Good morning, ladies and gentlemen, and welcome to the Q2 2021 Teradyne Earnings Conference Call. [Operator Instructions] As a reminder, this conference call is not -- this conference call will resume -- I would now like to turn the conference over to your host, Mr. Andy Blanchard.

Thank you, Phyllis. 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 second quarter along with our outlook for the third quarter of 2021. The press release containing our second quarter results was issued last [ evening ]. [ We're providing slides ] on the Investor page of the website that may be helpful to you when following the discussion. Replays of this call will be available via the same page after the call ends. Matters that we discuss today will include forward-looking statements that involve risk factors that could cause Teradyne's [ results ] to differ materially from management's current expectations. We encourage you to review the safe harbor statement contained in the earnings release as well as our most recent SEC filings. Additionally, those forward-looking statements are made as of today, and we take no obligation to update them as a result of developments occurring after this call. During today's call, we'll make reference to non-GAAP [ financial ] measures. We've posted additional information concerning those non-GAAP financial measures, including reconciliation to the most directly comparable GAAP financial measures and 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 KeyBanc, Rosenblatt Securities, Deutsche Bank and Citi. Now let's get on with the agenda. First, Mark will comment on our recent results and the market conditions as we enter the new quarter. Sanjay will then offer more details on our quarterly results along with our guidance for the third quarter. We'll then answer your questions, and this call is scheduled for [ 1 hour ]. Mark?

Good morning, everyone, and thanks for joining us. Today, I'll summarize our results for the second quarter and first half of 2021, update on current conditions in both Test and Industrial Automation, and comment on our view for the second half of the year. Sanjay will then provide the financial details on the quarter and our guidance for Q3. The strong demand we saw in Q1 accelerated in Q2 as both our Test and Industrial Automation groups grew substantially in the quarter. The long-term demand drivers we've discussed in the past continue to power demand for our products. In Test, it's device complexity and unit growth; in Automation, it's labor scarcity, the need for resiliency and productivity improvement. Opening further, performance at the company level from 2016 through 2020 saw our sales and non-GAAP EPS grow at a compounded rate of 16% and 32%, respectively. For the first half of this year, sales are running ahead of that rate at 21%, and non-GAAP earnings per share grew at 29% compared with last year. This demonstrates both the [ vitality ] of the markets we serve and the efficiency of our operating model. Significantly, in Q2, we saw Industrial Automation demand recover in all major regions, but particularly in North America. As a result, our production and [ operating teams ] operated at a high pace in Q2, and that pace is [ increasing ] in Q3. For the year, our IA business is on track to grow about 30% from 2019 and about 40% from 2020. Looking closely at the segment level. In Semi Test, SOC shipments grew 29% in Q2 from Q2 of 2020 with particular strength in both the compute and mobility end markets. For the first half, SOC sales grew 19%. For the first time in several…

Thank you, Mark. Good morning, everyone. Today, I'll [ provide details ] on our Q2 results, comment on current business conditions and describe [ our outlook for ] Q3. Now to Q2. Second quarter sales were $1.086 billion with non-GAAP EPS of $1.91, up 29% and 44%, respectively, from Q2 '20. Non-GAAP gross margins were 59.6%, and our non-GAAP operating expenses were $250 million, but $5 million below the high guidance due to the timing of some nonrecurring engineering expenses. Non-GAAP operating profit rate was 36.5%. [ Gross margin ] in the quarter was [ 59.6% ] compared with our plan of 58%. The increase was tied to favorable product mix in the quarter versus planned. For context, Q2 2020 gross margin was approximately [ 56.2% ]. In April [indiscernible], we noted that this level of gross margin would continue in the near term, driven by the introduction of several key test systems in early 2020, which would [ take several ] quarters to come down the cost curve. These new product introductions, coupled with heightened costs driven by COVID shortages, were impacting our gross margins in 2020. One year later, we have now most of these test solutions at volume, and they have come down the cost curve as expected. The result contributed to higher margins in the second quarter, which we expect will continue in the second half of the year. A component of gross margin improvement over 2020 is higher revenue yielding of leverage in the gross margin line. We had one 10% customer in the quarter [indiscernible] excluding discrete items of [ 14.5% ]on both GAAP and non-GAAP basis. Looking at the results from a business unit perspective. [ Semi Test ] revenue of $834 million was up 27% from Q2 2020. SOC revenue was $742…

[Technical Difficulty]

Thank you.

Ladies and gentlemen, we're going to have a quick transition.

