Teradyne, Inc. (TER) Earnings Call Transcript & Summary

Ladies and gentlemen, thank you for standing by, and welcome to the Third Quarter 2020 Teradyne Earnings Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded. [Operator Instructions] I would now like to hand the conference over to your speaker today, Andy Blanchard, Vice President of Investor Relations. Please go ahead, sir.

Thank you, Josh. 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 2020's third quarter, along with our outlook for the fourth quarter of 2020. The press release containing our third quarter results was issued last evening. We are 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 may 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 will make reference to non-GAAP financial measures. We have posted additional information concerning these non-GAAP financial measures, including the reconciliation to the most directly comparable GAAP financial measure, where available, on the investor page of our website. Also, please take special note of the safe harbor statement in the press release and slide deck for risks related to the COVID-19 pandemic and changes to the U.S. export regulations. Looking ahead, between now and our next earnings call, Teradyne expects to participate in technology or industrial-focused investor conferences hosted by Baird, Wolfe Research, Crédit Suisse and UBS. Now let's get on with the rest of the agenda. First, Mark will comment on our recent results, current market conditions, trade regulations and our future outlook. Sanjay will then offer more details on our quarterly results along with our guidance for the fourth quarter. We will then answer your questions, and this call is scheduled for 1 hour. Mark?

Thanks, Andy. Good morning, everyone, and thanks for joining us. My prepared remarks today will cover 3 topics: first, the highlights of our third quarter and first 9 months of the year; second, the impacts of the latest trade regulations on Teradyne; and third, I'll share with you how we're thinking about the test and automation markets as we close out 2020 and look into the next year and beyond. Our third quarter results were above guidance and reflect the continued strength of our test businesses. Additionally, our Industrial Automation businesses grew 17% from the Q2 trough, and we are now operating at 2019 quarterly levels as manufacturing activities in Europe and North America improve. At the company level, sales in Q3 were 41% above Q3 '19, and non-GAAP EPS grew 53% from the year ago level. Throughout 2020, and we have seen increased short-term upside demand across our Semi Test end markets. The impact of this is clear in both our above guidance results in Q3 and in our Q4 guidance, which at the midpoint is substantially higher than we forecast in July. Throughout 2020, the semiconductor ecosystem has seen somewhat cautious initial forecast, which were replaced with better-than-expected actual demand and test has been no exception. We continue to run a manufacturing pipeline that allows us to respond to this upside. Stepping back and looking at our performance through the first 9 months of the year, the results show the success of our new products and related designing efforts and the resilience of our employees, supply line partners and operating model. Teradyne sales through 9 months are up 44%, and our non-GAAP earnings per share are up 79%. Our test businesses collectively grew 52% year-to-date, while Industrial Automation revenue on an as-reported basis contracted 10%, reflecting the pandemic impact. In Semi Test, we estimate the SOC market will be about $3.3 billion, roughly flat with 2019's level as automotive, industrial and linear markets remain depressed. However, our SOC test business is up 53% year-to-date due to strong investments in mobility test and the shipment of our new UltraFLEXplus platform, which is ramping significant design wins. The principal driver of mobility test demand continues to be increases in complexity of cell phone silicon. This is especially notable in 2020 when smartphone unit shipments are expected to decline about 10% to 1.2 billion, yet the collective test intensity of each unit continues to grow at a rate in excess of this unit decline. Within smartphones, the mid-to-high tier is the place to be in test, and that's where Teradyne is solidly positioned. These phones are seeing disproportionate growth and complexity related to multiple high-density camera arrays, and the associated processing power and storage to manage this data. Another complexity driver is 5G, and these high-tier phones are early adopters of the extra silicon needed to enable these features. Less than 250 million phones are expected to be 5G-enabled in 2020, and only a fraction of those will support millimeter wave communication. So despite the bump in 2020, we are still in the very early stages of 5G adoption. Memory test is another bright spot. The market is likely to be up about 50% from $600 million in 2019 to about $900 million in 2020. The shipment ramp of our Magnum EPIC product LPDDR5 win last year, combined with continued strength in flash demand, has driven our year-to-date memory revenues up 70% from 2019. In System Test, revenues are up nearly 50% through 9 months on growth in Storage Test and