Good day and thank you for standing by. Welcome to the fourth quarter and full year Teradyne Earnings conference call. At this time, all participants are in a listen-only mode. After the speaker's presentation, there will be a question-and-answer session. To ask a question during this session, you'll need to press star one on your telephone. Please be advised that today's conference may be recorded. If you require any further assistance, please press star, then zero. I would now like to hand the conference over to your host today, Andy Blanchard, Vice President of Investor Relations. Please go ahead.
Thank you, Michelle. 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, and our CFO, Sanjay Mehta. Following our opening remarks, we'll provide details of our performance for 2021's fourth quarter and full year, along with our outlook for the first quarter of 2022. 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 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 undertake 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 reconciliations to the most directly comparable GAAP financial measures, which are available on the investor page of our website. Looking ahead between now and our next earnings call, Teradyne will be participating in technology or industrial-focused investor conferences hosted by Citi, Morgan Stanley, and Susquehanna. Now let's get on with the rest of the agenda. First, Mark will comment on our recent results and the market conditions as we enter the new year.
Sanjay will then offer more details on our results along with our guidance for the first quarter. We'll then answer your questions, and this call is scheduled for one hour. Mark?
Good morning, and thanks for joining us. In our call today, I'll summarize 2021's fourth quarter and the full year and then comment on our early view of 2022. Sanjay will then provide the financial details and review our updated earnings model and capital allocation plans. While we forecast 2022 to be another solid year for Teradyne, overall sales will likely decline in the first half, as we believe SOC test sales will be impacted by modest complexity growth in our largest market, as the jump to 3-nanometer production is pushed to 2023. More on that in a minute. Recapping 2021, we finished the year strong, bringing full-year sales, non-GAAP earnings growth to 19%, and sales growth of 19% and GAAP earnings growth to 29% respectively.
2021's performance was a result of broad-based growth across our test and IE businesses. Semi-test sales grew about 17%, with our UltraFlex family contributing to the expansion into the compute sector and Eagle product line serving the automotive and analog industrial sectors growing nearly 90%. For 2021, we estimate the SoC market was about $4.8 billion, up from $3.6 billion in 2020, which puts our market share at about 46% for the year. The memory market in 2021 was about flat with 2020 at approximately $1 billion. Our Magnum family continues to shine in the NAND segment, and we're reinforcing our position in the DRAM segment as the industry prepares for LPDDR5 and DDR5 ramps in 2022 and beyond. Our share remains at about 40%.
LitePoint's wireless test business had a great year, growing 25% from 2020. The global demand for connecting and tracking just about everything, machines, materials, and people, is nearly insatiable. LitePoint is well-aligned to this trend with products that simplify testing of the expanding range of wireless standards, whether in networking with Wi-Fi 6E and 7, in location tracking with ultra-wideband, in cellular with 5G or numerous other standards. The rate of technology change continues unabated, which is great for our business. Moving to our industrial automation business, at UR sales were up 41% from 2020. We continue to expand the number of UR+ partners and certified plug-and-play apps with assembly, machine tending, and palletizing among the most popular of the more than 375 available apps. We also continue to broaden our reach beyond existing markets, often with OEM partners.
One exceptional example is our expansion into welding applications, where we finished the year with a growth of more than 3x above 2020 levels. Welding applications now account for more than 6% of UR sales. At MiR, full-year sales grew 42% from 2020 on the strength of our new higher payload MiR 250 kg, 600 kg, and 1,350-kg AMRs. It's also notable that the value of MiR AMRs with advanced fleet management software is amplified as the size of the robot fleet grows. We saw this play out last year with large account sales, those with the potential to deploy hundreds of units growing nearly 50% faster than the installed base at large. Looking ahead to 2022, the long-term drivers that power our growth in test and industrial automation are strong.
In fact, as you will hear from Sanjay, our updated 2024 earnings model reflects an expected higher growth in sales and profits for both areas. However, in 2022, while we expect our IA business to power along with +35% growth, we see our SOC business likely contracting during the year as the shift to 3-nanometer volume production is pushed to 2023. You can begin to see some of this effect in our 1Q guidance as we usually see the beginnings of our summer ramp in March. We expect 2Q to show similar effects as it's usually our peak tooling period for our largest market. As a result, we are modeling first half sales down 15%-20%. We don't expect the impact of this to extend into the second half.
We expect demand to accelerate again in 2023 as we begin to see the complexity growth related to investments for 3-nanometer gate-all-around and advanced packaging. Overall, we expect the 2022 SoC market to be similar in size to 2021 at approximately $4.6 billion-$5 billion. Shifting to memory test, we expect the market in 2022 to be in the $900 million-$1.1 billion range, which, with a midpoint that's similar to 2021. We expect spending will be weighted toward DRAM as LPDDR5 adoption expands and DDR5 for server applications ramp. In the past, the shift to spending in DRAM would be a significant headwind for us, but we expect we'll maintain our share at about 40% in 2022 as our Magnum EPIC DRAM tester grows in market popularity.
