Teradyne, Inc. (TER) Q4 2022 Earnings Call Transcript

Teradyne, Inc. (NASDAQ:TER) Q4 2022 Results Conference Call January 26, 2023 8:30 AM ET

Company Participants

Andrew Blanchard – VP, Corporate Communications

Mark Jagiela – CEO

Greg Smith – President

Sanjay Mehta – CFO

Conference Call Participants

Brian Chin – Stifel

C.J. Muse – Evercore

Krish Sankar – Cowen

Mehdi Hosseini – SIG

Samik Chatterjee – JP Morgan

Sidney Ho – Deutsche Bank

Timothy Arcuri – UBS

Toshiya Hari – Goldman Sachs

Vedvati Shrotre – Jefferies

Vivek Arya – Bank of America

Operator

Greetings, and welcome to the Teradyne Fourth Quarter and Full Year 2022 Earnings Call. [Operator Instructions] As a reminder, this conference is being recorded.

I would now like to turn the call over to Andy Blanchard, Vice President of Corporate Communications. Thank you. You may begin.

Andrew Blanchard

Thank you, Daryl. Good morning, everyone, and welcome to our discussion of Teradyne’s most recent financial results. I’m joined this morning by our CEO, Mark Jagiela; President, Greg Smith; and our CFO, Sanjay Mehta.

Following our opening remarks, we’ll provide details of our performance for 2022’s fourth quarter and full year, along with our outlook for the first quarter of 2023. The press release containing our fourth 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 language 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 events 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 as it were 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 Citi, Susquehanna, Luke Capital and Morgan Stanley.

Now let’s get on with the rest of the agenda. First, Mark and Greg will comment on our recent results and the market conditions as we enter the new year. Sanjay will then offer more details on our quarterly results, along with our guidance for the first quarter. We’ll then answer your questions, and this call is scheduled for 1 hour.

Mark?

Mark Jagiela

Hello, everyone, and thanks for joining us this morning. I’m going to limit my remarks today as Greg will be taking full leadership of the company from next week. I’ll leave the outlook to Greg, and Sanjay will provide the financial details, including our updated midterm earnings model.

2022 was another good year for Teradyne with our second highest sales in history. Midyear, we saw a turn in our markets with the SOC test market softening after 6 years of growth and Industrial Automation growth slowing.

At the company level, our financial results were slightly under the long-term trend line after operating well above trend in 2020 and 2021. This oscillation around the trend line is a familiar pattern that we expect to continue. We manage our business investments according to this trend line and tend not to chase these excursions up or down. Sanjay will note how these trend lines roll into our updated midterm earnings model.

Looking at a longer-term perspective, over the last 8 years, Teradyne’s revenue has about doubled and EPS has grown about 4x. Growth of our core test markets is part of the story, as is expansion into new markets like Industrial Automation and System Level Test. This, combined with a rich cash flow and a balanced capital allocation plan, has been the recipe for shareholder returns. As you will hear, we see all of these drivers remaining in place for the years to come.

Since this is my last earnings call, I would like to thank all of you who follow, invest and influence our journey at Teradyne. As Greg takes on the role, I’m confident we won’t miss a beat, and he’s been a key part of developing and implementing our strategy for many years.

With that, I’ll turn things over to Greg for his deeper perspective on our results and outlook. Greg?

Greg Smith

Thanks, Mark, and good morning, everyone. Today, I will summarize the full year of 2022 and then comment on our early view of 2023. For the full year, we delivered sales of $3.155 billion and non-GAAP earnings of $4.25 per share.

2022 was the second-highest revenue in company history but down 15% from 2021’s record level due mainly to reduced demand in SOC test, specifically in mobility and compute. This offset was — this was offset somewhat by strength in the automotive end market, including substantial demand for ADAS processor test.

Overall for 2022, we estimate the SOC test market was about $4.6 billion, down 6% from ’21, and the memory market was down a similar amount for the year. Industrial Automation grew about 7% in dollar terms. Foreign exchange was a major headwind in that business. Our growth was 15% in constant currency.

I’ll divide my comments on market conditions into an early view of 2023, followed by our outlook for the midterm. In July of last year, we noted the Semi Test equipment market was entering a downturn, with demand declining in end markets such as smartphones. This began to create chip supply/demand and inventory imbalances.

We noted these types of corrections typically have a 4- to 6-quarter duration. Now we’re a bit more than 2 quarters in, and as the downturn continues, our customers continue to rebalance their production and inventory with end market demand. Much of the imbalance is in test-intensive end markets like smartphones, compute and networking, where we see lower utilization. At this point in time, with limited visibility into the second half, we estimate a market size for SOC tests to be 10% to 30% below 2022’s $4.6 billion level. Major SOC producers are expected to start the transition to 3-nanometer later in 2023, and this could mitigate the headwinds a bit.