defense-related investments. Recall Storage Test serves HDD and system-level test markets, and we expect sales to more than double in 2020 to over $200 million. And at LitePoint, sales are up 18% year-to-date due to increasing adoption of advanced connectivity standards like WiFi 6E and our growing share in 5G production test. As noted earlier, in Industrial Automation, we saw a significant uptick in demand in Q3 with growth of 17% off the second quarter trough. UR grew 23% as demand in Europe, North America and China showed steady gains. AutoGuide continues to win new accounts, and we expect over 50% growth in 2020 on a pro forma basis. Regarding trade, as we noted last quarter, the China military end-user restrictions require increased compliance work and costs. But we do not expect any material impact on our sales into China. In the case of Huawei restrictions, the fleet of testers previously installed at OSATs to support their device test are already being reabsorbed into the market to test the alternative sources of silicon supply that's growing to fill in the gap created by these regulations at Huawei. For example, in the third quarter, we have seen an increase in upgrade orders at these OSATs customers to reconfigure installed testers to meet the unique needs of new customers. This upgraded and repurposing continues in fourth quarter. Shifting to the future. It's difficult to make the call on how 2021 will shape up as it's been difficult to predict 2020 even on a quarterly basis. Customers will likely continue to forecast conservatively and respond close into demand. However, semiconductor complexity growth has proven itself resilient to COVID, and is the fundamental driver of our test business. With that in mind, I'll comment on a few of the key indicators that we are watching. In SOC test, we will be watching the smartphone market for complexity increases to support higher performance video and still photography, the adoption rate of 5G in millimeter wave, AI integration and handset unit growth. We'll also be watching the automotive and analog markets for signs of a sustainable recovery in test demand. Longer term, the increase in edge AI devices should drive billions of additional complex chip units into the market by 2025 so early design wins in that area are key. In memory, the transition to higher performance DDR5 standards is just underway and should accelerate in 2021, along with newer high-speed UFS and NVMe flash interfaces. The roadmaps for both flash and DRAM show continued growth in interface speeds, which is another driver of test intensity beyond the traditional bit growth, and should drive healthy memory test demand over the midterm. At LitePoint, the continued growth of WiFi 6, 6E and ultra-wideband connectivity standards, along with 5G will be drivers for continued growth. A bit further out, we expect the next-generation WiFi 7 standard will require another refresh of the entire existing connectivity installed base of testers. In System Test, Storage Test is the interesting wildcard. After [ toward ] growth in 2020, the underlying demand drivers remain in place. In HDD, both increasing complexity and 30%-plus annual exabyte growth and in system-level test, increasing device complexity and higher quality requirements are driving the additional test intensity. However, both our narrow markets and prone to swings in investment levels at individual customers. Our Industrial Automation businesses are well aligned to long-term economic and technical trends in manufacturing and material handling. So we are confident in their ability to return to high growth. The only question is how quickly the manufacturing economy returns to health. We continue to scale our distribution capability and invest in R&D to widen our leading position. Among other things, AutoGuide adoption by key logistics, e-commerce, retail and automotive customers in 2021 will set the stage for multiple years of double-digit growth. Despite these comments, as I've noted in the past, we do not spend too much time trying to predict the various short-term demand drivers as they generally don't affect our investment plans. We do spend a lot of time trying to predict the underlying long-term growth drivers. We want to be positioned with the right products at the right customers at the right time. We believe the use of semiconductors across the global economy will continue to expand and chip complexity will grow along with that expansion. Similarly, in Industrial Automation, the cost performance of the sensor, software and mechanical building blocks of advanced automation continues to improve, making our products economically attractive to an expanding universe of customers. We have built our strategy on these fundamental beliefs, and built our operating model with the flexibility to deal efficiently with the inevitable ups and downs of economic cycles. So while we can't predict what lies ahead in 2021, we finished 2020 on an optimistic note. The company -- across the company, our employees delivered remarkable results under very difficult circumstances. Our new test and IA products are seeing strong market acceptance, and our R&D pipelines are well-stocked with future products to drive future growth. With that, I'll turn things over to Sanjay for the financial details.

