Ladies and gentlemen, my name is Simon, and I will be your conference operator today. At this time, I would like to welcome everyone to the Teradyne Third Quarter 2019 Earnings Conference Call. All lines have been placed on mute to prevent any background noise. After the speaker's remarks, there will be a question and answer session. If you would like to ask a question during this time, simply press star then the number one on your telephone keypad. If you would like to withdraw your question, please press the pound key. Thank you. Andy Blanchard, you may begin your conference.
Thank you, Simon. 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 the 2019 third quarter, along with our outlook for the fourth quarter of 2019. The press release containing our third quarter results was issued last evening. We're providing slides on the investor page of the website that may be helpful to you in following the discussion. Replays of this call will be available via the same page after the call ends. The matters that we discuss today will include forward-looking statements that involve risk factors that could cause Teradyne's results to differ materially from management's current expectations.
We encourage you to review the safe harbor statement contained in the earnings release, as well as our most recent SEC filings. Those forward-looking statements are made as of today. We take no obligation to update them as a result of developments occurring after this call. During today's call, we'll make reference to non-GAAP financial measures. We've posted additional information concerning these non-GAAP financial measures, including reconciliation to the most directly comparable GAAP financial measure, where available on the investor page of our website. Between now and our next earnings call, Teradyne will be participating in investor conferences hosted by Baird, Cowen, Credit Suisse, and UBS. 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 fourth quarter.
Sanjay will then offer more details on our quarterly results, along with our guidance for the fourth quarter. We'll then answer your questions. This call is scheduled for one hour. Mark?
Thanks, Andy. Good morning, everyone, and thanks for joining us. In my prepared remarks, I'll cover three topics. Highlights from the third quarter at the company and business unit level, how our recent agreement to purchase autonomous mobile robot maker AutoGuide fits into our industrial automation strategy, and how we're looking at next year and beyond. Sanjay will provide the financial details of the quarter, our view of the 5G test opportunity, and our guidance for the fourth quarter. We delivered results above our third quarter guidance on continued strength across all of our test businesses. Our MiR mobile robot business also performed well in the quarter. Sales at Universal Robots were roughly flat year-on-year, as the impact of softening manufacturing activity in Europe and the U.S. were a strong headwind. Our year-to-date revenue is up 4% from 2018, and non-GAAP EPS is up 13%.
For the fourth quarter, we expect strong test demand to continue, combined with single-digit growth in IA on a year-over-year basis. At the business unit level, SemiTest SoC test demand for digital broadband and RF devices related to 5G infrastructure remained very strong. That infrastructure demand, combined with strength in handset-related test, is driving our estimate of the SoC market size for 2019 up to a $3 billion-$3.2 billion range compared to last year's $3 billion level, and above our earlier estimate of $2.6 billion-$3 billion. Geographically, China's SoC test market will be up $300 million year-on-year, while the rest of the world's market will likely be down. This reflects strong tooling for 5G infrastructure and handsets as the combination of device complexity and base station unit growth are fueling demand.
On the product front, in early September, we introduced an extension of our flagship UltraFLEX product, the UltraFLEX Plus. The Plus extends our performance for an emerging class of AI and big data devices that require a new test architecture to efficiently and comprehensively test the rapidly growing complexity of these devices. Multicore chip architectures are driving billions of transistors per die, which operate in parallel to process the vast data streams associated with machine learning and 5G networks. This is yet another example of the complexity growth we've been describing that drives up test seconds and drives our markets. The UltraFLEX Plus puts in place an architecture with the headroom to test this complexity growth for the next decade. At the same time, it maintains software compatibility to leverage the accumulated FLEX family software experience of over 10,000 trained customer test engineers.
Systems are now in use at multiple customers. We recognized our first revenue from product in Q3. In memory test, demand remains strong in Q3, driven by both flash package and DRAM wafer test. Although the memory test market will likely be down about 40% this year to around $600 million, our market share will be up from about 29% last year to the low 40% level this year. Building on our strong foundation in flash package test, our Magnum product family now covers the entire memory test market, wafer and package test for both flash and DRAM. LitePoint and the System Test Group sales were about flat with the second quarter and up 23% and 48% respectively from the third quarter of 2018. Advanced connectivity standards and early 5G-related buying powered LitePoint's results.