In system test, after five years of high growth driven by the storage test product line, we expect 2022 to be a digestion year with sales softening slightly. Wireless test at LitePoint, however, is expected to fill that revenue gap, so we expect that combined system test or wireless test will be about flat with 2021. Shifting from test to industrial automation, our business outlook is brighter than ever. As I noted in our last call, the penetration rate of both collaborative robots and autonomous mobile robots is under 3%. The economic environment is favorable with worker shortages, the movement of production capacity closer to end markets, and a relentless drive for higher quality and safer operations, all helping to drive demand. The opportunity in front of us is immense, and we are investing to exploit it to the fullest.
At the IA segment level, we've increased our long-term revenue growth rate and expect our sales in 2022 to grow more than 35% off of 2021. Sanjay will provide the long-term modeling details, but the key point is the investments we've made and will continue to make position us for both short and long-term success in this expanding market. Expanding OEM relationships and served markets is a key part of our strategy. Like the welding initiative that is bearing fruit, we have several others in flight. One example is in e-commerce. One of our partners, Nimble Robotics, uses AI, unique grippers, and clever software on our UR cobot platform to pick consumer goods and high-volume warehouse operations for numerous national brands. You may have seen the recent Wall Street Journal article, complete with photos of the solution in action.
Their innovative solution, dubbed Goods to Robot, complements automated storage and retrieval systems widely used in e-commerce. Our cobots ease of use and durability are a natural for this application. Well over 15 million items across 500,000 unique products have been picked to date as Nimble executes an ambitious growth plan in the e-commerce space. Another driver of growth in IA has been the growing use of UR cobots to improve the competitiveness of local manufacturing to support reshoring of production. Pentik, a Finnish maker of high-quality ceramics, is a good example. They've added automation to allow skilled craftspeople to focus on high volume and high value tasks, while UR cobots do the repetitive and physically demanding ones such as glazing and finishing of ceramics.
To summarize, 2021 was another year of impressive growth at Teradyne and caps a five-year stretch where sales and non-GAAP earnings have grown at a compounded rate of 16% and 32% respectively. As we've said before, we've managed the business to a trendline model. Our updated 2024 earnings model shows improved growth trend lines reflecting our increasing confidence in the business. These projections are not hockey sticks. They are consistent with our past performance and correlated to investment trends in semiconductor capacity and automation market drivers. Along the way, we expect some years will perform above the trend line and other years below. In test, 2020 and 2021 were above trendline years, while 2022 will likely be below. These year-to-year swings in customer buying patterns are a part of our market dynamics. Our underlying business model with outsourced manufacturing and test is designed to efficiently absorb these dynamics.
Our good and improving gross margins give us increasing leverage within this dynamic. The growth in volume and complexity of semiconductors that propels the test market is stronger than ever. Our industrial automation portfolio is well-positioned against macro trends and back to high growth. All these dynamics should net out to attractive midterm growth of both sales and earnings. With that, I turn it over to Sanjay for more details. Sanjay?
Thank you, Mark. Good morning, everyone. Today, I'll cover the financial highlights of Q4 and review the financial details of 2021. Looking forward, I'll provide our Q1 outlook and update our midterm earnings model and our capital allocation plans. Now to Q4. Revenues were $885 million, and we delivered a non-GAAP operating profit of 31% and EPS of $1.37. Semi test revenue of $592 million was strong across the board, with notable demand for apps processor and RF test and SoC and higher speed flash test in memory.
System test group had revenue of $127 million, up 23% year-over-year, driven by higher sales and storage test and defense and aerospace. LitePoint revenue of $52 million was up over 31% from the year ago period on Wi-Fi 6E and early Wi-Fi 7 test demand. Industrial automation revenue of $113 million was up 23% from the fourth quarter last year on strong demand at both UR and MiR. Non-GAAP gross margins were 59.5%, on plan and down quarter-over-quarter due to lower volume and product mix. You'll see our non-GAAP operating expenses were up $11 million- $253 million from the third quarter, due primarily to ongoing industrial automation investments. We generated $302 million in free cash in the fourth quarter.
The non-GAAP tax rate, excluding discrete items for the quarter, was 13.6%, and on a GAAP basis was 14.75%. For the full year, it was 14.5% on a non-GAAP basis and 14.75% on a GAAP basis. We ended the year with cash and marketable securities of $1.5 billion. Regarding debt, to date, $343 million of our convertible bond has been redeemed early and repaid. We have a remaining face value of $117 million. Turning to the full year results of 2021. Teradyne revenues of $3.7 billion grew $581 million or 19% year-over-year. $485 million of the growth was from our test portfolio, while IA delivered $96 million of growth.
We had one customer that directly or indirectly drove approximately 19% of our revenue in 2021. Gross margins for the year were 59.6%, and operating profit was 33%, up from 57.2% and 30% respectively in 2020. Non-GAAP EPS was $5.98, a 29% increase over 2020. We generated $966 million in free cash in 2021 and returned $666 million through share repurchases and dividends. Looking more closely at the components of 2021 revenue growth. As Mark outlined, SOC test revenues grew $371 million or 20% on strength in all market segments, but with particular growth in the compute, industrial, and automotive markets. In memory, revenues were $396 million, up 3%.