Thank you, Mark, and hello, everyone. In my remarks, I'll review our Q3 financial results, comment on how COVID is impacting our business, provide Q4 guidance and comment on our full year financial outlook at the midpoint of our Q4 guidance. Our third quarter sales of $819 million were just above the high end of the guidance range, which enabled a 30% non-GAAP operating margin and $1.18 non-GAAP EPS, which is also above the high end of our range. Strength in Semi Test and Storage Test were the revenue highlights. Improved profit was a result of higher sales, partially offset by a higher-than-forecasted tax rate. Our non-GAAP operating expenses were $211 million, and our non-GAAP diluted share count in the quarter was 175 million. In Q3, our non-GAAP tax rate was 17.4%, which included a year-to-date catch-up as our estimated 2020 annual non-GAAP tax rate increased to 15.5% from our prior forecast of 14.5%. The increased tax rate is driven by higher foreign earnings, which resulted in an increase in the U.S. minimal tax on foreign earnings. We generated $280 million in free cash flow in Q3. We paid $17 million in dividends, had capital expenses of $63 million and ended the quarter with cash and marketable securities of approximately $1.3 billion and no short-term debt. Inventory decreased to $191 million. DSO in the quarter decreased to 65 days. We had 1 10% customer in the quarter. At the business unit level, Semi Test sales were $592 million, up 49% from Q3 '19. SOC shipments were $449 million, and memory test had record shipments at $143 million. Semi Test saw strength in mobility and compute applications in SOC where we continue to ramp our UltraFLEXplus test system. In memory, we saw broad shipments across flash and our Magnum EPIC solution enabled continued strength in DRAM. We expect these products to continue to gain new applications as market acceptance has been very encouraging to date, and they're well aligned to technology trends in both SOC and memory test. As Mark noted, the automotive microcontroller and analog markets remained at historically low levels in 2020. But we did see some pickup in the analog markets in Q3 versus our expectations. Several analog companies have outperformed their expected results, and we are seeing some unexpected short lead time demand as a result. Shifting to System Test. Sales in the quarter were up 61% from Q3 '19 to $118 million, Storage Test was the star at $76 million as both HDD and system-level test delivered strong results. Defense and aerospace and production board test combined to deliver $43 million in the quarter. Rounding as the test portfolio, after a very strong second quarter, LitePoint sales softened to $41 million, down 4% from the Q3 '19 level. Industrial Automation revenue was $69 million, flat from Q3 '19 and growth of 17% quarter-over-quarter. UR contributed $53 million, near $10 million and AG and Energid made up the remainder. While the COVID pandemic has negatively impacted our go-to-market efforts in Industrial Automation, we are seeing signs of improvement in several geographies across the globe and some seeing year-over-year increases in sales. These positive signs in different territories should be balanced by a continued uncertainty tied to COVID-19 in predicting the pace of the global recovery in IA over the short term. Turning to the impact of COVID-19. At Teradyne, our priorities remain consistent during the coronavirus pandemic: safety of our employees, supporting our customers and a focus on execution to achieve our financial objectives. In line with my prior earnings call remarks, I want to acknowledge the continued challenges during the pandemic that our employees, customers, suppliers and their families are going through. Operationally, we continue to work through supply line issues in the quarter. And again, I must recognize the incredible work and scale of our operations team and our supplier partners to successfully overcome a wide variety of challenges to meet our customer requirements. Great execution also by our engineering teams in introducing new products and delivering products during the quarter, while overcoming significant supply chain issues along the way. In late October, we're in a much better spot than 6 months ago. That said, there is still uncertainty of how this pandemic will impact global supply chains and market demand going forward. We still expect to encounter spot shortages and other issues through the remainder of the year and likely into 2021. Our guidance range continues to be wider than typical to reflect the potential impact of these uncertainties and some short lead time business noted earlier. Our test portfolio continues to execute, grow shares and revenues. As a result, we are investing, along with our contract manufacturers, to increase capacity and resilience in our supply chain. Specific actions include building larger buffer stocks and some components, and increasing the geographic diversity of our supply and contract manufacturing. During the pandemic, we reduced our OpEx spending tied to travel, trade shows and other go-to-market activities. The savings were approximately $8 million to $10 million per quarter. Post the pandemic, when returning to normal, we expect these expenditures to come back to our P&L. Moving to the fourth quarter -- moving to the outlook for the fourth quarter. We expect revenue of $680 million to $740 million and on a non-GAAP -- and a non-GAAP EPS of $0.90 to $1.06 on 175 million diluted shares. This guidance excludes the amortization of acquired intangibles and noncash imputed interest on convertible debt. In prior guidance, I noted headwinds on gross margin in the second half of the year before returning to historical levels in 2021. I'm happy to report we're ahead of that plan, and gross margins are now expected to be 58% to 59% in Q4, up 56% in Q3. The earlier-than-expected improvement in margins is driven by increased volume, improved mix and supply chain execution in our new product ramps. In Q4, we expect operating expenses to be 29% to 31% of sales. The operating profit at the midpoint of our Q4 guidance is 29%. As Mark noted, you can see we've been positively surprised by both SOC and memory test demand since our last call. Continued strength in mobility test solutions for several customers was the primary driver, but we've also seen unforecasted demand with expedited lead times elsewhere in our test portfolio. For example, analog solutions utilizing our Eagle Test systems and DRAM memory applications for our Magnum EPIC systems have also driven up demand in Q4. We have seen this pattern play out over the last several months. Looking at the full year from a financial perspective. Short-term slowdown in IA has been more than offset by the growth in our test businesses. At the midpoint of our Q4 guidance, our 2020 revenue should be above $3.1 billion and non-GAAP EPS will be approximately $4.50. Gross margin for the full year should be above 57%, down from 58% in 2019, reflecting the impact of high semiconductor shipments for mobility, along with the short-term impact of faster-than-expected ramps of new products. Our 2020 non-GAAP operating profit rate will be in the high 20s, up from 25% in 2019, and our full year tax rate is expected to be 15.5%. Through 9 months, we spent $147 million on CapEx, and we expect we'll spend $193 million for the full year. At the start of the pandemic, it was unclear as to the depth and the duration of the economic consequences. As a result, we took actions to strengthen our liquidity and cash position. For example, we suspended share buybacks on April 1, and expect to end the year with approximately 175 million diluted shares. Given the share repurchases have historically been a part of our balanced capital allocation strategy, we'll update you in our January call on our capital return plan for 2021. Looking ahead at 2021 from a modeling perspective. Macroeconomic conditions, timing of pandemic recovery and political environment all add uncertainty. The environment and demand uncertainty prevents us from speculating on 2021 market sizes. But at a high level, we expect gross margins to return to historical levels and OpEx to grow. From a longer-term model perspective, you can see that at the midpoint of our Q4 guidance will be inside the revenue range and above the EPS range of our 2022 earnings model. We'll provide you an update of the earnings model on our January call. To summarize, we closed out Q3 with outstanding financial and operational performance in a difficult working environment, and I again, thank our employees and partners for their incredible efforts. We entered Q4 with a forecast of higher-than-expected sales and earnings on the strength of mobility and memory. For the full year, share gains in Semi Test and increased traction in our storage business enable us to exit the year stronger than when we entered. The Industrial Automation market is showing early signs of improvement, and we continue to invest in this segment as we monitor the market progress closely. While we're not immune to the macroeconomic shocks and our market visibility is limited, as you've seen since early July, overall, we feel good about our execution during 2020 in a challenging environment. It's too early to estimate demand in 2021, but we'll be prepared from a product, operations and balance sheet perspective for whatever comes our way. With that, I'll turn things back to Andy.