While in System Test, our storage test products were the standout, as revenues more than doubled compared to a year-ago quarter on terabyte-class HDD demand and system-level test shipments. Shifting to our Industrial Automation segment. Universal Robots sales were flat with the third quarter of 2018 as demand in our largest geographic markets, Europe and North America, and our largest end market, automotive supply chain, remained weak. Conversely, this was balanced by strong demand in China and Japan, with year-over-year growth in the quarter running 30% and 44% respectively. MiR, on the other hand, continued to grow at a healthy clip, posting year-over-year sales gains of 47% in the quarter. We believe the lower year-to-date growth at UR is a short-term phenomenon, consistent with the slowing of industrial activity in Europe and North America.
While our automation solutions are not immune to the decline in industrial output, we are faring better than the market in general. We also see opportunity to expand our market coverage in the U.S. and Europe to offset these headwinds, and we'll be investing to support underserved verticals and geographies in these regions. In addition, we continue to expand our product offerings. In September, we introduced the UR16e, which opens up new applications by extending the payload capacity of the product line. We also shipped our first industrial bin picking application in the quarter utilizing a UR robot, 3D vision, and Energid's unique path planning software to ease deployment and open new applications at new customers. UR also passed the 200-product threshold of UR+ certified plug-and-play peripherals, expanding the range of applications and reducing the deployment time for our cobot customers.
MiR has also introduced a similar plug-and-play strategy for peripherals called MiR Go. Both programs are consistent with our strategy of making automation safe, low risk, and easy to deploy by shop-level technicians with short ROI. With the announcement earlier this week of our plan to acquire AutoGuide, a maker of autonomous mobile robots for industrial and warehouse material handling, we are expanding our capabilities to include high-payload pallet and material moving. The global market for manually driven forklifts is primed for automation, as technological advancements have enabled cost-effective autonomous solutions. AutoGuide's products use many of the same sensing, guidance, software, and drive technologies as MiR to provide autonomous operation to higher payload tugging and pallet moving applications. Additionally, AG's unique modular architecture offers customers greater flexibility than either a dedicated forklift or tugger with reduced cost and safer operation.
Like our other IA investments, AutoGuide is a small but differentiated and fast-growing player in a nascent market. We expect full-year sales to more than double to over $10 million this year. There are natural synergies in engineering, marketing, and distribution with MiR, and we expect our broad lineup of low and now high-payload autonomous material handling robots to be attractive to global customers. As you can see from our guidance, we expect to finish 2019 on a very strong note. As we've described in the past, our visibility in the test business is a quarter or so, and in industrial automation, less than a quarter. We expect volatility in our markets, we also expect overall trend line growth consistent with our midterm growth plan and earnings model. Teradyne's cost structure and business plans are architected with this volatility and growth in mind.
In SemiTest, SoC test has seen another strong year despite an overall soft semiconductor device market. As I noted at the outset, this year's strength in China for testing smartphone handset and base station devices has more than offset the decline elsewhere in the world in automotive, linear, and MCU test. As we look to 2020, with the aggressive 5G base station ramp, will the aggressive 5G base station ramp continue, or will there be a pause for digesting? Will 2020 phones adopt 5G mmWave in meaningful volumes? Will the health of the memory business improve? Will the automotive and industrial market stabilize and recover or further contract? All of these questions have implications for our 2020 business levels, all are hard to predict. Frankly, we don't spend too much energy on trying to predict this as it doesn't really affect our plans.
Our strategy and plans are based on the fundamental belief that the pervasiveness and complexity of semiconductors will continue to grow on a trend line and continue to drive growth in our test business. Similarly, we believe the inflection point we've reached in industrial automation, where new, powerful, yet cost-effective technologies like LIDAR, 3D vision, AI, and others have opened new markets, will drive opportunity and growth for decades to come. We strengthened our competitive position in our test and IA businesses with a steady flow of new products through the year, and AutoGuide adds another exciting and technically differentiated building block to our stable of advanced automation tools. We don't know what 2020 will bring in terms of test and IA demand. We are confident that our product lineup, financial position, and growth strategy positions us well for the long-term success.