I'll go back -- okay, it's Sanjay again. I'll go back to looking at our operating model. I'd like to make 3 quick points. First, our performance over time from 10 years ago, our gross margin percent has expanded from low to mid-50s to nearly 60% today and our operating margin from the teens to low 20s to greater than 30% today. Second, our flexible business model shifts fixed cost to variable costs, where appropriate, through contract or outsourced manufacturing, variable compensation and other means, which enables resiliency on the downside and is accretive when business is strong. Third, the operating leverage in the model is evident in our Q2 results. Our Test businesses are dropping through over $0.50 per revenue dollar through to the profit line in 2021, even while we continue to increase our R&D support -- and support investments to strengthen our Test business. This enables Teradyne to continue to invest in both our Test and IA portfolios. For IA, we are building a deeper product ecosystem and distribution differentiation while achieving the Rule of 40 in our IA portfolio. Now to our outlook for Q3. As Mark noted, the demand environment across the business remains strong. Our guidance assumes no significant changes, positive or negative, in the availability of materials and also assumes that we won't see additional pandemic-related issues. With that said, sales in Q3 are expected to be between $880 million and $960 million with non-GAAP EPS in the range of $1.29 to $1.55 and 176 million diluted shares. Third quarter guidance excludes the amortization of acquired intangibles and noncash imputed interest on convertible debt. Third quarter gross margins are estimated to be between 59% and 60%. OpEx is expected to run at 27% to 29% of third quarter sales. The non-GAAP operating profit…

Thanks, Sanjay. And everybody, thanks for dealing with our small technical issue there. Phyllis, we would now like to take some questions. [Operator Instructions]

[Operator Instructions] Your first question comes from the line of Brian Chin with Stifel.

A couple of questions here. Maybe just to kick things off. I'm just curious to what degree, if any, are the extended lead times in ATE in this sort of the current strong semiconductor environment impacting your shipment outlook or pattern in the second half? Obviously, Teradyne no longer discloses bookings, but your competitor appears to be building backlog into the December quarter. So I was wondering if you have better-than-typical backlog visibility beyond September.

Yes. So lead times have definitely been moving out. I'd say auto and industrial demand is still outstripping supply. But our lead times have pushed out, obviously, based on the very, very tight supply chain environment we're dealing with. And so what used to be, let's say, within a quarter has pushed out to 5, 8 weeks incremental to where we were in the past to be, in some cases, in the 20s in weeks of lead time. So with that, we have been seeing, I guess, improved bookings from our customers, and we're working with them to manage through it.

Got it. And then maybe for Mark. I think you touched on how, at least in the smartphone market, units are kind of a bigger tailwind this year, which hasn't always been the case in recent years. But I think in the 2022, what do you think -- to what degree could increased packaging complexity across foundries and maybe some IDMs be a more meaningful driver of incremental test intensity next year?

Yes, we -- that's a good question. We bundle that into the complexity side of life. And there's already been advanced packaging technologies deployed for several years now in phones, but it's increasing. So that trend of multichip, multi-die packaging incrementally adds test time above and beyond what you would get if you were putting all of that silicon on a single integrated die. And there's not a good rule of thumb of how much of an adder it is, but it is a trend that's going to continue to grow and drive tests.

Your next question comes from the line of Atif Malik with Citi.

I have a similar question on the mobility side end market. It sounds like your mobility outlook has improved for the year versus 90 days ago despite millimeter wave weakness. You talked about units helping. Can you talk about the confidence in test intensity staying elevated for both mobility and compute into next year? And if you can also highlight the steps you're taking to improve your market share in the areas that you're not strong, like x86 GPU and display drivers.

Yes. So I think the complexity trends looking even into the next year are all very strong and positive. In the case of mobility, the newer lithographies that are coming online are being widely adopted by the manufacturers of silicon for phones, which portends more transistors, which portends more test time. So that, looking into next year, all looks positive. And we get, at this point in the year, early glimpses of what the silicon for next year might actually look like. The -- on the compute side, similarly, we not only have the trend of lower lithography nodes, but the new interface standards related to LPDDR5 and DDR5, as I mentioned in my remarks, the early adoption of that has been pushed out a little bit as the processors that go with it have been delayed a bit. But all of the complexity required to run at those higher bandwidths is coming in 2022. So those are 2 positive things I look at that gives me confidence that this driver of complexity growth is definitely going to be strong next year. In terms of share, what we said before is that penetrating the traditional x86 GPU stalwarts is going to be a long-term endeavor for us that's going to hinge on some kind of technological discontinuity like that shift to DDR5 or like a shift to PAM4 interfaces to crack into. So that one is episodic and will play out over, let's say, 3 to 5 years. On the other front is -- are the emerging hyperscalers and new people coming into the market of creating complex silicon. These are automobile manufacturers, hyperscalers that we've talked about in the past, the Googles, the Facebooks, the Microsofts, the Amazons, that's where we're focused on getting a position to grow with them as they launch their products into the market in the shorter term.