NAND test demand drove the growth in 2021. In system test sales grew for the fifth consecutive year to $468 million, up 14% from 2020. All subsegments of system test grew in 2021, with storage test delivering the largest increase, moving from $242 million in 2020 to $288 million in 2021. LitePoint is also on an impressive growth path. Sales increased to $217 million, 25% above 2020, and has grown at an 18% CAGR since 2016. Growth drivers included the production ramp of Wi-Fi 6E, Wi-Fi 7 engineering systems, and UWB expansion. It's notable that UWB is early in its market penetration, and we expect a long-term growth trajectory as new applications for this technology emerge.
In 2021, IA revenue was $376 million, up 34% from 2020's $280 million level. UR grew 41% to $311 million, while MiR grew 42% to $64 million. AutoGuide sales were under $2 million in 2021 as we continue to reposition that group for sustainable long-term growth. In 2021, we managed through numerous supply constraints at UR and MiR without a material impact on sales. We expect those constraints to become more severe and expect they may impact the timing of some shipments in the first half of 2022. We've considered that in our guidance. Now to our outlook for Q1. Sales are expected to be between $700 million and $777 million.
Sales are expected to be between $700 million and $770 million. Non-GAAP EPS range of $0.76-$0.98 on 175 million diluted shares. First quarter guidance excludes the amortization of acquired intangibles and the non-cash imputed interest on the convertible debt. First quarter gross margins are estimated at 58.5%-59.5%. First quarter OpEx running at 33%-36% of first quarter sales. The non-GAAP operating profit rate at the midpoint of our first quarter guidance is 24%. Looking at 2022. Our first half 2022 outlook is down from 2021. This will be felt in the first and second quarters, with expected Q2 company sales to be about 25% lower than Q2 2021 levels. Regarding our OpEx plans for 2022.
OpEx growth of 11%-13% is expected, with the majority of the incremental growth in IA. This will likely put our OpEx above model in 2022 before coming back to model in 2023. The OpEx investment this year is targeted at R&D to improve ease of use, broaden our robot product lines, and strengthen our go-to-market capabilities, all in support of long-term sales and earnings growth. Turning to capital expenditures. We expect CapEx to be approximately $160 million. For 2022, our GAAP and non-GAAP tax rate is estimated at 15% based on current tax laws. Moving to our earnings model. We expect test and IA growth will drive 2024 company revenue to $4.9 billion and non-GAAP EPS to $8 at the midpoint of our updated model.
Gross margin is estimated at 59%-60%, a 100 basis point increase from the prior model, and is based on product mix and operating leverage from higher revenue. OpEx as a percent of sales will decline to 26%-28% and non-GAAP operating margin expected will be 31%-34%. The model assumes current tax laws. As shown on the supporting slide, the new model is in line with the trend line performance. Now let me provide a little color. In 2021 non-GAAP EPS of $5.98, we've achieved our previous 2024 earnings model three years early on the strength of our test businesses.
As we've done in the past, when building our earnings model, we considered the year-on-year demand swings that occur in both test and IA, but we evaluate our growth on a trend line versus any specific year. Therefore, as in past models, we use the average of 2020 and 2021 revenue as the baseline for our projections. From that baseline, we expect test revenue to grow at 7%-11% compounded and IA revenues to grow 32%-45% through 2024. Both growth rates are significant increases, both in absolute and relative terms from our earlier models. In Semi Test, we expect long-term growth will be driven by the steady increase in device complexity across our businesses, along with high single-digit semiconductor unit growth.
The transition to 3 nanometer and gate-all-around technologies will drive complexity growth and higher transistor densities, increase test data collection for process yield learning, and longer test times. The complexity point is illustrated by looking at the recent market size trends. When taking an average of 2018 and 2019, the ATE market size was approximately $4 billion. In 2021 and 2022, the average will be approximately $6 billion. Our earnings model assumes continued growth at, in the ATE market, and we believe WFE CapEx is a leading indicator of future test demand. This capacity is being put in place to serve as favorable semiconductor end market trends. First, the compute market is benefiting from the growth of both X86 and Arm-based solutions with higher performing solutions at a faster cadence than in the past.
The Arm solutions include new hyperscale applications along with AI devices for the data center and edge compute. Second, automotive market has elevated to a higher level. Increases in electric and autonomous driving are pushing the growth in complexity and number of chips per car. Third, the analog industrial end market is expected to continue to operate at a higher level as the digitization of industry continues to broaden. In IA, we're attacking an attractive and sustainable long-term growth opportunity. The market penetration of cobots and autonomous mobile robots remains below 3%, while the range of applications they economically serve continues to expand. Our strategy is to drive penetration higher in existing markets and grow the number of new applications in new industry verticals. To enable this strategy, we continue to lean into our OpEx investments.