With that, I'll turn things over to Sanjay for financial details.
Thanks, Mark. Good morning, everyone. Today, I'll review the third quarter results and performance of each business, comment on the full-year results at the midpoint of our Q4 guidance, provide some insight on how we're looking at mobility as a long-term driver of the test business. Then provide additional detail on the AutoGuide acquisition. Lastly, I'll offer some early financial information for modeling 2020. Now to our results for the third quarter. Third quarter revenue of $582 million was above the high end of the guidance due to strength in our Semi Test business. Sorry, due to our strength in our test businesses, partially offset by lower industrial automation revenues. We had one customer greater than 10% of sales in the quarter. Gross margin of 59% was driven by favorable product mix in Semi Test.
Non-GAAP operating margin of 28% and non-GAAP EPS of $0.77 were above the high end of the Q3 guidance, driven by slightly higher revenues with improved product mix and lower spending relative to the third quarter plan. You'll see our non-GAAP OpEx were $185 million, down $5 million from the second quarter due to seasonally lower spending in Industrial Automation and lower project spending in Semiconductor Test, slightly offset by higher variable compensation accruals on higher profits. Turning to the individual businesses. Semiconductor Test revenue of $398 million grew 6% over Q2. Notably, memory test revenue of $72 million grew 23% over the prior quarter, driven by flash, final, and DRAM wafer test demand.
As I've met investors over the last several months, one of the common questions asked revolves around the surprising strength in the handset test demand in an environment where volumes are down year-over-year, and the transition to 5G is in its infancy stage. I see two primary drivers of this trend. First, devices at each tier of the phone are getting more complex. Typically, leading technology gets introduced at the premium tier or devices greater than $600. Innovations such as lower process nodes for power and performance, increased compute capability, more camera sensors, and increased memory requirements on the device are a few examples of higher complexity. This complexity enters at the premium tier, and then in one to two years, key technologies waterfall down to high-tier devices. This waterfall down of technology continues from high to mid-tier devices and mid to low-tier devices.
A continuous increase of device complexity drives the need for more test time per device. Second, in the smartphone market, historically, people who purchase a low, mid, or high-tier smartphone typically upgrade at some point to the next tier up. These two drivers increase the test TAM. While the mobility market is not without volatility, we do expect it to be a multi-year positive driver of our SemiTest business. Moving on to system test. The group continued to deliver strong results in Q3 with revenue of $73 million. Within the System Test Group, defense and aerospace grew sequentially on increased defense-related buying. Production board test grew quarter-over-quarter and was slightly ahead of plan.
As Mark noted, on a year-over-year basis, we delivered 48% growth in System Test in the third quarter, while storage test more than doubling, benefiting from growth in terabyte drives and cloud storage. Defense and aerospace grew 30% year-over-year as we are seeing continued growth in long-term projects within many branches of the military. For the full year, System Test is on track to grow over 25% from last year's $216 million of revenue. At LitePoint, we saw continued momentum with sales of $42 million on the strength of new Wi-Fi standards and early buying of 5G development test. For the full year, LitePoint will grow about 10% from $132 million last year. In Industrial Automation, sales were $69 million, down 8% from Q2 and up 4% year-over-year. Within IA, UR was $58 million of revenue, a decline from Q2 and about flat year-over-year.
MiR revenue of $10 million was slightly down quarter-over-quarter and up 47% year-over-year. As Mark noted, we view the slower growth in IA so far this year primarily as a result of a slowdown in global industrial investments, especially in Europe and North America, and not a long-term market detractor. However, we also recognize we have to continue to improve our organizational capability in order to fully capture the opportunity ahead of us. We will continue to invest to strengthen those capabilities. Turning to the acquisition of AutoGuide. The upfront purchase price was $58 million, with an earn-out of up to an incremental $107 million if certain revenue and profit targets are achieved through 2022. We expect the transaction to close later this quarter post regulatory approvals.