Thanks, Sanjay. Josh, we'd now like to take some questions. [Operator Instructions].

[Operator Instructions] Our first question or comment comes from the line of Vivek Arya from Bank of America.

Congratulations on the strong results and your execution. Mark, I'm curious, how much did 5G contribute to your results in calendar '20 versus what you thought at the start of the year? And the context here is that your large mobility customer now already has all their models with 5G, and I think the U.S. ones already feature millimeter wave. So does it mean that going forward, you are more dependent on unit growth rather than content growth, at least when it comes to mobility-related functions? So just what is the 5G TAM as you see it now and where are you in that journey?

Yes. So just to refresh on the 5G TAM, we've been talking about a $400 million adder to the market, and we're roughly $250 million into that $400 million adder, and we expect that will peak out in the 2023, '24 time frame as the -- in that time frame, we would expect most -- greater than 50% of cellphones would have millimeter wave capability built-in, and that would run at that $400 million level for 3 or 4 years. So that's sort of the baseline. Where we are this year was -- just on the 5G content of cell phones, it's probably about $50 million higher than we expected this year. So certainly, more than we expected, but not dramatically more. So instead of the market being at what we might have thought a $200 million market, it's maybe closer to $250 million. Now the other point you make about, without speaking to any specific customer situation, the market in terms of test equipment for 5G is facilitated to support about 300 million, maybe 250 million units of 5G capability. And so it's still a small fraction of what will happen over the next few years. So now Teradyne's concentration is a different matter. But I do think for the market that remains out there, the other 1.25 billion phones that are going to move to 5G over the next few years, you can assume that our market share in that space is roughly 50%. So we're pretty concentrated in what happened this year, but we have a very good position in what's happening over the next few.

And for my follow-up, I know you're not giving specific 2021 outlook. But I'm curious, after this very strong year that you had across the board in almost every division outside of IA, what of those markets do you think face a tougher comparison next year? That was there -- were there any one-offs, either from an end market or customer or product perspective that you think face tougher comparisons year-on-year as you look into next year, just trend-wise?

Yes. It's probably -- it sounds like a copout, but no, I don't. All the markets that -- in test that showed growth this year have their own story. And the one thing I would maybe cite that I said in the script, that's a bit of a wildcard is the storage test piece, where we've had a phenomenal year, more than doubling of sales there. If -- again, if we listen to what customers are telling us, we think that still has room to run. But it's a thin market in terms of the customer base, and that one is a bit more of a wildcard than the others. But I don't see any other one-off issue with what we've seen in 2020 that would sort of say, "boy, that's going to be tough to beat next year."

Our next question or comment comes from the line of Brian Chin from Stifel.

I guess, first, maybe for you, Mark. Just curious, again, does -- what you're seeing from a test utilization and/or market breadth standpoint, give you at least a degree of confidence that these stronger market conditions are likely to sustain into next year. And I realize that -- again, it's early to provide any official outlook on the test TAM. But can you at least give us a directional sense? Because in a vacuum, if smartphone unit demand grows stay 5% to 10% next year instead of contracting 10%, and 5G mix continues to shoot up, is there a reason to think that yours and the market TAM would not grow next year?

Yes. I think in that scenario that you cite, it would be hard to imagine the market TAM wouldn't grow. And so -- but there's so many variables. We've proven to ourselves and to you that we're not even that good at predicting the current quarter or the next one out. And so we're a little more gun-shy about speculation, I would say. But the issues that you cite around the suppressed unit demand for automotive, cellphones, both are significantly down in units this year and test is still jumping out of the gym. So a recovery next year in those 2 unit volume areas should bode well. But COVID is still out there, the economic situation of the world is unknown. So we're just -- we're not going to go too far out on a limb here and try to speculate.

Okay. That's fair enough. And maybe second, deciding the recovery you're seeing in UR sales in 3Q, and it sounds like that's ongoing in the 4Q. I would assume maybe this is occurring absent any real sizable recovery in key markets such as auto. And maybe you can provide more color on the applications and markets that have begun to exhibit some improvement.