We'll operate the IA business with a target operating margin of 5%-15% while we invest to drive growth and reinforce our competitive position. We don't expect the growth rate to moderate in the midterm, but at the point in the future that it does, we'll dial back the OpEx growth and expect operating margins above 20%. In 2022, we expect we'll operate towards the low end of the 5%-15% range. Shifting to capital allocation. We'll continue our balanced strategy. This year, we'll increase our quarterly dividend by 10% to $0.11 per share in 2022. Regarding our share repurchases, in 2021, we bought back 4.1 million shares for $600 million at an average price of $125.74.
In 2022, we expect a minimum of $750 million of repurchases, reflecting our confidence in our operating model and the end markets we serve. In summary, 2021 was a record year for sales and earnings. Our global team continued to deliver extraordinary results in a challenging environment. The team's success in balancing customer demands and business needs amidst a constantly changing collection of pandemic-induced constraints was impressive. As we look ahead, long-term growth trends in our end markets remain strong, and we've updated our earnings model to reflect that outlook. While we expect 2022 will be below the long-term revenue and earnings growth trend lines, we're confident in the long-term trends in our markets and our ability to thrive in them. With that, I'll turn things back to Andy.
Thanks, Sanjay. Michelle would now like to take some questions, and as a reminder, please limit yourself to one question and a follow-up. Thank you.
Thank you. If you have a question at this time, please press star then one. If your question has been answered or you wish to remove yourself from the queue, please press the pound key. Our first question comes from the line of Mehdi Hosseini with SIG. Your line is open. Please go ahead.
Yes, sir. Thanks for taking my question. I think it would be great if you could give us some more detail as it relates to your largest customer. Your commentary on 2021 and 19% contribution is insightful. Given your guide for 2022, you suggest that we're in a two-year digestion period. I understand the 3 nanometer, which has been known for six months, but I think expectation was for higher unit shipment for that particular customer to make up for push out in 3 nanometer. It seems like that's not enough. Again, we go back to two-year digestion, and I'm wondering if you can comment on whether I just misunderstood. I have a follow-up.
Sure, Mehdi. Of course, I'm not going to divulge too much about our specific customer because that wouldn't be appropriate, but I'll give you some additional color. I wouldn't call last year a digestion period, first of all. Last year, we had tremendous growth outside of that concentration, which sort of brought their portion of our revenue from the low 20s down below 20%. It's really growth elsewhere that drove them down a bit versus digestion. You know, the phenomena this year is something that has been in, you know, whether we attribute it to a variety of things such as complexity growth due to unit volume. I think I don't wanna comment too much about that because, again, it gets into some proprietary information.
The latest news we have is that as we've indicated, there will be a pretty significant reduction here in the first half of the year, which is usually the big ramp of tooling for the fall launch of new products. You know, we've quantified that as best we can. The second half, we don't see. Typically tooling starts in March, it peaks in second quarter and a little bit dribbles into July. That's where we'll feel the impact.
Got it. Thank you. Just a quick follow-up on IA. Sanjay mentioned that for 2022, operating margin is targeted at 5% at the low end. Perhaps you can give us some color on the ROIC difference between IA and automated test, and what is it that you're looking at in the longer term that would make a continued OpEx investment in IA attractive and perhaps maybe ROIC or some other metrics could help us to better think about the investment that you're making.
Hi, Mehdi, it's Sanjay. I think, you know, the latter part of the question first. You know, as we see revenue continue to grow and with a penetration of the market of less than 3%, we are gonna continue to invest in driving competitive advantage, broaden the portfolio, go-to-market, you know, across the board in engineering and go-to-market to continue to capture the market share. From a profitability or leverage perspective, moving more towards the ROIC comment is, you know, fundamentally, when the market starts to slow and growth starts to slow at some time in the future, and we haven't considered that within our midterm, we don't see that growth moderating over the midterm with our 32%-45% growth. At that point is when we will lower the investment.
I will comment, you know, another point on the ROIC is our gross margins in that segment are above the corporate average. You know, we'll start to see a much higher return on the invested capital when revenue starts to moderate, the growth starts to moderate. But fundamentally, we are leaning into OpEx to really capture that market.
Thank you.
Thank you. Our next question comes from the line of John Pitzer with Credit Suisse. Your line is open. Please go ahead.
Yeah, good morning, guys. Thanks for letting me ask my questions. Mark, last night, Advantest reported overnight, and I think it's always a little bit dangerous to compare because starting points can be different. I think they guided kind of their view of the SoC market up about 16% this year versus yours of flat. It's off the same growth rate that you have for 2021. I guess I'm just trying to square the circle as to whether or not I should view your guidance as being particularly conservative, or is the way you square the circle just your relative customer concentration? I guess if you look outside your largest customer, how do you see the SoC market trends for this year?
Yeah, it's a good question. I saw those numbers too. I think there's two phenomena that you cited going on. I think we are a bit lower than them in terms of the annual market size guidance. Part of that is due to the fact that probably what our largest concentration is doing this year is somewhat opaque to them. That's a factor that's probably not in their numbers. You know, I do think if you look at the rest of the market, excluding that high concentration of ours, there is growth this year in the market, is what we expect. You know, it's probably almost 50/50 that gets the difference.