Our historically largest end customer is expected to lead this transition, and revenue driven by this customer is expected to grow in 2023, moving from less than 10% of our revenue in 2022 to low double-digit percents of revenue for this year. Demand won’t be finalized until Q2 and will be weighted towards the second half of the year.

In general, our models factor in 2 key demand drivers, unit volume and device complexity. We’ve read the same press reports that you have, that smartphone volumes are likely to be down in 2023. The complexity growth associated with the 3-nanometer transition is likely to offset the test demand impact of the unit decrease, and the net effect is the increase in revenue that I’ve described. Our assumption that this transition is going to be gradual is likely to mute the peaks, but it will also fill in the troughs as the full portfolio migrates to 3-nanometer.

Overall, averaging across the entire life of a process node, like 5-nanometer or 3-nanometer, we expect that each new node will drive continued higher test investment in total than the prior node. On top of that, increased unit growth and the addition of new part types to the portfolio will drive further increases in test investment.

In the memory test market, technology transitions continue to drive demand. Faster interface speeds in DRAM with DDR5 and in flash with UFS 4.0 cannot be tested on existing testers and are driving purchases of next-generation ATE. Offsetting this technology-driven replacement demand is a difficult end market for our memory customers, which is likely to reduce capacity buying for this year. We are seeing some impact of that in the first quarter.

On balance, we expect the memory test market size to be flat to slightly down from 2022. We expect Storage Test will be weak in 2023 due to excess capacity in the HDD end market. And Wireless Test demand will be soft on lower smartphone shipments and a demand lull in advance of the transition to WiFi 7 beginning in 2024. Shifting to Industrial Automation. As we expected, the macro-outlook in industrial markets is cautious, with weak industrial PMIs. We expect this will be a growth headwind in the first half of the year.

Rolling up these headwinds and offsetting factors over the first half of the year, our current judgment for the total company has our second quarter about flat with Q1. However, in Industrial Automation for the full year, we have 3 notable factors that should help offset these headwinds later in the year.

First, we do not expect the currency exchange impacts we experienced in 2022. Second, growth initiatives that began in 2022, including a channel transformation at UR, will gain traction. One component of the channel transformation is supplementing our traditional distributor channel with a focused OEM channel. That effort delivered 26% growth in 2022, and we expect this to continue in 2023 as our existing OEM partners continue to grow and we add additional OEMs and targeted verticals.

The third factor driving IA growth is expanding the served market through new products. Most notably, in 2023, shipments of the new long-reach, heavy-payload cobot at Universal Robots, the UR20, which will ramp in the second half of the year. Barring a significant deterioration in the macro economy and reasonably stable currencies, we expect channel expansion, combined with new products, to drive greater than 20% growth for IA in 2023, weighted to the second half of the year.

Now shifting to the midterm outlook. The short-term changes in customer buying patterns in semi cap equipment can be abrupt. We built our flexible operating model to accommodate those cycles. Our midterm plans track the long-term historical trends and the future demand drivers in each of our businesses rather than the short-term cycles. In any given year, we will land above or below trend, but that trend line has provided a reliable baseline for planning.

Sanjay will be going through a quantitative view of our 2026 earnings model. To set up that discussion, I’d like to make a few qualitative comments to provide some context. Our 2026 earnings model shows significant revenue and EPS growth, and it’s reasonable to ask why we assume midterm growth when the short-term environment is so weak. There are several factors that give us confidence in our midterm outlook. End markets, like AI and cloud computing, mobile processing and automotive, including ADAS and EV, are driving increased semiconductor content and increasing chip complexity.

The deployment of advanced wireless standards will support ever higher data volumes and the pervasive deployment of edge AI. This end market demand will drive the timeline for new semi fab nodes and packaging technologies, like 3-nanometer, chiplets and gate all around.

We have seen these technology transitions drive demand for test as the new nodes enable more complex chips and multichip packaging technologies like chiplets drive higher quality level requirements. Both of these factors drive longer test times and higher ATE TAMs. But the landscape is changing. These complex chips are increasingly developed by a new class of vertically integrated producers, or VIPs, including hyperscalers and automakers.

We’ve had good design-in success to date with this emerging customer type, and these VIPs provide Teradyne with an opportunity to grow share in a space long dominated by legacy x86 architectures, where our share has historically been lower. These large, complex devices are used in uptime-critical applications and will require exceptionally low defect levels. To achieve this quality level, our customers will increasingly adopt an additional test step, system-level tests or SLT. We have a strong footprint in this growing market, and our design-in success with new customers is expected to be a growth driver over the midterm.

Over the midterm, we expect to see WiFi 7 ramp and the rapid expansion of UWB-enabled devices for both precision location tracking and security. These new standards obsolete existing test instrumentation, and this replacement cycle will be a growth engine over the midterm for both SOC test and Wireless Test.

Turning to IA. Global labor shortages and converging regional wages will continue to be unrelenting demand drivers. Market penetration for collaborative robots, including AMRs, is under 5%, providing enormous opportunities for long-term growth.