Yes. It's Sanjay. So just, obviously, we've seen growth quarter-over-quarter and a little bit more color. In UR, we grew, in IA, 23%. And actually, it was one of the reasons why we exceeded our guidance this quarter. What we've seen as manufacturing kind of comes back online, we've seen the U.S., Europe and China, bode strength. And really, it's a function of both growth in the quarter. And a couple of those markets, we're actually seeing year-over-year growth. And so -- and then also from an AG perspective, that continues to grow.

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

Congrats on the strong quarter and guidance. Mark, I wanted to ask you on market share in both SOC test and memory test. You're clearly executing really well in both segments. On the SOC test side, I guess the question is, is the growth that you're seeing in market share this year, primarily due to your largest customer having a strong year? Or are you picking up share elsewhere within SOC test? And then similarly, on the memory test side, again, a very big quarter in Q3. How should we think about sustainability into 2021? Are you seeing wins that would support continuity in the current trajectory? Or should we consider this to be more of a one-off? And then I've got a quick follow-up.

Yes. It's a good question because the whole market share year-to-year is very volatile. So how do you make any sense out of it? Just to give you an example, Teradyne's market share last year in SOC was probably around 40%. This year, it's going to be somewhere in the mid-50s. So what's real? So here's how I would characterize it. Last year's 40% wasn't real. It was probably in terms of real market share closer to the mid-40s. We have -- and this year, it will probably be in, like I said, the mid-50s. But what we have done is pick up real noncustomer buying pattern share year-over-year due to design-ins we had last year. And that's to the tune of, let's say, 5 or 6 points of real market share. So off that base of 45 last year, this year, we're probably really normalized running in the low -- 50 to low 50s, is the market share position I'd say we're at today. And there's been a lot going on that influences that. Some buyers of equipment have exited the market like Huawei. That causes some systemic shift in share. And in Teradyne's case, we -- certainly was a huge customer of Teradyne's. But we were -- it was a bit below our average share point. And as that silicon supply redistributes, that helps us. The UltraFLEXplus has designed into some new segments that we've really not participated in for the past decade, that's real new share. It's also designed into some other segments that are new to us. So under the covers of what looks like, mathematically, a 40 to 55, let's say, point share gain in 2020, what's really, I would say, happen as we've gone from mid-40s to 50 to low 50s based on that. In memory, now memory is a market where we see the TAM growing significantly this year. It's gone up $300 million from $600 million to $900 million. About $200 million of that's in DRAM and about $100 million is in flash. And fortunately, we were successful in entering that DRAM segment last year with our new Magnum EPIC. And so we've been able to grow as that DRAM segment has grown this year. And that brings us to, I'd say, our true market share and our reported market share are going to really be close in memory in the low 40s. And what we've said is that given our footprint now with our products and customer base, we legitimately see a way to get that to that 50% level over the next few years. So maybe that's a long-winded answer, but those are the 2 touch points I'd give you.

And as a quick follow-up, I wanted to ask on Industrial Automation profitability. I suppose over the past several years, you've been in investment mode in the business, growing your distribution network, investing in R&D. I think currently, you're around breakeven, given muted revenue levels. But how should we think about kind of through-cycle profitability in IA? And if you could remind us what your long-term margin target is for the IA business, that would be super helpful.

Sure. It's Sanjay. So long question there, but maybe I'll give you a little bit of color. Recall in 2019, our IA business had an operating profit of about 10%. We entered the year in 2020, a big investment year, but we expect it to be around that same operating profit. COVID hit a lot of the go-to-market OpEx, we reduced, obviously, because it was run rated towards the higher revenue levels. And in the first half of the year, we lost money. In Q3, we were slightly profitable. And as we've said in -- or in Q4, we expect revenues to continue to increase and to grow that profit level. We haven't concluded on our 2021 plans, but my expectation, assuming that revenue continues, we're going to then obviously increase some of the go-to-market, obviously, with travel, trade shows, et cetera. But the engineering investment has not really wavered from that standpoint. And so we expect to continue to heavily invest in Industrial Automation. So over the next several years, you would expect that our profitability wouldn't be at kind of the company average because it's still in investment mode as we grow. And then what we said over the horizon is that -- sorry, over the midterm, is that we expect the revenues to grow between 20% to 35%.

Our next question or comment comes from the line of C.J. Muse from Evercore.

I guess, Mark, wanted to, I guess, go back in time and use your memory because I can't remember. But if you look back to AP and all of mobility, in the kind of '14, '15, '16, '17, '18 period, despite, I guess, some consternation around Apple fully buying new testers. You continue to see strength in that business as other players came in. So curious, as you look back in time around the 4G ramp and then look forward to 5G and what you're seeing in terms of silicon content on the RF side, in particular, millimeter wave as well as some of the processing capabilities required. How does that make you feel around what you think the trajectory of mobility tests could look like over the coming 1, 2, 3 years relative to the strength that you're seeing here in 2020?