That's helpful. Sanjay, just going back to Mehdi's question on the IA business. I think I understand the dynamic of growth rate to op margin, but is there sort of a revenue level of IA where you would expect to see scale take over? Or is this a business that the faster you're growing, the more you have to build out channels? I'm just kind of curious, you know, is there a revenue level that we can expect you guys hitting it sometime out in the future that's just gonna drive some natural operating leverage in that business?
Yeah. Over the midterm in our model, we are seeing a little bit of that play out, as revenues grow. Sure, certain fixed costs are gonna gain on that leverage. But I'd ask you to think about if, as we grow into different verticals and the applications within those verticals, we're really building both engineering capability in our product, reducing friction, in those different verticals of how we can get to, you know, reducing the time of implementation significantly. A significant portion is tied to R&D, as well as scaling the go-to-market in different territories and in different segments. The quick answer is, as I said earlier, there is some of that scale we are seeing in the midterm, and we will see that come in.
I just wanted to reinforce the point that we continue and plan to continue to invest in broadening the portfolio, reducing friction of implementation and really, go-to-market investments.
Thanks, guys.
Thank you. Our next question comes from the line of Vivek Arya with Bank of America. Your line is open. Please go ahead.
Thanks for taking my question. For my first one, I'm curious, Mark, how much does test intensity change as you move from, you know, 5 or 4 nanometers- 3 nanometers and gate-all-around? Is this a step function in 2023 and then it stays flat in 2024, or is it a growth from 2023- 2024? Just, you know, conceptually, how much does test intensity or complexity, whichever way you want to quantify it, you know, change from, you know, 5 or 4 nanometers- 3 nanometers and gate-all-around?
Okay. There's a couple of ways you can look at it. Maybe the economic way to look at it would just look at history a little bit. When we see a major node transition in the industry, you can see that the tester market, the ATE market, and our revenue sort of grows faster than, let's say, two years into a node. You know, economically, if you look at that sort of, you can kind of see that in the charts we publish with the earnings deck that shows the revenue trends against the trend line over time. Certain years are above the trend line, so those tend to correlate to when new nodes get introduced.
The new nodes enable a big jump economically, a big jump in transistor count, and transistor count is what drives our business. More transistors mean more test times, means more testers. The last part of this is, it takes a couple of years for a new node to sort of reach, sort of its normalized volume. In the first year of production, let's say 3-nanometer next year, maybe 3-nanometer will grow to somewhere around ± 12% of semiconductor revenue next year. Maybe it's 10%, maybe it's 13%, something like that. In 2024, it will continue to ramp, and maybe it'll represent closer to 20%-25%. We'll see a couple of years of benefit from 3 nanometer as it sort of grows in its contribution to the overall semiconductor revenue stream.
All right. For my follow-up, I'm curious, what is your current exposure to 5G modems, and how should we think about how it evolves from 2023, both from a unit and a content perspective?
You know, modems are a strong area for us. The dynamics are changing a bit there because more and more cell phone manufacturers are starting to build their own modems. So, there's been a sort of a disaggregation of that market. And by and large, that's, you know, a positive trend for us because the relative share position we have at, let's say, the established suppliers of modems, was below our average share in the market, and our share at the disaggregated newcomers that are building their own modems is higher. So, we look at that as a positive.
Is that contemplated in your 2024 outlook, or would this be incremental?
Yes, it is. It is contemplated, yes.
Okay. Thank you.
Thank you. Our next question comes from a line of Timothy Arcuri with UBS. Your line is open. Please go ahead.
Hi. Thanks a lot. I'm wondering, Mark, can you break down the $4.8 billion SOC TAM in 2021 into the different segments, mobility and compute and autos and industrial and all those kind of things? Also, can you give us some sense of if you assume the market is flat, which I'm gonna ask you a follow-up, it's like hard to believe that the market would be just flat. But if you assume it is, can you talk about how you would see the mix shifting in 2022? I had a follow-up as well.
Okay. I'll give you the numbers by the segments as we currently kind of see them today, and of course, these are estimates for 2022 and 2021 for that matter. The compute market, as an example, we think last year was about a $1.1 billion ATE market, and we think that's gonna grow this year to about $1.3 billion. There's growth in the compute segment. Mobility last year was about $2 billion. We think that will come down to $1.8 billion. Much of that is due to you know, our concentrated area. Automotive and MCU, we think, is flat at about $500 million year-over-year. Industrial's down slightly, probably $600 million-$500 million. The service business will grow maybe $100 million from $600 million- $700 million.
Got it. Okay, thanks. I guess I had a question just on, I mean, having followed this industry for, you know, 25 years, the relationship, I mean, it doesn't hold true every year between WFE, and you know, testers you know, market size. Some of maybe what you're seeing in the SoC market is because some of the you know, WFE suppliers are struggling to get tools installed and whatnot. Once those tools get installed, you're gonna have to test those chips. If you compare the size of the SoC TAM to the size of the you know, you compare non-memory to non-memory, you know, you have a $60 billion-$65 billion non-memory WFE TAM this year.