Great, very helpful. And Sanjay, for you, the largest U.S. phone maker and your biggest indirect customer talked about supply constraints impacting smartphone sales in the September quarter yesterday. Have the materials and parts tightness gone worse over the last 90 days for your business?

Well, first of all, I won't comment on any particular customer, but I will speak in general about the environment. Last year, we were really working through demand increases, but we had fairly good, robust inventory strategies, and we are working through the impact of COVID. You have thousands of components that go into these testers, and we did a lot of resiliency improvements there. You fast forward to today's environment, and the demand has really kept accelerating. And the environment is tighter, I would say, today. And frankly, we don't see it letting up until the second half of 2022. So I think as the semiconductor industry goes through continued growth, you're seeing supply chains really getting tested. And then with the increase in the infection rates of COVID, especially, I'd say, in Southeast Asia, we're working through and managing the best we can. But I believe the net summation is it's a tighter industry now than it was, say, 3, 6 months ago.

Your next question comes from the line of Mehdi Hosseini with SIG.

Two questions, one on ARM-based ASIC design. I'm trying to get a sense of how you see the TAM. You can either elaborate as a mix of SOC tests or perhaps you can tell us how is the test time for an ARM-based ASIC chip compared to like an app process or maybe that way we could give a sense of how demand looks like. So either qualitatively or quantitatively, you can -- if you can elaborate on ARM-based SOC test will be great. And I have a follow-up.

Yes. So I guess the first thing I'd say is that a ARM-based high-end map processor, which is what's in most of our phones today, driving most of our phones, has a transistor count that's equivalent to any laptop, x86 kind of product you might have in your computer. And the test times comparatively between those 2 are not that different. It's kind of proportional to test count. The new ARM processors that are coming to market for compute applications, not smartphone applications, have perhaps anywhere from a 25% to 60% at or, on transistor count, above what's at the highest end of smartphones. And the test time associated with them, I would say, is proportionately longer at this point. And so I think it's, generally speaking, scaling with transistor count, and the transistor count on the ARM side is running a little bit at a faster clip than it is on the more traditional architecture side, if that helps.

Sure. So the follow-up has to do with ARM-based. Actually, I want to just dig in a little bit more. Would you, at some point, break this out so we could better understand how kind of ARM-based SOC tester is tracking or scaling versus the rest of the SOC market? And number two, is there anything you can give us to better understand the competitive landscape for this specific application?

Yes. I guess I haven't thought about trying to find a way to break out ARM because most of what mobility is, is ARM; and then there's, let's say, compute applications for ARM; and then there's some processors that are almost dual-use, and they go into either application. So I think it's fair to say that any compute business that we have at Teradyne is ARM-based compute. And perhaps maybe in the future, we can look at some way to sort of characterize both the market and our associated revenue for that. But I don't have a good number for you now on that. And in terms of the applications coming to market, in the short term, we know about the phone applications, we know about some early adopters of ARM for compute. Those are all coming to market or are in the market now. The hyperscalers that we're working with are coming into market at various points next year with products that are quite interesting but highly speculative as to whether they'll latch in the market. And obviously, I'm not going to talk about those because they're pretty confidential. But we'll see what latches. If one of these applications can become a 100-plus million unit application, which most of these design teams are targeting, then that's a significant adder to the market.

Your next question comes from the line of Vivek Arya with Bank of America.

I was curious, how is your kind of visibility for Q4 as you kind of stand today versus what it is usually? And if you could give us some color by end market in terms of what is in your assumption for Q3 and the second half, that would be very helpful.

Sure. So Q4, if I look back at my last 2 years here, we've been surprised mainly on the positive side. And I'd say that given the tightness of supply and customers providing a little bit more in the way of backlog, it gives us a little bit more insight. However, obviously, in the near term, we have much stronger visibility in Q3 than in Q4. And so -- but we do have, I'd say, a little bit of incremental visibility. In color -- in your question in regards to color for Q3, we see continued strength in semiconductor. Obviously, the different components within Semi Test are going to be a little volatile, but we do see a trend down. It's still what we believe is very strong demand, but we see a trend down. And we see in our IA portfolio, really, a strength. Obviously, with the COVID impact in 2020, industrials are really coming back. I'll cite some PMIs in the 60s at the U.S. and Europe. And we really see that in our, as Mark noted, in our growth in the U.S. But we're seeing that growth in China as well as in Europe and the U.S. So continued expected growth in IA and then, from a test perspective, coming down a bit.