Yes. Good question. So just to put it a little bit in perspective. So the TAM or the test market for mobility, if you go back to, let's say, 2017, '16 time frame was roughly $1 billion of test equipment for mobility. This year, we estimate it will be close to $1.6 billion. And that's 1 data point. And that's in light of lower unit volumes. Another data point I'll give you is that if you look at a 4G phone and add up all the silicon in a typical 4G phone from a few years ago and say, "how much time does it take to test all that silicon?" And then you take, let's say, a high-end 5G phones today that are just coming on the market and ask the question, "how long is it to test all of that silicon?" And all I'm talking about now is the RF content and the apps processor. I'm not talking about the cameras and the power management and all that other stuff. It's up about 60% on a per unit phone basis, and we're just getting into 5G content phones. So you can see that there should be, as long as the units stay reasonable on cell phones and 5G keeps rolling out, a lot of tailwinds around continued growth of the mobility camp.

That's very helpful. I guess as a follow-up question. You talked about mid-50s share here in 2020. On the last earnings call, you talked about a target of 60%. And I guess, can you kind of speak to what can bridge those 5 points, particularly around -- anything you can say around wins with UltraFLEXplus as well as the fact that auto and linear are probably down 35% this year versus last year. So I would think simply right there, that could bridge you to 60% in 2021.

Yes. So just on the share, first of all, I'm not going to use, as I mentioned earlier, 55% is our normalized share right now. I'd say it's somewhere around 50, plus or minus. So how do we get 10 more points is probably the right way to think about it. And there's a clear path there in my mind. So 1 piece of it is, as you cited, when automotive and linear comes back, that's a place where we have higher-than-normal market share, and that will be a bit of an inflator to our share. The UltraFLEXplus, though, is a key component of the strategy. We've already secured some segments that we've not participated in, in a long time. They're just beginning to ramp this year. So that platform designing into those segments alone should bring us halfway at least to that 60% or extra 10 points we need to get there. So if analog brings us a couple of points, and the UltraFLEXplus brings us maybe another 5, now we're talking -- we're up in that sort of 7-ish range. And then the last piece that I alluded to in my script is there's this -- and this is another UltraFLEXplus story, I think. But edge AI devices are growing. Our position there is encouraging. And I see that, that segment as it establishes itself, if we can get that early design wins to kind of ramp with that growth of devices at the 60% share kind of level, that can pull us the rest of the way. By the way, we're talking -- maybe this is a 5-year, 6-year journey. It's not something that's around the corner.

Our next question or comment comes from the line of Mehdi Hosseini from SIG.

A couple of follow-ups regarding your Q3. So if upgrading some of the installed testers that last year were used for Huawei helped with the Semi services. How should I think about looking forward and its impact on the SOC tester? To what extent system upgrades of the past 6 months would have adverse impact on the new SOC test demand? And as a second question or a follow-up, can you elaborate on the mix of HDD system? And to what extent the strength in advanced packaging like chiplets had the impact there? Any detailed color that would help us separate HDD from system-level test would be great.

Yes. It's Sanjay. I'll take the first one. So from a services perspective, yes, in the quarter, we did see some drop in or supplies' demand. And really some of it was tied to the upgrade of kind of existing testers that are being kind of reconfigured at OSATs to test other silicon -- providers of silicon. And it was a very high quarter from a services perspective. We did have some increases on new testers as well. But we expect that business to kind of continue. At or -- once we get through, I would say, the migration or repurposing of existing testers of the OSATs for other customers. So we expect that to be a little bit up in the very short term but get back to normal over the midterm.

And then the HDD and SLT.

Yes. So on HDD and SLT. So roughly, the way to think about it is, today, the revenue split there is almost 50-50. It fluctuates, but roughly 50-50. Going forward, there's a lot -- there's actually upside on both sides of this. So on the HDD side, what really has been driving that business historically has been exabyte growth rates. But the -- this is a similar story to test intensity. These 18-plus -- giga, terabyte-plus drives that are being produced and the roadmaps that we have, have rather sophisticated electronics to allow for high-density error-free read/write operations that require more test intensity. So we're getting a multiplicative effect on the HDD side a bit on top of exabyte growth. So we are optimistic that, that leg of the stool in storage test can continue to grow. SLT is a little more -- a harder thing to... [Technical Difficulty] So the -- on the other side of it, the SLT side, chiplets and things like that, certainly, that complexity presents an opportunity for more system-level test, testing the module, so to speak, before they go into a higher value-added subassembly is going to be a premium. So that's the theory. And we're in various trials around various customers to see if economically, it makes sense to -- for those class of parts, add a test insertion. It's made sense for some mobility parts. But remember, those are billion of units a year. So the question really that we haven't gone out on a limb yet and really answered for ourselves is how far down the unit volume equation can SLT go to make sense. So that one, I'd say, is still a bit of a wildcard.

Our next question or comment comes from the line of Atif Malik from Citi.

First one, for Mark. Mark, if you can remind us about your opportunity on the compute side. There's a discussion that the die sizes for the CPUs for a North American customer notebooks next year, it could be 3 to 5x larger than what's used in tablets. Can you just talk about the compute opportunity for Teradyne this year and next year?