AutoGuide is on track to achieve revenue of over $10 million on a standalone basis in 2019, up from $4 million in 2018. We expect AutoGuide to grow over 100% in 2020. Their sales today are primarily in North America, with a plan to broaden their distribution to serve global customers. We expect them to incur a small loss this year and be slightly positive on an EBITDA basis in 2020. Relative to share buybacks, we forecast this acquisition will be about neutral in EPS in 2021 and will be accretive on an EPS basis in 2022. AutoGuide will operate as a standalone business within the industrial automation reporting segment.
Partnership between UR, MiR, Energid, and now AutoGuide, coupled with our central core competencies, will enable them to efficiently and effectively scale their business, taking advantage of technologies, processes, external relationships, and experts across all of Teradyne's businesses. As stated in prior calls, our industrial automation acquisition strategy is to acquire growth companies with leading technologies serving nascent markets that have short ROI and make the workplace safer and more productive. AutoGuide is another example of Teradyne executing on the strategy. AutoGuide has a modular, purpose-built architecture for their autonomous mobile robots, along with intuitive user and fleet management software. We believe this approach differentiates them from the competition, enabling flexible, scalable solutions with compelling economics to be efficiently implemented at a customer's location.
We've provided a slide in the earnings deck which summarizes the transaction along with the schedule of the earnouts so you have a sense of the revenue growth required to reach the performance targets. Additional AutoGuide details are included in the deck's supplemental information. Back at the company level, our operating model continues to deliver strong results. We generated $162 million in free cash flow in the third quarter, and we ended the quarter with $1.040 billion in cash and marketable securities. We repurchased $122 million of shares in the third quarter. This brings the year-to-date share repurchase total to $369 million at an average purchase price of $41.93. We paid $15 million in dividends in the quarter. For the full year, we expect to repurchase $500 million of our shares. As usual, we'll update our capital allocation plan for next year in the January earnings call.
Turning to our guidance for the fourth quarter. Revenue is expected to be $590 million-$630 million, and a non-GAAP EPS range of $0.73-$0.84 on 174 million diluted shares. Q4 growth will come from both our test and industrial automation businesses. Historically, our SemiTest business has been seasonably lower in Q4, but this year, continued strength in 5G infrastructure and memory will drive growth. The guidance excludes the amortization of acquired intangibles and non-cash convertible debt interest. Fourth-quarter gross margin should run approximately 58%, and OpEx should run from 31%-33% of revenue. The operating profit of our fourth-quarter guidance is forecasted to be 25%-27% of revenue. A quick comment on diluted shares. There is a GAAP requirement to measure and disclose EPS with the most dilutive calculation.
We have applied this concept to the non-GAAP EPS with respect to the treatment of our convertible debt. Diluted shares are higher in both Q4 guidance and Q3 results versus Q2 at current share price levels. It is more dilutive from an EPS perspective to assume the convertible debt has converted and include the associated shares than to include the interest expense related to the debt and not include the shares. Shifting to taxes. Our non-GAAP full-year tax rate is expected to be 16.5%. We expect the tax rate to be 16% in 2020. Related to modeling 2020, as in past years, you should expect Q1 industrial automation sales to be seasonably down versus fourth quarter. From a full-year perspective, at the midpoint of our Q4 guidance, we expect sales to grow 7% and non-GAAP EPS 16% from last year.
From a longer-term perspective, this is the sixth consecutive year of non-GAAP earnings per share growth with a CAGR of 17%. Over this period, we averaged a non-GAAP operating profit of 23%. In summary, we are delivering very strong results across all of our test businesses. In particular, SemiTest over the past few quarters has enjoyed a wave of 5G infrastructure investments that will continue in Q4. However, we expect the global 5G rollout to drive lumpy test demand. As Mark described, we don't have much visibility into 2020 test demand. Our operating model has been tuned to handle any potential volatility. The industrial automation portfolio is well-positioned to get back to consistent levels of growth when global industrial spending recovers and will be further bolstered by AutoGuide's rapid growth. We're continuing to take actions necessary to drive incremental demand and investing to strengthen our position.
With that, I'll turn things back to Andy.
Thanks, Sanjay. Simon would now like to take some questions, and as a reminder, please limit yourself to one question and a follow-up.