Got it. And then on the UR side, Mark, I'm curious, what are the top 3 applications you're serving today? And how do you expect these applications to evolve? You are maintaining a very strong growth rate in that business. And I'm curious, what is driving that. Is it more number of customers? Is it more applications within the same customer? So what's giving you the confidence you can maintain this very strong growth rate in the UR business? And is -- on the gross margin side, is it accretive? I know on the operating margin side, you've given a range. But I imagine on the gross margin side, it might be accretive to your business.

Yes. It's interesting because there's no silver bullet as to what's driving the growth at UR. So for example, we're going to be up about 30% from revenue in 2019, 40% up from 2020. And in the script, I mentioned this application for welding, which is a brand-new application that is now driving about 6% of our sales for the year. Last quarter, I talked about this service application for servicing high-voltage power lines that's also running at around that 6% of sales applications that was nonexistent -- essentially nonexistent 2 years ago. So already, we have 12% of our growth attributed to new applications in new markets with new customers that we didn't have 2 years ago. So it's a combination of expanding applications like those 2 examples as well as established markets growing that's driving this. And our top 3 applications tend to be the same. It's automotive supply chain, it's industrial machine tending and it's electronic assembly tend to be the ones that are at the top of the pack. But frankly, they are shrinking as a percentage of sales as these new applications come online. So what gives me confidence looking forward is that ecosystem of partners who are developing these application solutions on our platform, not our competitors, is just continuing to grow and prosper. And they're not all going to be as successful as high-voltage line tending or welding. But it only takes 10% of them to fuel the kind of growth numbers that we've been seeing. And the activity there is very strong, and the technology is maturing to the point that more and more applications can economically be served. And over the horizon, when you look at what AI can bring, there's a whole new set of, let's say, features that will enable yet another expansion of the market. It's what gives us the confidence to talk about these kind of decade-long growth rates of 25%, 30%, 35%.

Your next question comes from the line of John Pitzer with Credit Suisse.

Mark, maybe another way to ask the calendar fourth quarter question is, over the last 3 years, the business has run in such a way that second half has been greater than the first half from a top line perspective, which has broken kind of a trend, where if you go back the prior 7 years, it was first half stronger than second half. Did you have any commentary on sort of how you think the second half over the first half looked this year just given how strong the market environment is and how tight test capacity is?

You're right, perfect student of the history. And we have been, as Sanjay said, always a bit surprised in the recent years on how strong the fourth quarter has come in and driven second half to be a bit higher than first. So if you asked us today, we'd say first half is a little bit stronger than second, but that's no different than kind of what we thought for the last couple of years, too. And that just is a testament to the visibility in Q4 as -- the lack of visibility in Q4. As Sanjay said, though, it's a little bit more visibility now than before because lead times are a little longer. But all the upside that could come in is the, obviously, the piece that's not visible. So we feel probably better now about Q4 than we felt in any prior 2 or 3 years, but it's sort of what might happen between now and October that's invisible.

That's helpful. And then, Mark, as my follow-on, I apologize, the audio quality was a little poor during your prepared comments. I wanted to go back to some of the commentary you made around DDR5 adoption. I think the point you were making is perhaps it's a little bit slower than you thought. Is that correct? And I guess, more importantly, can you give us an update of how you think you're positioned for DDR5 when it does start to ramp relative to share in memory test?

Yes. Yes, sorry about the audio quality. But in fact, yes, DDR5 for server applications has pushed out a bit as the server chips themselves have been delayed. We've seen the impact of that in the demand for DDR5 testers. And similar things happened with LPDDR5 going into mobility applications. So despite all that, the share position for Teradyne, we think, is unchanged from our earlier estimates at around 40% of the market, and it's going to be a tailwind next year when the DDR5 and LPDDR5 ramp as was originally thought would happen toward the end of this year. So we're well positioned there, and I think that's a tailwind for us next year.

Your next question comes from the line of C.J. Muse with Evercore.

I guess first question, Mark, you've taken SOC test market size higher again. And I guess, curious, if we were to hit the high end of the range, what would be the driver there? And would your market share still be 48%? Or could you see greater contribution from what's driving potentially the market to the higher end of the range?