The steady application of new technologies in our products will continue to expand our served market, and the transformation of our channel will enable us to serve a broader range of customers and drive revenue to the $1 billion level in 2026. Summing it up. Over the midterm, our strong core test businesses will support share gain and trend line growth, while IA will grow to be about 20% of company sales and become a meaningful contributor to earnings.

In 2022, we have planted the seeds for future growth. We expanded our design-in footprint in the vertically integrated producer space and recognized substantial revenues from this emerging market. We expanded our customer base in SLT. We grew our IA sales in challenging business conditions and set the foundation for higher long-term growth at both Universal Robots and [Newark].

We’re in a cyclical downturn in the semiconductor capital industry, and visibility in downturns is always a challenge. We expect sales and earnings to be below our midterm trend line in 2023. While we don’t have line of sight to an inflection in demand, that’s typical in these cycles. The market will recover, and we expect to return to historical growth rates driving strong earnings over our midterm planning horizon, as Sanjay will describe.

Before turning it over to Sanjay, I would like to thank Mark for his more than 40 years of service to Teradyne and his 9 years as CEO. He has steered the company through an extraordinary period in the semiconductor industry and helped to assure our future growth through our investments in robotics.

More personally, it’s been an honor to work for Mark during that period. His candor and insight has made me and all of us at Teradyne better. Although we’re facing a cyclical downturn, we’re facing it as a company with tremendous financial strength, a great team and a clear strategic vision, thanks to Mark.

Now Sanjay?

Sanjay Mehta

Thank you, Greg. Good morning, everyone. Today, I’ll cover the financial summary of Q4 and the full year 2022, provide our Q1 outlook and review our updated earnings model and capital allocation plans. Now to Q4. Fourth quarter sales were $732 million, which was approximately $20 million above our mid guide with non-GAAP EPS of $0.92. Semi Test revenue of $481 million was strong in Automotive and Industrial and SOC.

System Test group had revenue of $100 million, down 22% year-over-year, driven by lower sales in Storage Tests serving a weaker HDD end market, slightly offset by higher defense and aerospace and Production Board Tests. LitePoint revenue of $40 million was down 23% from a year ago due to lower cellular and WiFi demand. Industrial Automation revenue of $110 million was down 2% from fourth quarter of last year, but up 7% in constant currency.

Non-GAAP gross margins were 57.4%, above our guidance, due to favorable product mix and some resiliency costs deferred until the first half of ’23. Non-GAAP operating expenses were $252 million, about flat with third quarter OpEx. Non-GAAP operating profit rate was 23%. We had one customer — we had one 10% customer in the quarter. The tax rate, excluding discrete items for the quarter, was 12.3% on a non-GAAP basis and lower than planned because of geographic mix. We repurchased $2 million of shares in the quarter.

Turning to the full-year results. We had revenue of approximately $3.2 billion. QUALCOMM was our one 10% customer for the full year. Gross margin for the year was 59.2%, OpEx was $1 billion, and operating profit was 27.5%. Non-GAAP EPS was $4.25. We generated $415 million in free cash flow in 2022. We returned $822 million to our shareholders through share repurchases and dividends and ended the year with $1 billion of cash and marketable securities. Our tax rate for the full year, excluding discrete items, was 16.3% on a GAAP and non-GAAP basis.

Semi Test revenue for the year was $2.1 billion, with SOC revenue contributing $1.7 billion and memory $373 million. Our SOC sales contracted in 2022. But I’ll note a bright spot for us in Automotive, where year-on-year shipments increased over 60%. In memory, our sales declined about 6% and were roughly evenly split between flash and DRAM. System Test group had revenue of $469 million, which was flat to 2021. Strength in defense and aerospace and Production Board Test was offset by a decline in our storage business. In Wireless Test, revenue of $202 million in 2022 was lower year-on-year, with declines in cellular and connectivity partially offset by strength in UWB.

Now to Industrial Automation. IA revenue was $404 million, with UR contributing $326 million and MiR $77 million, which includes. Notably, MiR grew 26% for the year in constant currency, 19% in dollars. Large portion of IA sales are outside the U.S. and sold in local currencies. For the full year, 40% of the IA sales were in Europe, 35% in the U.S., 11% in China and the remainder in the rest of the world. From a profitability perspective in IA, the group was slightly above breakeven on a non-GAAP operating basis for the full year.

Now to our outlook for Q1. In October, we expected Q1 revenue of approximately $640 million which was projected to be down 10% from our Q4 midpoint. Since then, 3 declines have occurred: First, approximately $20 million of revenue was pulled in from Q1. Second, we expect approximately $20 million of supply constraints in Q1, which has not been included in our guidance range. Third, approximately $20 million of memory related to capacity additions has shifted to later in the year, including other offsetting factors.