Ladies and gentlemen, at this time, I would like to remind everyone that in order to ask a question, please press star, then the number 1 on your telephone keypad. We'll pause for just a moment to compile the Q&A roster. Your first question comes from the line of Vivek Arya with Bank of America. Your line is open.
Thank you for taking my question. Congratulations on the strong results. Mark, I'm curious, a number of semiconductor companies had suffered this year because of pull-ins from China last year. I'm curious, how sensitive is your 2020 outlook if China demand normalizes versus what we are seeing right now? I understand that the visibility at this time is low. Do you have some sense of what the utilization is of the testers bought by your Chinese customers so far this year?
Yeah. A couple of comments. Utilization of the existing installed fleet of testers is very high. As we described in Q4, we have still continued shipments and strength into China, so the pull for demand through Q4 is strong and unabated.
I also mentioned in my remarks that China is the one geography this year that's up for test, where the rest of the world is down. The key question is, what's normal? I think there's been a heavy investment for 5G infrastructure test tooling this year. We've been surprised as well by how long it's continued. The view going into next year is that there should be a little bit of a pause in some of the rate that they're adding capacity, but that the long-term trend line is still intact, and that could pick up again in the back half of the year. Frankly, the forecasts that we've been looking at throughout this year, as you can see from our results, have been somewhat weak, and the actuality's been stronger. It's really hard to predict at the moment.
All right. As a follow-up, I think in the past you outlined, I believe, around a $300 million-$400 million opportunity if all smartphones get converted to 5G. I think from what TSMC said, they expect 5G penetration to be about mid-teens next year. Does the math work in a linear way? Let's say mid-teens smartphones are converted, then does it mean it's like a mid-teens part of the $300 million-$400 million TAM? Do you have some similar math from a base station perspective also?
That's a good question. The $300 million-$400 million increment is based on 5G in general. What we've said in the past is that somewhere around 2021, 2022, when the majority of smartphones cross over with 5G mmWave capability, we would see that kind of bump. Much of what we may see next year in smartphones for 5G is sub-6 GHz, which has an impact, but not as much as mmWave. However, the infrastructure demand this year has been so strong that what I would suggest is that of that $300 million-$400 million, we're probably this year seeing $200 million or so of that bump from 5G. The infrastructure piece comes first.
That will likely wane a little bit as we go through the next few years as the capacity gets built up to run out a lot of base stations, and then the phone thing will kick in. It'll kick in a little bit next year for sub-6 GHz, and then in subsequent years, to a greater extent with mmWave, is how to think about it.
Thank you.
Your next question comes from the line of Timothy Arcuri with UBS. Go ahead, your line is open.
Thanks a lot, Mark. I just wanted to follow up on what you just said. We began the year for the SoC TAM was about $2.5 for this year, then we went to $2.8, and now we're at $3.1. If 5G is adding maybe $200 to that $600 increment, where's the other $400 coming from?
Well, there's strength in, I would say, general computing, and sort of AI-type big digital applications out there. I think there's also strength in handsets in general, not 5G related, but just 4G. The complexity of the processors, the extra sort of MIMO and RF side of the 4G phones this year has driven a lot of test demand as well. It's sort of all on the digital side, and I'd say phones, 4G phones, are the principal other half of the equation.
Okay. Got it. Thank you. I guess I had a question in industrial automation. With respect to UR, obviously the business is suffering from some industry trends, but we've also heard about some price competition. There's a lot of offerings now coming at a lower price. I think Techman is now partnered with Omron. How do you think about that in the context of you went out and spent a decent amount of money for a company that's not really going to be accretive to earnings for a couple of years? There seems to be a lot of sort of investor concerns. I get a lot more questions about this, that people are getting a bit concerned that there's some deterioration happening in your core business. Can you sort of talk about that and what your longer-term strategy is?
Also, whether you think that this changes your view of this being a billion-dollar business by 2022 or something like that? Thanks.