Yes. Certainly, if the market goes up towards the high end of that range, our revenue would grow probably proportional to that. Would our share go much higher? I gave a number -- I threw in a number of roughly 48% at the midpoint of our market size guide. I would expect our market share is probably going to stay around that number, plus or minus 0.5 point at this point in the year. But it's hard to sort of prognosticate too much about where the share comes from. And as you know, one customer's buying capacity can swing multiple points of share in any given year. And so one of our customers could come in and drive a demand that we don't see right now. We have the manufacturing capacity to serve it. And yes, I guess we could be up at 50% share if that happened. It's just a little opaque at the moment.

Okay. That's helpful. And I guess a follow-on question to a few of the questions that you got earlier. You spoke to DDR5 now being pushed to a tailwind next year. You've spoken to high-performance compute, particularly the non-x86 world as a tailwind for you guys next year. I think the concern out there is that your one large customer could decline meaningfully and that would cause SOC for you guys to be down next year. So I guess, as you sit here today and you're setting up your supply for customers and lead times have extended, how are you thinking about the world into 2022 for SOC for Teradyne?

Very bullish because I think the concern -- let's see, I'm not going to speak too specifically around our customer. But what I would say is that the portfolio of devices that are being developed by our hyperscaler customers continues to expand. And that's going to -- in addition to the complexity growth of the existing portfolio, you add more chips on top of that, it portends a growing market, a growing customer. So there's going to be ups and downs, but I think the cataclysmic or sort of a cliff concern, we're now a decade into this almost, so -- and it just hasn't happened. So the trend line, I think, is pretty clear, and I think that trend line will continue.

Our next question comes from the line of Toshiya Hari with Goldman Sachs.

I had 2 as well. Mark, I guess somewhat related to C.J.'s question. Just curious in terms of the auto SOC test market, I think in your prepared remarks, you mentioned that you expect the market to exceed $500 million this year and now that's the highest mark since 2017. How are you thinking about sustainability there into '22? Near-term demand is clearly very strong. And when we speak with your customers, they're all kind of complaining about extending lead times. So I guess the bias is to the upside, but curious how you're thinking about that market. And then as a follow-up, a question on Storage Test. It seems like both HDD and SLT are trending very nicely. I think the business was up more than 2x last year in 2020. Curious what you're thinking about the business full year 2021. If you can differentiate between SLT and HDD, that would be super helpful.

It's Sanjay here. So I'll take a cut at them. So from an auto market perspective, it's true this year, we've seen tremendous growth and just about $500 million or plus or minus a bit from a market size, and yes, as far as we've seen even back to 2017. And fundamentally, I think in the last call, there was commentary around the auto industry, for years, has lived on a just-in-time manufacturing. And fundamentally, in back half of 2019, sales weren't so high. And coming into '20, obviously, demand has really picked up. And so you're seeing just a significant replenishment of inventory back into the system. And we're seeing that continue. We see, obviously, the lead times. And currently, the demand is outstripping our supply. And we see that good visibility until the end of 2021. 2022 will be interesting because we'll have to take a look at what are the inventory levels and what is the market demand in 2022 in first half. And so it's a little bit opaque, I'd say, in 2022, but it will depend on what I believe to be the inventory level and, obviously, the end market demand. And from a storage perspective, a little bit of color around 2020 and 2021. If I go back to 2020, HDD and our SLT business was kind of split relatively even. One was a little bigger than the other. And from an HDD perspective, we've seen continued end-market demand fueling that business. But when we look into 2021 from an SLT perspective, we're seeing a broadening of devices being adopted for SLT testing, which is providing a tailwind, which is enabling a larger growth in that segment of the storage business this year relative to the HDD business. So the HDD business is still very strong; end-market demand, very strong. But the broadening of the SLT -- or sorry, the ASICs being tested under SLT is increasing, which is really good news. So it's growing a little faster than the HDD shipments from a storage perspective.

Your next question comes from the line of Timothy Arcuri with UBS.

I had a question about your commentary, Mark, about your SOC share for the year and sort of what it implies for SOC in Q4, and more broadly, what it implies for total company revenue in Q4. So you've sort of given us all the different pieces. You guided wireless for the year, you guided systems for the year and you gave us the pieces on the Test businesses. So if I just look at what that implies for SOC in Q4, I mean you're going to do about $575 million roughly for SOC in Q3. That's not a big mystery. But the Q4 number implies it goes down to like $425 million, somewhere in the low 4s. That will be down year-over-year. So I guess my first question is like why would that be given that you have all this visibility? And obviously, you're quite bullish about the market. So why would Q4 be down so much? That's my first question in SOC.