Q1 sales are expected to be between $550 million and $630 million, with non-GAAP EPS in a range of $0.28 to $0.52 on 165 million diluted shares. First quarter guidance excludes the amortization of acquired intangibles. First quarter gross margins are estimated at 55% to 56%, down from Q4 due to lower volume and higher spending on accelerating our manufacturing and service resiliency. OpEx is expected to run at 39% to 44% of first quarter sales. The non-GAAP operating profit rate at the midpoint of our first quarter guidance is 14%.

A few points to assist you in the modeling in 2023. First, the gross margin profile. In 2023, we expect gross margins will be below our model in the first half, driven by lower volume and additional spending on accelerating our manufacturing and service resiliency. We expect second half gross margins to return to our 59% to 60% model range.

Regarding OpEx for the full year. We expect full year 2023 OpEx to be roughly flat, compared with 2022. IA profitability is planned to be in the range of 5% to 15% profit level on higher volume. Our GAAP and non-GAAP tax rate is forecasted to be 16.75% in 2023.

Turning to capital allocation. Our strategy remains consistent as we take a balanced approach and prioritize free cash flow to maintain a minimum cash level to run the business and to have reserves set aside in the event of a significant downturn. Excess free cash flow will be prioritized for M&A, share buybacks and dividends. Prior share buyback authorization was replaced with a new $2 billion authorization. We expect up to $500 million of repurchases in 2023.

Moving to our earnings model. We expect test and IA growth will drive 2026 company revenue to approximately $5 billion and non-GAAP EPS to $8.75 at the midpoint of the updated model. Gross margin is estimated at 59% to 60%. OpEx as a percent of sales will decline to 26% to 28%, yielding a non-GAAP operating margin of 31% to 34%.

Some context on the model. Over the last 6 years, Teradyne has grown revenue at a compounded 10% rate and our non-GAAP EPS at a 19% rate. Our model midpoint, we — at our model midpoint, we expect the 2016 to 2026 extended trend line for revenue and non-GAAP EPS CAGRs will be — to be 11% and 19%, respectively, roughly in line with the historical rate and reflecting an increasing contribution of revenue from Industrial Automation. We have included a backup slide in the deck that illustrates these trends.

In prior calls and Greg’s comments today, we have outlined the key drivers for the test portfolio, including device technology trends, complexity and unit growth, which we anticipate will drive ATE growth over the 4-year horizon. Our test revenues are expected to grow at a CAGR of 8% to 13% from 2022 to 2026. For IA, we believe this market is still sub-5% penetrated. Greg has outlined the drivers over the midterm for this market, including labor shortages, new products and applications and channel transformation. Our IA revenues are expected to grow 20% to 30% from 2022 to 2026.

Summing up. After 6 years of growth, the SOC test market slowed in the second half of 2022, and our financial model flexed down as designed. Despite the smaller test market, we delivered 27.5% non-GAAP operating profit, generated over $400 million in free cash flow, returned over $800 million to shareholders, reduced our share count by 4% and ended the year with $1 billion in cash.

Going forward, our updated model builds on proven foundation of historical growth in test and IA and includes new drivers of future growth. While there is uncertainty in the short run, we are confident of the growth opportunities in the markets we serve and our team executing our strategy to capture that growth.

With that, I’ll turn the call back over to Andy. Andy?

Andrew Blanchard

Thanks, Sanjay. Daryl, we would now like to take some questions. [Operator Instructions].

Question-and-Answer Session

Operator

[Operator Instructions] Our first questions come from the line of Mehdi Hosseini with SIG.

Mehdi Hosseini

The first one has to do with the Semi Test cycle. You did mention in your prepared remarks that perhaps we are in the second quarter of a typical downturn that would last 4 to 6 quarters. But wouldn’t that be fair to say that the situation with Teradyne Semi Test is different? If I just look at your guide for Q1, it would suggest that Semi Test is already down by more than 50% since the peak some time in early ’21, and perhaps Teradyne is well into the semi cycle correction. I would appreciate if you could help me reconcile, and then I have a follow-up.

Greg Smith

Thanks for the question, Mehdi. Our thought in terms of the Semi Test cycle is that we really don’t have great visibility. Where — we believe we are about 2 quarters in. And one thing to remember is that the peak that you’re talking about was well above our trend line. So there was a fair amount of contraction to happen before we got into a region, where we would have considered it to be sort of a true downturn. So we’re really measuring from that sort of July of ’22 time frame. And we think we are just about 2 quarters into it.

In terms of when it will inflect back up, we really don’t have good visibility. There are a number of factors that make us feel a little bit better about the second half of the year than the first. But we don’t see the kinds of leading indicators that the demand is returning. So I don’t think we can give you a lot more color than that.

Q – Mehdi Hosseini

Got it. As my second as a follow-up and as a follow-up to your comment, would it be fair to say that you have pushed out your targeted model by 2 years because you just — we don’t know the rate of recovery into ’24, and yet you have a new target 2026. We may go through another downturn before we get to that target. But the main valuable here is the rate of recovery into ’24, and we just don’t know. And I also want to thank Mark for all the calls that he has hosted and wish him the best of luck in his next endeavor.