If the SoC TAM is only $4.8 billion, I mean, we've, like, never seen a ratio anywhere close to that low. It would either say that the, you know, number, you know, this year has to be a lot higher as that stuff gets installed or you make it up next year. Can you just sort of talk about that? I know that you guys look at those numbers, but it just seems sort of hard to believe that the ratio would be, you know, be that low. Thanks.
Well, you're absolutely right. The reason is kind of what you hinted at, the time it's taking to bring up 3-nanometer is longer than traditionally it's taken to bring up a new node. You know, we're kind of stretching it into a three-year window. A lot of the tools, by the way, have been put in place. Some are a little bit constrained by supply, but a lot of them have been put in place. Tuning the recipe, getting it ready for ramp, and then the cycle, the process cycle time for the wafers themselves is longer. You know, once mass production switch turns on, it's incrementally longer to crank the wafers through the whole recipe than the prior node. There is a throughput.
All of that kind of moves the impact of 3 nanometer to 2023. But the ratios that you talk about is why we've, you know, we look at this carefully too, and our mid-term earnings model has increased so significantly compared to a year ago. Because all of this equipment going in to support this new technology is a bow wave coming our way. The only unfortunate thing is it, you know, it didn't make it for 2022.
Got it. Okay. Thanks, Mark.
Thank you. Our next question comes from the line of Atif Malik with Citi. Your line is open. Please go ahead.
Yes, thanks for taking my question. Mark, as a follow-up to Tim's prior question about the WFE relation. Given that you have a very low or almost no exposure to X86 related spending, which is a very big portion of WFE this year and next year, I mean, how should we think about your test growth? I mean, I think the correlations it probably breaks down because you don't have a lot of exposure there, and they're spending a ton of spending from government and whatnot. Then, what about things like advanced packaging and heterogeneous compute? If you can talk about when the test attaches rates become a meaningful driver for those end markets.
Yes. You know, there's, let's say, Teradyne's growth and the market growth to think about. The X86 world, for example, in 2022, was quite anemic, whereas Teradyne's business was quite strong that year, because of where we're concentrated and which customers are tooling. In 2022, the X86 world, I think, is gonna be more robustly tooling than last year even, and we won't benefit much from that this year. That in fact is a true dynamic. Then, you know, other things happening with advanced packaging and chiplets and all of that, those are moving, I would say, modestly into the market. They moved in cell phones a while ago, but in compute, that's something I think that we're probably not gonna feel until 2024 maybe.
3 nanometer will come first, and then the sort of more ubiquitous use of different kinds of advanced packaging will follow shortly thereafter. The two things won't jump on top of each other just because the, I think, the yield concerns of two new technologies at once will keep them staggered.
Great. As a follow-up for Sanjay, is there any impact from supply constraints in your guidance for the March quarter, as other suppliers have talked about getting hit from labor shortages and freight-related, you know, issues?
Sure. You know, I'd say the supply impact in Q4, there's really no material supply impact on sales. In Q1, I'd say we have about $30 million-$40 million worth of supply risk, which we've contemplated in our updated revenue guide. None of this will result in a share loss. It'll just be, you know, pushed into Q2. You know, a little bit of commentary on the supply environment. You know, fundamentally, we see the supply tightening in 2022. Many more semiconductor players are going into formal allocation, which incrementally indicates incremental tightness in the market in 2022. We were under the impression that in the second half of 2022, with the fab and substrate investments, that the supply-demand dynamic would alleviate and the tightening would loosen.
However, our view currently is that that's being pushed out to the first half of 2023, really based on the discussion with our supply chain partners. While the fab and substrate capital is being deployed and increasing supply, we still see tightness over 2022.
Thank you.
Thank you. Our next question comes from the line of Toshiya Hari with Goldman Sachs. Your line is open. Please go ahead.
Good morning. Thanks so much for taking the questions. I had two as well, one on the semi-test side and the other on IA. Mark, I think in your prepared remarks, you talked about Eagle test as sort of a platform growing 90% year-over-year. I think that's very consistent with some of the TAM numbers you just gave out in response to a prior question. Just curious how you're thinking about Eagle test in 2022. When we listen to your customers talk about their CapEx intentions, not only in wafer processing but also in test, you know, the posture is really aggressive, right? Whether it be, you know, TI or ST or ON or Microchip, I mean, long list of customers. Just curious how you're thinking about Eagle test and why is your TAM estimate flattish year-over-year?
Yeah. There's a couple things going on there. The Eagle test product and the thing that we usually talk about called automotive tends to be more the traditional power-related semiconductor content in an automobile. As the supply chain needed to be replenished last year and automobile production started to come back up, a lot of the traditional airbag control, anti-lock brake control, all of those things drove Eagle demand. We think this year for Eagle, for those kinds of devices , will be similar to what it was last year, so not growth. As I said, the market is, well, not growth. However, there's another thing that we're having a harder time classifying as automotive, which is the digital content of automobiles. We tend to put that at the moment in the compute market.
