference, and you may disconnect your lines at this time. Thank you for your participation.