Sanjay Mehta

Yes. Mehdi, it’s Sanjay. Yes, I think it’s a reasonable comment to say that the model has shifted to the right, really tied to the economic downturn. Really, as Greg mentioned, the inventory in the channel tied to the demand in the end markets coming down and the supply in the market.

Mark Jagiela

And maybe, Mehdi, thank you for those kind comments. I’ll just make one quick comment on your first point, which is every downturn cycle, as we try to measure, it’s different. But if you classify it around when does the imbalance between semiconductor supply and demand start to inflect and cause an industry-wide downturn, from that point of view, we see it as something that started last summer. So that’s how — why we mark it back to that date.

Operator

Our next questions come from the line of Timothy Arcuri with UBS.

Timothy Arcuri

So the SOC TAM in 2022, it came in at the high end of what your range was. So what segment drove that in 2022? And then, can you also segment for us, in 2023, how you see the mix? I mean, I think this year, you’d been — you sort of gave us some numbers for each of the markets. Can you sort of tell us — if you just take the midpoint of that down 20%, can you tell us sort of what the mix will be, and how the different markets will change?

Sanjay Mehta

Yes. It’s Sanjay. Yes, 2022, the SOC market, we estimate about $4.6 billion and continued strength in compute, a little bit down in ’22 from ’21 in mobility, and then strength in auto that we referenced in prior calls, and industrial continued along with service. And in ’23, really across the board in SOC, we see a decline with the exception of service across the board. And most notably, of course, in compute and mobility, as those markets we’ve talked about, we see a supply and demand imbalance, as well as Automotive and Industrial. So really, it’s across the board from a decline perspective, is our view.

Timothy Arcuri

Okay. Okay. Got it. And then I guess, Greg, since you’re taking over for the company, I know you’ve been running IA. And I guess the indictment I would have over the past 5 — well, really more like 7 years in IA, the growth expectations sort of have consistently been downticking since 2017 or 2018. And I get there’s been a ton of macro headwinds.

But can you just sort of take a step back and sort of assess maybe the competitive element in the market? Do you think some of it is competition? I know you’d like to maybe do some M&A, it sounds like in IA. So I think some — I mean I just get a lot of questions from folks that just question whether there really is as much growth in IA as you think there is because the model does keep on pushing out.

Greg Smith

Yes. So it’s a really good question. The thing that I would note is the growth of the collaborative robotics, including AMRs, has been slower than we or other players in the market or analysts have predicted. So essentially, the TAM has grown at a slower rate than we were originally envisioning. So there are other competitive entries in the field, but in general, what we’re seeing is that they’re sort of raising the profile of the market and potentially helping to accelerate how fast we increase market penetration overall. So especially in cobots, where we have a really strong share position, we think that market entrants are both sort of a blessing and a curse, that they are certainly giving customers options to compare.

We think we’ve got a good product, and we believe that we’re going to be able to hold our share despite those entrants. And that’s going to sort of increase market awareness and help to accelerate the growth of the whole market. In AMRs, it’s much more of a toss-up, that we are one of the leaders in that market, but their — the share is spread quite wide. And the market is still forming and people are still figuring out how to really make money in that space. We think we have a winning strategy over the midterm for that, but we have less certainty of the market growth in AMRs.

So our plan is to take both of those things into account. But to sort of sum up, the growth has been slower in TAM. We don’t think that it’s a sign of loss of share due to competition.

Operator

Our next question has come from the line of C.J. Muse with Evercore.

C.J. Muse

First off, Mark, congrats. And we’ll miss seeing you on Valentine’s Day, but I’m sure you’ll find better dates from here. But first question, for SOC test market share in ’23, I guess I would love to hear your thoughts on the moving parts. It looks like roughly 37% share in ’22, which was a very off year for your Cupertino account. What kind of recovery are you assuming there? And I guess, what kind of offset on the negative side should we be thinking around Eagle from the slowdown in Auto, Industrial? Would love to hear your thoughts there.

Greg Smith

Sure. No, it’s a good question. We’re very cautious about making predictions about 2023. We have a reasonable line of sight, but a lot of uncertainty around even just the first half. In the second half, there are some things that could break our way and could drive both the market larger and our share higher. And there has also been — in 2022, there’s been quite strong buying in compute with traditional customers that has tended to push our share down. So I can’t really give you sort of a target number for share in 2023, but we believe that we’re going to see some share recovery.

C.J. Muse

Very helpful. And then maybe more importantly, as you think about the underlying growth trends for SOC test, can you speak to 2024? And if you kind of had to rank-order the positive drivers for your business, whether it’s 3-nanometer, a recovery in auto, hyperscalers starting to really expand, could you kind of walk through the 3 most important drivers that we should be thinking about into 2024?