Yeah. Hi, it's Sanjay here. I'll comment on the UR pricing competition first. I think, first of all, UR margins actually ticked up this quarter slightly. When we look at our win-loss and evaluate the competition, pricing is obviously out there from other competitors that's much lower. However, we haven't lost significant business on that front. I think we have a quality product offering. It's recognized well by the customer base, and it's being priced appropriately. I think from a UR standpoint, yeah, we are planning that competition eventually will get the quality levels up, and we'll have to deal with a little bit of pricing. Currently, that's not necessarily the case where we're losing business from a price perspective.
From a long-term perspective on AutoGuide, I think in both Mark's and my prepared remarks, I think you need to consider this as a complement set of offerings in our portfolio of higher payload, of which there'll be some potential product go-to-market synergies with MiR and an entire portfolio. That coupled with the scale of Teradyne and the core competencies, we believe very strongly in completing out the portfolio of the industrial automation, and this is just another example.
I would also, just on the specific comment around Techman and Omron, we've talked about them as, it's an interesting product. It's really not a price deflator. Even in its home turf of Japan, that's our strongest growing region this year with UR. I think the thought that somehow price competition or competition in general is causing the growth sort of slowing in UR is really not how we see it. It's more the macroeconomic effects.
Got it, guys. Thanks so much.
Your next question comes from the line of C.J. Muse with Evercore. Go ahead, your line is open.
Yeah, good morning. Thank you for taking the question. I guess first question regarding gross margins and mix as we move from September to December. I think in your prepared remarks, you talked about overall System Test growing about 25% for the calendar year, which would imply storage test, I believe, roughly flat. That's your lower-margin business. Curious what's driving the sequential decline in gross margins. Historically, it's been more mobile-based. I'm curious if there's other drivers within your core SoC business that's getting to that.
Yeah, it's Sanjay here. I think Q3 gross margins had favorable product mix, specifically tied to the Semiconductor Test business. The same Semiconductor Test business has a little bit of mix change, and that's driving kind of a margin % down a tad, so 59%-58%, roughly. It's kind of in our plan and as expected.
I guess, just curious, does that mean we're seeing a pickup in mobility, or there are other drivers within SoC that have similar margins to mobility?
Overall, the mix is as we planned. We are seeing continued strength in mobility. That is a key driver. It's not anything more than that.
Okay. I guess as the follow-up, share of north of 40% in memory is pretty incredible. I guess, can you kind of walk through how you're thinking about 2020, particularly as we move to UFS3 standards and the requirements for higher-speed NAND tests, as well as the move to DDR5 and your competitive positioning there?
Yeah. Those two trends are great. We love it. That helps our product line. I would say that it's likely that the low 40% share position we will have in 2019 is a bit inflated because the DRAM portion of test buying this year is quite low. NAND's been better than DRAM, and DRAM's sort of the new space we're entering. Our share there is still relatively small. If we think about 2020, if the market this year is $600 million, we've talked about a trend line market of $700 million for memory. If we think about the trend line, the market going back, it's likely to be on the DRAM side. Our share next year could come down into the high 30s.
Given the products and the trends that you cited, if we think about another couple of years out, we should be routinely operating in that sort of mid-40% range, is our view.
Thank you.
Your next question comes from the line of Brian Chin with Stifel. Go ahead, your line is open.
Hi. Good morning. Thanks for letting us ask a few questions. Congratulations on the nice current results. I guess my first question would be on sort of the UR and industrial automation business, sorry if I missed this, what is the current expected growth rate for UR and for IA as a whole in 2019? Can you flesh out some of the underserved verticals and regions you expect to invest more heavily in as it relates to UR? Just curious kind of where you think margins on the IA business will come in at this year.
Yeah. I think from an industrial automation perspective, you should expect us to be in the teens for 2019.
Operating margin, you're talking about?
Growth.
Growth.
Growth.
Okay.
Growth. Yeah. Expected to be in the teens. That's a full year, 2019 over 2018, on an as-reported basis. Pro forma, you think it a little bit lower as we acquired MiR in April, maybe around 9%, 10% on a pro forma basis. From an investment perspective, I think, should think about different verticals and consumer electronics, and the like.
Yeah, I think geographically, as we look at United States as an example, we've had some areas of North America and the U.S. that we've been a little light in our distribution network, so we're going to fill that out. On the verticals front, we do see that electronics remains strong. It's not like we haven't invested in electronics, but we haven't leaned into it disproportionately. I think that's one area where we will have more concentration looking forward.