Those advanced controllers are something that is more targeted and put on our UltraFLEX platform. We are expecting to see some growth there for applications, automotive applications that are digital content. For the moment at least, because it's somewhat nascent and growing, it's bucketed in compute.
Got it. Thank you for that. Then as my follow-up on the IA side, I guess I'm curious, why wouldn't this business grow faster in the near term? I realize you've got supply constraints, but when you think about the business on a three to five-year basis, 40% growth is certainly not bad growth. But given the penetration rate of cobots that you guys have talked about, the labor shortages with wage inflation, I'm guessing payback periods are shorter. So, from a user perspective, I feel it's a no-brainer, particularly with COVID and all these things going on. So, is it manufacturing capacity? Is it distribution?
I think in the past, you guys have talked about, you know, getting the word out as kind of an important dynamic as you grow that business. You know, why wouldn't this, you know, grow 60%, 70%, 80% as opposed to 35%, 40%? Thank you.
Yes. That is the million-dollar question that I ask every day. Because you're absolutely right. The dynamics are there, the ROI is there, everything is there. The biggest obstacle to sort of faster growth is the unfortunate fact that to deploy a cobot takes a human, and it takes a skilled human, a technician today, to spend some number of weeks, depending on the application, putting it into production. Therefore, we're somewhat constrained by the same labor supply issues that all global industry is experiencing. We're growing that, you know, +30% a year, that army. It's not an army that we ourselves hire as much as it's an army that our distribution partners hire. They are, but there's some practical limitation on how quickly they can hire and train.
The way out of this bottleneck is to make that deployment time quicker and quicker and quicker, which is why we are actually growing faster than we're growing the human fleet of people to install these things. We've got to widen that gap. We have to get, you know, to grow 50% only takes 10% more people, for example, in the applications deployment space. But that's the biggest single thing. We're working on that. We're planning, you know, in the model we presented here, continuing to grow at this sort of 30%-45% rate through 2024, we're assuming that every year we're going to make progress on that efficiency.
Very helpful. Thanks, Mark.
Thank you. Our next question comes from the line of C.J. Muse with Evercore. Your line is open. Please go ahead.
Yeah. Good morning. Thank you for taking the question. My first question, Mark, if I go back to kind of your commentary, you know, three to four months ago, it certainly sounded like the breadth of spending at your largest customer combined with diversification and share gains elsewhere at kind of new ARM-based emerging players gave you the confidence that even if your top customer were to decline in 2022 that you know, you'd still be well-positioned. Now, you know, you're guiding to a SoC test market, you know, relative to Advantest that is $700 million lower. You told us that your largest customer was 19%, which is about $700 million. Are you basically telling us that you're taking your largest customer to zero this year?
I guess, you know, why would that change so dramatically kind of in the last three to four months? I think that it was widely known that 3 nanometer was gonna be a very kind of small node in 2022, much larger in 2023. Curious what really has changed in the last three to four months, in your view?
Well, I think certainly the macro number around how much our largest customer represents as revenue last year is right, around $700 million at 19% of revenue. No, it's not going to zero. It's gonna probably be below 10%. The magnitude, I think that the change that has occurred represents is larger than we would have expected. I think that the transparency around and even now, you know, the amount of visibility into what that will be is still vague. It's not something that we typically get full confirmation of it until April of this year, so to speak. We're kind of working on the most recent inputs that we've received.
I think they're fundamentally, you know, as I've said in the past, there's always multiple scenarios in flight with the area we're concentrated in as to how the year can play out, and there's a range of forecasts that we get that get updated along the way. Those can swing dramatically as they've done this time. I don't have any more insight than that or anything that I think is prudent to share around what changed with that large account. I do think, you know, if you go from $700 million below 10% probably this year, that's a big portion of the delta between perhaps what Advantest is guiding and we're guiding. Then we're down into, you know, 5% difference on a forecast for the year. It could be either way.
It could be either way. I don't think there's a lot of precision in that on our side or theirs, but I think we're pretty close.
Very, very helpful. I guess my follow-up question on your target model. You're talking about test revenues growing at a 7%-11% CAGR. I'm assuming that's kind of your underlying growth rate for semi-test. If I make that assumption, then your system test plus LitePoint plus HDD, SD revenues essentially are like down $200 million, you know, in three years time. I guess, are you assuming a slower kind of growth CAGR for semi-test, or is there something else going on in kind of the other bucket?
Hi, it's Sanjay here. Overall test is growing 7%-11%. Obviously, semi-test is the largest component of our test portfolio but think of our other businesses as growing marginally in test, marginally at the same level or marginally a little bit higher.
Thank you.
Thank you. Our next question comes from the line of Krish Sankar with Cowen and Company. Your line is open. Please go ahead.
Hi. Thanks for taking my question. I have two of them. Mark, thanks for the color on the SOC market size by segment. You mentioned that compute should grow to about $1.3 billion from $1.1 billion last year. I'm just kind of curious if you can help give a little more color on that on the split between X86 and non-X86. And as you know, one of the large U.S. IDMs start doing more foundry business, is that an opportunity where they would start looking at outside platforms versus in-house designed SOC testers? And then I had a follow-up.