Greg Smith

Yes. So I think you hit many of them. So I think that we are more bullish about 2024 than we are about 2023, certainly. There’s — it’s a cloudy crystal ball to try and figure out what’s going on. But the positive drivers, I think you hit an important one, that 3-nanometer is going to have much broader adoption across the industry, and 3-nanometer enables a significant increase in device complexity. That’s always been a really good driver for ATE TAM. The hyperscalers and automakers, we think that, that is going to be a — we tend to think of it more as a share gain opportunity than a real TAM driver.

So if you think about — the end market has a demand for a certain number of semiconductors, those semiconductors are going to come from traditional suppliers or these vertically integrated producers. We think that we have an opportunity to do well with that class of customers and potentially shift some share.

Overall, I think I hit kind of the high points around AI, cloud computing, a recovery in mobile and especially increasing semiconductor content in automotive. The one thing that could happen in 2024 is we’re seeing unit declines in smartphones in 2023. And typically, if people take time off, if they slow down their refresh cycle for their phones, the following year tends to have a little bit more positive task to it. So that may also help with the rebound.

Operator

Our next questions come from the line of Samik Chatterjee with JP Morgan.

Samik Chatterjee

I guess for the first one, I know you’re sort of seeing the test market, particularly on the SOC side, being down, particularly in the first half. But — I mean, you called this sort of the early part of a down cycle as well. But when you sort of rely on your experience going through prior down cycles, how would you sort of think about the recovery path between the different pieces here? Like should we expect memory test to sort of rebound sooner? Or are there certain pieces of SOC that you would expect to sort of rebound a bit faster than the others? How should we think about sort of the recovery path? And then I have a follow-up.

Greg Smith

That’s interesting subtlety. So I think your question is really, how should we expect the shape of the recovery? Do we expect it to come first in SOC, first in memory? The first thing that I’ll say is that, right now, it appears that the SOC market is impacted more than the memory market in terms of 2023 TAM. And that’s mainly because of these technology transitions in memory that are driving tester purchases, even though there isn’t a need for more capacity.

What that says to me is that, when there is a recovery, it’s probably more likely to be SOC-led than memory-led because there’s sort of a capacity overage in memory that is going to need to be absorbed as the end market for memory recovers. So that’s just my opinion. I think it’s pretty murky. So it could turn out a different way. But I think SOC is likely to snap back a little faster and harder than memory would.

Samik Chatterjee

Okay. And for a follow-up, and I apologize if you’ve discussed this already on the call, I jumped a bit late. But in terms of sort of your expectations for what sounds like a bit better sort of second half versus first half, I mean, is — how much of that is driven by sort of one of the primary customers moving to 3-nanometer? And have you — are you able to provide a split of how you’re thinking about sort of first half versus second half in terms of revenue?

Sanjay Mehta

All right. It’s Sanjay here. Yes, I think from a revenue perspective, right now, our plans in the first half, as we provided, are lower, obviously, than the second half of ’22. And we do expect an increase through the year really tied to IA and including in Semi Test. And we do expect that to occur in the SOC market as well.

Greg Smith

Yes. But we’re not able to sort of predict a 2023 in full, so we really can’t give you an exact percentage, this percent first half, this percent second half. We’re cautiously optimistic that it will be bigger.

Operator

Our next questions come from the line of Brian Chin with Stifel.

Brian Chin

Again, best of luck, Mark. And Greg, congratulations and look forward to working more with you. Maybe first question, just to clarify on the complexity. You outlined again that you expect your largest customer to grow year-over-year and expect — I believe that complexity-driven revenue will occur mostly in the second half of the year. And so while the sizing sounds fairly similar, is this later than you anticipated maybe 3 months ago from a timing perspective? And just how would you describe why that those shipments could occur later towards the second half versus sort of maybe in 2Q, but where I might normally expect it to be concentrated?

Greg Smith

Yes. Thanks, Brian. Yes. I think it is a bit later than we’ve seen in some other years. Our customer always tends to firm up their demand during the second quarter. But oftentimes, they have sort of a directional idea of what they’ll need. I think the combination of softness in the smartphone end market and this technology transition is pushing those decisions a little bit later in the year. But that’s sort of our outside observation.

Brian Chin

Got it. And I mean this is probably just geared up to a plan at the moment, which obviously is subject to change or revision. But do you have sort of a — or could you give us a sense of where test cell utilization is broadly across our device end markets? Maybe where you think it gets to by midyear and where it could end the year, based on your current planning?

Greg Smith

Yes. So it’s — measuring utilization is something that’s challenging to do, and so the absolute numbers aren’t perfectly trustworthy. The thing that we try to do is we try to use the sort of a consistent method to look at it each quarter, so that we can trust the deltas. So if things are moving down, then that’s not great. If things are moving up, then that is a good leading indicator.

And what we’ve seen across the fourth quarter of 2022 is that utilization is declining, especially in OSATs. It’s helped — holding up a little bit better inside of IDMs. And we are — we expect it to sort of bottom out around where it is, and potentially, utilization to increase through the back half of this year. But it’s definitely softened in the second half of ’22.