Yes. I don't know if it will be down so much or not. It's a portfolio of businesses, and all of those numbers have a margin of plus or minus, let's say, 1% around them. So that creates a large Monte Carlo simulation. And if you go right down the middle, I think you probably -- I'm sure you did the math right, that's probably what you come up with. But that's typical, as I said earlier, of where we stand at this point in the year looking into Q4 as in past years where it's come in higher. There's a significant unknown around what we'll book between now and through October that can drive Q4 shipments. And while we have backlog that extends in some product lines, as Sanjay mentioned, all the way through the end of the year, there's others that's more of a turns business. And we've positioned ourselves to be very responsive on the SOC front, in general, to capitalize on these sort of short demand request to come from our customers. That's why we've in the past been enabled to, in the quarter, exceed our guidance and even in the fourth quarter and prior years exceed what we thought when we were talking back then in July. So it could very well turn out that the demand that we've seen in the last couple of years in Q4 yet materializes again and we end up, as C.J. was hinting at, having a second half larger than first. It's just that at this point in the year, given what customers are talking about, we just don't see that as the most likely outcome. And it's episodic, too. What happens in Q4, a lot of it happens, in the past couple of years, in the December quarter, preparing for product launches that occur in February time frame, new phones and things like that. And so -- and again, with a rash of phone introductions in February, although we don't see it now, we could see a demand for SOC testers to support that driving December shipments.

Yes. No, it's -- I mean the total company revenue is sort of implied to be below what people are expecting for Q4, so that's kind of why I was asking. So anyway, on your revised TAM for Semi Test, so can you just update us on the segment? So I think before, you were thinking compute would be $1 billion, mobility would be $1.6 billion. Auto, you just gave that number, is like $500 million. And industrial, I think before, you were talking about $500 million. So is the revision mostly in the compute side, can you sort of update us on those numbers?

Yes. The numbers for compute are about the same. Mobility is probably up a couple of hundred million. Auto is up probably $50 million and industrial up about $50 million.

Your next question comes from the line of Krish Sankar with Cowen.

And Mark, thanks for the color earlier on the ARM test opportunity. I just want to ask the question in a different way. When do you think the non-x86 compute TAM or maybe the total test compute TAM be similar or bigger than the mobility test market? And then I had a follow-up.

Bigger than the mobility test. So your question is when will the ARM compute test TAM be bigger than the ARM mobility test TAM?

Yes, mobility test TAM is about $1.6 billion or $1.7 billion, and you said compute is about $1 billion. When do you think the total compute TAM or maybe just the non-x86 compute TAM get to be bigger than mobility?

I don't think it's going to be in the next 4 or 5 years. I think much of what's being developed in the ARM space by the hyperscalers will more likely fall into the category of mobility. So I don't see compute growing beyond mobility in the next 4 or 5 years.

Got it. Got it. And then, Mark, on the auto test market, you said it's over $500 million. I'm guessing that includes auto microcontroller, linear, et cetera. If that is the case, is there any way of giving some more color within that? Or you think it's too hard to segment it?

Yes, it does include microcontrollers. And I would say that auto is outstripped -- most of the growth is auto more than microcontrollers. Microcontrollers is growing too, but it's mostly auto. Auto -- and of course, there's microcontrollers in auto. So when I say auto, I mean microcontrollers for auto versus microcontrollers, let's say, for white goods and things like that.

Your final question comes from the line of Joe Moore with Morgan Stanley.

I wanted to ask about the millimeter wave commentary. It seemed like the expectation was primarily one customer in one region. So when you say that was kind of short of your expectations, what are you referring to there?

So it's not one customer, one region. Millimeter wave last year for us grew significantly. We probably had 90% share of the early buying per millimeter wave test in 2020. And I would say that was spread out across 6 or 7 customers. Anybody who's making a chipset related to millimeter wave was pretty much using a Teradyne tester in 2020. The deployment of millimeter wave by the telecommunication companies has ground to a halt. And that has proportionately ground to a halt the need for incremental test in 2021.

Okay, folks, thanks so much for joining us today. That concludes the call. We apologize for the audio difficulties at the front end of it. We look forward to talking to you in the days and weeks ahead. And those in the queue, I'll get back with you directly here. Thank you.

Ladies and gentlemen, this concludes today's conference. Thank you for your participation, and have a wonderful day. You may all disconnect.