Well, compute is an area that Teradyne really hasn't participated in since the '90s. The UltraFLEXplus was specifically designed to give us that opportunity to break back into that segment. And again, roughly, just to give you numbers, you can think of compute is somewhere in that $600 million a year TAM in the total market right now. I expect that's going to grow because of the diversity of suppliers of compute devices is growing. There's a bit of a disaggregation of the supply chain for compute devices. So the UltraFLEXplus is targeted there. It's had some success already in that mission, and we expect that will continue. So I do think it's a rich area for us. But to put it in perspective, it's a market that's about maybe around $600 million on average now, could grow to $800 million in the next year or 2.

Great. And then Sanjay, there was no material impact to your sales from military end use China restrictions, like you talked about. But can you talk about if the demand from domestic China semiconductors, memory and logic was higher than what you expected in both Q3 and Q4? And how big is domestic China semiconductors as a percentage of your total sales?

Yes. So maybe I'll start with the end. So roughly 15% in Q3 of our sales were China based. And so it's true that with the regulations, either tied to Huawei or discussions about SMEC or the military end use. Really from a military end use perspective, there's a tremendous amount of internal compliance work that we're doing to ensure that we are compliant with the regulations. And it has not had a material impact on our revenue stream. And I think in earlier remarks we've made is that as the end market remains the same. And the different silicon providers are providing solutions for that end market, we feel good about our position on that front. So it's a -- there are some short term disruptions. We talked about service upgrades for testers that are being redeployed at the OSATs that we're servicing high silicon solutions and now going to be servicing other solutions. So that's the type of stuff we're seeing in the short term. But in the long term, we feel good about our position because we see it as a share shift where we're positioned well.

Our next question or comment comes from the line of Krish Sankar from Cowen & Company.

I have 2 of them. Mark, thanks for the qualitative color on calendar '21, and I understand it's very hard to quantify it at this time. But if you look at the SOC test market, it's kind of grown for the last 4 or 5 years. A lot of it is because of increased complexity. But the memory test market in the last 4 years has been more cyclical up and down. Do you think memory is at an inflection where you could see as LPDDR5 and everything else comes along that you might be in a prolonged strength for memory like SOC? And then I have a follow-up.

Yes. So on memory, you're right. Absolutely, it's been cyclical for the past few years. There are good compelling arguments that it could be pretty less volatile and more sustained high-level investments over the next few. Certainly, our memory team believes that. And because of LPDDR5 and DDR5 transitions that are imminent, that's going to compel a lot of investment on the DRAM side. So I just don't see the downside in memory nearly as typical as we've seen in the past decade. I think the DRAM story alone is going to keep the memory market pretty healthy. So then the question is flash. And flash is a tougher one to read. The only thing I'd say that's encouraging is this relentless innovation around the interface speeds of flash drives a lot of tester demand and drives obsolescence of the fleet. And the roadmaps there are pretty robust. These cellphones and the camera systems in the cellphones just generate an incredible amount of data that needs to be quickly moved in and out of flash. So I'm kind of optimistic that memory in the next 4 years won't be as volatile as it's been, let's say, in the last 4 or 5.

Got it. That's helpful. And then as a follow-up, on the Industrial Automation side, how much of your IA revenues are from the auto vertical? And how much are coming from small and medium businesses?

35%.

Yes. It's Sanjay here. So from the auto perspective, it's in the neighborhood of 35%-ish. And sorry, your second question?

How much from SMB?

Small and medium size business.

I don't know. We'd have to dig that one out. But it's been running in that sort of, I'd say, 75% range. So it's still the vast majority of our sales. The larger enterprises have been a growing percentage over the past 3 to 4 years. But that's a rough number.

Our next question or comment comes from the line of John Pitzer from Crédit Suisse.

Mark just a follow-up to Mehdi's earlier question on Huawei. To the extent that they were historically a large customer, a lot of those testers are sitting at OSATs and can be repurposed. I'm just kind of curious, do you -- how much insight do you have into kind of that installed base and its ability to move to additional customers or new customers? And I guess, are you putting up these results despite the fact of that? Or is that something that we need to kind of worry about in future quarters?

So we have decent visibility into the utilization of those testers. But the high-level picture I'd give you is, let's say there's 100 testers out there for Huawei that we're testing 1 billion parts a year. As Huawei ends up unable to supply those 1 billion parts, the first immediate reaction is, oh, that's 100 idle testers. They've got to go somewhere. But the fact of the matter is that somebody else has to supply those 1 billion parts because the end-user demand isn't changing that much because of this. So -- and typically, by the way, those suppliers at the moment are outside of China suppliers that are filling in that billion unit gap to use this analogy. So that -- and then those testers that were idle, just sort of shift from testing parts that had the high silicon Huawei brand stamp on them to testing parts that have these other suppliers' brands on them, but they're very similar in nature. So the upgrades to facilitate that shift of the fleet over to those new sources of supply started in earnest in Q3. And so despite the fact that there's already repurposing happening at a pretty good clip of those testers, we're putting up the numbers we put up, and we're guiding the numbers we're guiding, and we expect that will continue in Q4. When will that entire fleet get repurposed is kind of the last end game question in that, and that's harder to tell. We may end up not seeing that fully flushed out until somewhere in the second quarter of next year.