Brian Chin

Got it. I appreciate the comment. Maybe if I could just sneak one thing on Automation. Greg, you provide a growth rate for the OEM business through UR, 26% last year. Can you give us a sense of how large that is, the significance relative to UR sales? And also, is that margin profile similar or different relative to traditional channels?

Greg Smith

Let me take that in backwards order. So the margin in the OEM channel right now is the same or better than our traditional channels. So that is a potential operating margin improvement for that business. In terms of the portion of sales, if I’m recalling correctly, I think it was 16% of UR sales in 2022, but we probably should double check that number.

Operator

Our next questions come from the line of Toshiya Hari with Goldman Sachs.

Toshiya Hari

The weakness in SOC test this year, Greg, you mentioned it was pretty broad-based. I guess I’m a little surprised that Eagle Test, Auto and Industrial could be weak as well, just given how your customers sound as of today. Are you seeing weakness kick in, in the near term? Or are you embedding some sort of conservatism as you progress into the second half?

Greg Smith

Yes. We’re — I think it’s safe to say that we’re baking in conservatism into the 2023 plan for Automotive and Industrial. It’s been a very, very strong year for that in 2022, and it sustained very well through the whole year and into the first quarter. But we’ve definitely modeled declines in that market as well.

Toshiya Hari

Got it. And then as my follow-up, a high-level question for you, Greg. I know you’ve been with the company for a long time, you’ve partnered with Mark for a long time. From a top-down kind of strategy standpoint, how you manage the portfolio? How you think about capital allocation? Buybacks versus M&A? What do you intend to inherit? And what are some of the changes you might introduce going forward?

Greg Smith

Yes. So it’s an interesting question because I helped to bake this cake. The strategy that we have in terms of a capital allocation strategy that prioritizes M&A, share buybacks, dividends, nothing is going to change in terms of that strategy.

The thing that’s happened is we’ve been on the sidelines since 2019 in terms of M&A, mainly because when we did our analysis of the valuations, it didn’t seem like we would be able to do something that was beneficial to our shareholders by using our capital in that way. It’s a different market now. And we assume that valuations are going to become more reasonable. And so you may see different outcomes across the next few years than you’ve seen in the past few years. That’s not a reflection of a change in strategy, it’s more of a change in the business conditions.

Operator

Our next questions come from the line of Vivek Arya with Bank of America.

Vivek Arya

Congratulations and best wishes to both Mark and Greg. For my first one, from what you have described about the different segments, it seems like full-year sales are in the ballpark of being down 10% to 15% year-on-year. I just wanted to make sure I’m in the ballpark. And if yes, I wanted to check how you expect your mobility sales to do this year versus last year.

Sanjay Mehta

It’s Sanjay, Vivek. Thanks for the question. Yes, we do expect full-year sales to be down. And really, as you can see with our first quarter guide and our comments throughout the call, visibility is really unclear in the second half. And that’s something we’ve learned from prior downturns, the depth and the duration of this downturn is questionable. So it’s really challenging to provide a full-year view. But I would give you the directional — yes, we agree. We expect it to be down.

From a mobility perspective, again, in earlier questions, how do we view the TAM? We view the TAM in mobility to come down. And again, our revenues are expected to come down as well, as that market significantly, we believe, is softer, just given the inventory in the channel and demand coming down and volume.

Vivek Arya

Got it. And for my follow-up, I’m curious, for the 3-nanometer transition, do I absolutely need new testers? Or is it possible to just reuse existing tester capacity?

Greg Smith

Yes. Vivek, that’s an excellent question. So unlike the memory market or the Wireless Test market, where you need new equipment for new standards, and that drives replacements, in node transitions like 3-nanometer, for the most part, the installed capacity is perfectly capable of testing the new part. And so the tester demand is really driven by increased complexity.

So if you are testing a part that takes longer to test, you’re going to need to add to your test capacity, test the part on both the stuff you already own and then the new testers that you buy. And so we look really carefully at how much of the industry is on a particular node because that drives basically the number of transistors that have to get tested. So it’s definitely — there’s a high degree of reuse in the SOC market as nodes transition.

Vivek Arya

Got it. So there is a scenario where, if 3-nanometer volumes are low enough, that they don’t really need to buy new testers in the back half, right?

Greg Smith

Yes. That’s certainly one scenario. It’s not our current best view, but things could certainly play out that way because of this balance of units versus complexity.

Operator

Our next questions come from the line of Krish Sankar with Cowen.

Krish Sankar

Mark, thanks for all your help, and welcome, Greg. Two quick questions. First one, Sanjay, you mentioned about gross margins in the back half of the year kind of going back up. Kind of curious, what gives you that comfort, especially given the fact that if auto does roll over, those are high-margin Eagle testers that get out of your profitability equation? So I’m just kind of curious on the second half gross margin. And then I have a follow-up.

Sanjay Mehta

Sure. Thanks for the question. So I think a couple of things are included in our plan. The first thing is we expect to gain some operational efficiencies, some tied to specific programs, I would say, mainly in our IA portfolio. There are also some price increases baked into our — that are rolling through in the second half, and some volume that is anticipated in the second half.