Yeah. I wouldn't attribute it all to X86. We use that as shorthand sometimes in these conversations, but there's other areas like graphics, chips, and FPGAs that are also part of that compute thing that are growing this year. The collection of X86 plus graphics and FPGAs would be the driver of that. What was the second part of the question?
Foundry.
Oh, do you think f oundry.
Yeah.
Yeah. Yeah. You know, the thing is I do believe there's no doubt that if a company's getting into the foundry business, they're gonna have to accommodate external test platforms. The only thing is that just because there's a new foundry player doesn't change the overall global market. They're just stealing share from some other foundry who would also buy commercial test equipment. The real question is, for their internal product development, will they shift to commercial equipment where today they use in-house equipment? That, you know, well, let's wait and see. We're not assuming that in any of our midterm plans, so this plan through 2024 does not assume any of that.
You could imagine that the cost of continuing to invest in that is high, and the rate of the roadmap change now for X86 manufacturers is quite high. It used to be a very slow, predictable cadence, and now the competition's heated up and it's fast and furious. It's harder and harder for the internal tester group to keep up with that. We don't assume it's gonna happen, but I think the cards are sort of played in a way that suggests it's probably going to happen, maybe not by 2024, but certainly out in the next four or five years. That would be my bet.
Got it. Fair enough. Super helpful. Just a quick follow-up, maybe a two-part question, if I can. On the IA side, you know, is throughput an issue for you when, you know, as the industrial robotic folks start getting into, you know, down the payload, it seems like they might be faster than the cobots. Is that an issue from your vantage point? Just a quick follow-up is, you know, I think Advantest, when they talk about SOC market, they include services in it. You know, you guys do not. Is there a way or what do you think services would be this year versus last year maybe that can explain $200 million difference maybe?
Yeah. First, on the speed of the robot's question. Yes. You know, traditional industrial robots tend to have much faster cycle times. They can pick and move and place things at a lightning speed. That's partly why they're dangerous, partly why they're in cages, they're kept away from people. And those products have been around for decades, frankly, decades. Collaborative robots is really opening up a whole new application space where those 20-year-old fast products couldn't perform for some reason, either because the cost of isolating them from humans was too high, or the flexibility that they had in terms of repeatability and such wasn't good enough, or the time to change over for a higher mix environment was onerous.
You know, if it's picking up one thing and moving one thing 1,000 times a minute, for years, it's the right solution, and it's been solved. C obots is a brand-new thing. It's all the places those things didn't make sense that we're opening up with cobots, and I think that's what exciting. So, I don't think there's any news there. Your last question was-
Service
In the TAM. Yeah. I think we estimated somewhere in the $600 million-$700 million range. I frankly don't know what Advantest estimates on that, but it's probably close.
Thanks, Mark.
Operator, we have time for just one more question, please.
All right. Our last question will come from the line of Brian Chin with Stifel. Your line is open. Please go ahead.
Hi. Good morning. Thanks for squeezing us in and letting us ask a few questions. I guess first on the SemiTest side, on a quarterly basis, Mark, it looks like the year-over-year declines might bottom out in 2Q and certainly first half. I guess off this lower first half revenue base, is it then unfair to expect muted seasonality or do you think a typical sort of 52%, 48% first half, second half relationship still applies given sort of your views for the TAM this year?
Yeah. You know, we've proven the last two years that we're not good forecasting the second half of the year. Because of this abnormality in the first half, I would expect on average that you're gonna see a little more of a second half weighting because we're kind of have this a little bit of an abnormal hole in the first. It wouldn't be a typical year. It would probably be a little more back weighted.
Okay. Not even level set, but maybe even a stronger bias to second half.
Yes.
Interesting. I guess, you know, on the industrial automation side, based on your commentary, there's constraints and I'd imagine you expect year-over-year growth to slow near-term but accelerate moving across the year. Can you give us a sense of the magnitude of revenue impact this is creating in first half? Do you view that demand as perishable or shifting to second half?
Brian, are you asking a question about supply?
Well, it's the overlay of supply constraints against the demand profile you see in industrial automation. It sounded like there was a talk of some, you know, constraints or even severe constraints in first half in that business. I'm kind of wondering, you know, what that revenue impact then is, and also if it just shifts to second half or if it's perishable.
Yeah. Demand is quite strong. In Q1, out of the $30-$40 million I noted, about $10 million of it is tied to IA. Really just given the allocation for Q2, you know, our commentary in my prepared remarks, it had a little bit of the impact there. It's true, we are seeing some supply impact in the first half. I would say that, you know, from a first half, second half, we do expect demand in and revenue in the second half for IA to be larger than the first half.
Okay. Great. Thanks, Sanjay.
All right, folks. We are out of time. For those in the queue, I'll follow up with you offline, and thanks for joining us, and we look forward to talking to you in the weeks ahead. Bye-bye.
This concludes today's conference call. Thank you for participating. You may now disconnect. Everyone, have a great day.