That's really helpful. And then Mark, as my follow-up. In the quarter, you guys created a new role in the IA division and you put in place there a very incredible guy in the form of Greg Smith. I'm wondering if you could spend a little bit of time just kind of helping us understand the rationale and how we should think about this -- Greg's sort of efforts in IA and what it means to the overall strategy there?

Yes. Thank you for asking that because Greg is a very proven exceptional leader and has driven our semiconductor test business here for the past 5 years. And it's a significant testament to both what we, as a company, and Greg personally believes the opportunities are in IA for future growth. And Greg's mission really is in looking at the portfolio we have is to look for areas where we can get leverage, enhance our M&A strategy, look for common investments that will give us both in software and hardware, further differentiation. And sort of bring more of the resources of the company to bear on those businesses. Most of them are now through their earn-out phase, which gave us a little bit of -- there were a little bit of handcuffs on there of how we could think about operating those businesses. But now that that's behind us, we have a lot more degrees of freedom. And Greg is going to be a great resource to figure that out for us.

Our next question or comment comes from the line of Sidney Ho from Deutsche Bank.

I have 2. My first question is on the SOC test market. The mobility is roughly half of the SOC TAM today. I think, Mark, you mentioned $1.6 billion of the $3.3 billion this year. Can you remind us what are the other main buckets within the SOC test market? And which segments are still below the normal run rate? Maybe help us size how much below the normal levels in some of those markets, please?

Okay. I'll give you sort of the rough numbers. So there's a compute segment that is CPUs and FPGAs and AI and all that good stuff, that's roughly $600 million that we just spoke about. And it's actually a little bit hot, too. I'd say it's running -- it's another growth segment for some of the reasons I talked about earlier. There's automotive and linear and microcontrollers. That this year -- and again, these are rough estimates, but it's roughly around $200 million to $250 million market. That's less than half of its normal run rate. A normal run rate for that business would be $500-ish million. So that one is certainly -- and it's down from last year as well. So it's trending down and it's half. And then the last one we track is industrial. Industrial is somewhere like around $300 million. And it's also down from a normal run rate that might be closer to -- it's not as volatile, but it's probably up in the high 3s to $400 million.

Great. That's super helpful. My follow-up question is as you talk about an increase in demand for short-term lead time -- lead time businesses in Q3. Does that -- does your 4Q guidance assume any kind of headwinds from any kind of supply constraints in any of the product lines?

Right now -- it's Sanjay here. So right now, we're not -- we believe that there could be some issues that arise. But the last 6 months has clearly shown us that we can overcome what those hurdles are. Nothing is perfect, but we feel that we're in a very good spot in replenishing our components and manufacturing. And if there are things that come up, we've got some capability to solve those problems. So at this point, there's no supply constraints.

Our final question or comment comes from the line of Timothy Arcuri from UBS.

So I had 2. The first thing, Mark, I'm still trying to figure out on the SOC share point. The TAM of $3.3 billion, it's not a whole lot different than what you thought it would be at the start of the year. And at that time, you thought your share would go from roughly 40% to mid -- to roughly mid-40s. Here we are, we're kind of mid-50s this year. So I guess I'm trying to understand where the incremental 1,000 basis points of share came from. Was this your biggest customer ordering more? Was this something else? Maybe you can double-click on that a little bit, and then I have a follow-up. Because it sounds like it's not really 5G, you said that 5G is only like a $50 million swing factor for the whole market. So I'm still trying to figure out where that share came from versus your expectations?

Yes. So at the highest, highest level, it's just that the market -- the supply side of the semiconductor market has shifted more favorably to our customers. But underneath that, why is that? Well, one thing we didn't see happening when the year started was that further regulations would come into place on Huawei and drive them out of the market for silicon and test. And since that occurred, what happened, as I described a bit earlier, was other suppliers have had to step in and fill the void, and we're better positioned at those other suppliers. So that is an upside. Our customers, our traditional customers, independent of that particular situation, are also buying more than we had forecast. So that is absolutely another piece of the upside. I'd say those are the 2 most significant ones. And then the third one, which isn't an SOC thing, but I think is a total company thing is on the memory side, the fact that the market has been so hot, and that our LPDDR5 ramp has been probably better than we expected, has helped on that side as well.

Got it. And then I guess for my second question, just on the point about Huawei. So you saw the original bump in the capacity required for Huawei from the OSATs. And now you're seeing some boost from upgrades to that fleet. Can you sort of quantify how much of a boost these upgrades are giving you? So I guess, how much of the [ $449 million ] SOC revenue was related to the upgrade of this fleet for non-Huawei OEMs?

Yes. It's less than 10%, but more than 5%.

Okay. Folks, we are out of time. Thanks so much for joining us again this quarter, and I'll follow-up with folks individually that are still in the queue. Again thanks so much. Take care.

Ladies and gentlemen, thank you for participating in today's conference. This concludes the program. You may now disconnect. Everyone, have a wonderful day.