Krish Sankar

Got it. Very helpful. And then as a follow-up, maybe for Greg. Kind of curious, in the past, you folks have mentioned how — I understand you’re not impacted by some of these China export controls because it’s more of a front end. But the Chinese OSATs have been a big component of your China sales, and they might also see a slowdown if some of your front-end counterparts are seeing weakness. I’m kind of curious how to think about China on a go-forward basis, especially given the fact that through 2021 and the first half of 2022, Chinese OSATs seemed to have ordered a lot of equipment.

Greg Smith

Yes. So that’s — it’s a good question because most of the business that’s going through Chinese OSATs, it comes from 2 places. There are local Chinese fabless specifiers that are driving volume through that value chain. And then there’s also companies that are from other places that are utilizing those facilities. What we expect is that, if there are issues in terms of producing chips in China because of any sort of restrictions, the non-Chinese specifiers will move their supply chains. They still need to build their chips. They will move to other parts of the world, other facilities to do that.

The indigenous Chinese sales, that could be affected by the new trade regulations that are in place. But I’d like to point out that, that’s not the majority of our revenue in China, and it’s not the majority of the business that — of the TAM in China. So our exposure there is limited to probably less than 5% of our total revenues.

Operator

Our next questions come from the line of Vedvati Shrotre with Jefferies.

Vedvati Shrotre

So I just wanted to dig in a little bit on the compute SOC side of things. So you mentioned that you’re seeing a lot of digestion that’s going on in the traditional x86 markets, where you have lower share. So does that, in a way, put you in a better position? Are you seeing things — are things different on the hyperscalers versus the traditional x86 architectures?

Greg Smith

So that’s a really good question because 2022 was a terrific year for traditional compute suppliers, and they drove a lot of tester purchases. We are seeing that soften in the back half of the year, and we expect that to remain soft, going into 2023.

The new vertically integrated producers, including hyperscalers, are running to it on a different track. They have new parts that they’re designing, that they’re releasing to production. And those parts typically need different tester configurations that are driving capacity in different places than the traditional compute specifiers. So we do think that, that’s going to be a stronger part of the compute market in 2023 than the traditional — there’ll be softness in the traditional market.

Vedvati Shrotre

That’s great. And then for my follow-up. On the automotive side, there’s, I think, portion of your revenues are tied to legacy parts and some would be more on the ADAS side of things. Could you help me understand like what that split is, essentially? Is it a 50-50 split? Or is it more levered to traditional and less towards ADAS or other way around?

Greg Smith

Yes. So it’s definitely a market in transition. If you look back in time, legacy automotive parts, airbag controllers, engine controllers and MCUs, was sort of the majority of that market. And last year, that market would have been in the $600 million to $800 million range. There are new part types, and there are 2 types of new part types. One are ADAS, which are complex digital devices. And the other would be power devices for battery management and for high-power management going into electric motors, in EVs and hybrid vehicles.

That part is, in 2022, would probably have been maybe 2/3 of the size of the traditional automotive. Well, actually — so the number that I gave you in terms of the size of the automotive market includes these new classes, and I would say that in 2022, it was probably a little bit more than half in these new stuff that’s growing, the EV and ADAS, and slightly less than half in legacy.

Moving forward, we expect the legacy stuff to track vehicle sales. So as vehicle sales increase, that would increase at a relatively slow rate. But the adoption of ADAS and the transition from internal combustion to electric vehicles is going to drive much higher growth rates in ADAS and power.

Andrew Blanchard

And operator, can we sneak in just one last question, please? Operator Our final questions come from the line of Sidney Ho with Deutsche Bank.

Sidney Ho

I just have one question. So you talked about your largest customer going from less than 10% of sales to, I think, [’22]. So call it $300 million, going to low double digits in ’23. So maybe $100 million increase. But that’s still quite a bit lower than in 2021. First, are those numbers right? And second, do you think that this, call it, $400 million range is the correct run rate, going forward, but maybe with less volatility than in the past?

Greg Smith

Yes. So I don’t think I want to comment on the specific numbers. But directionally, I think you’re not wrong. And your comment about smoothing out the troughs, that’s our belief as well. As to what the sort of steady state average is, I think we’re going to have to wait and see.

We’re pretty — we have some optimism that, that average is going to be strong because of both new parts coming into the portfolio and the transition of the portfolio of all processors going to 3-nanometer and continuing to march in that way. So that’s — I think that’s kind of our view, that we have some optimism that the peaks are going to be lower, the overall is going to be stronger.

Andrew Blanchard

All right, folks. And we’ve overstayed our welcome. So thank you so much for joining us today, and we look forward to talking with you in the weeks ahead. And those remaining in the queue, I’ll get back to you later this morning. Thanks so much.

Operator

Thank you. This does conclude today’s teleconference. We appreciate your participation.