Okay. Yeah, that's helpful. Got it. It sounds like with the IA maybe in 4Q, some seasonality, but maybe not quite the seasonal bump maybe you've seen in recent years. Then maybe switching gears to the acquisition in terms of AutoGuide, just curious, will you go to market differently with those autonomous vehicles, perhaps more of a direct sale than some of the extensive distribution integration partners you've leveraged for UR and MiR? I would imagine if that was the case, then that might feed into sort of the timeframe you've talked about from an accretion standpoint.
Yes. That's exactly right. It's the classic customer or the big opportunities for AutoGuide tend to be larger enterprises, and the direct sales approach there will be the predominant approach. Not that we won't have channel partners, we do and will expand that, but in the early going here, that will be the majority of the story versus MiR and UR, which come at it from the other way, where today 90-plus % is through the channel, and we're just developing a few enterprise direct sales accounts.
Okay, great. Thank you.
Your next question comes from the line of Toshiya Hari with Goldman Sachs. Your line is open.
Hi, guys. Thanks very much for taking the question. Mark, I wanted to follow up on the memory test business. The improvement in market share this year to the low 40% range, how much of that is due to mix shift within the TAM, DRAM being lower and NAND being better, and how much of that is due to real share gains at the expense of your competitor? Kind of related to that, given memory test has been sort of the bread and butter for your competitor, have you seen any price retaliation from them?
I would say very little pricing activity that would be abnormal in this environment. What I said earlier about DRAM is kind of weak this year. If you normalize it out DRAM, maybe the way to think about it is our 42%-43% share would be more like 37%-38%. That's probably more our natural share position at this point. We're basically picking up business at existing memory manufacturers, so there's some real share gains happening at the traditional suppliers that you might think of for memory. There's emerging players in China, which are kind of a jump ball that we're competing with. Our share in China is above our kind of average share globally. Those two pieces are what combine to sort of put us at that natural 37%-38% point.
Got it. I assume most of the activity in China in memory so far has been on the NAND side as opposed to DRAM?
Well, there's activity in both. I don't think I would break it down too much, but both areas we see activity.
Okay. As a quick follow-up, on analog test, can you remind us how big that subsegment is within SemiTest today? I know it's a business that's been in sort of correction mode for the past few, if not several quarters. How big was that in Q3? Is it fair to say that you're pretty close to the bottom, or could there be further downside given some of your customers' commentary overnight? Thank you.
Yeah. I think when we say analog in the purest sense, linear analog is a market that traditionally sort of in the $300 million-$400 million range. This year, it's probably at the low end of that. There's another segment that's related to the discussions going on in the sort of general linear analog area related to automotive and MCU and logic. That's a market that probably oscillates between $300 million-$500 million and is also down at the low end of its range this year, probably even a little bit below $300 million. We had a couple of strong years of automotive leading up coming into this year. This year is obviously down quite a bit in automotive. If past trends sort of hold, it'll stay down for another nine months to a year, perhaps, before we start to see that recover.
Got it. Thank you.
Your next question comes from the line of John Pitzer with Credit Suisse. Your line is open. John Pitzer with Credit Suisse, go ahead. Your line is open.
Moving.
Okay, let's move on, please.
Certainly. Your next question comes from the line of Krish Sankar with Cowen and Company. Go ahead, your line is open.
Yeah. Hi. Thanks for taking my question. I had a couple of them. Mark, I think I asked this last time, too. Is there a way to quantify how much of your sales in SoC test or SoC test plus wireless test is coming from 5G? I had a follow up.
From 5G. I don't think we have that broken down well for you. No, I can't do that for you on the fly here. I'm not sure that's something we're going to be able to break out too easily, but I can't do it off the top of my head. Sorry.
Got you. No worries. I had a question on the industrial automation and your strategy. It looks like UR growth is slowing. I understand, there are external factors like macro and competition, over the last couple of years, you acquired MiR and now AutoGuide. Does it make sense to go down more downstream to a defensive software or a optics vision-based technology acquisition than a robotics hardware asset that has the potential to be commoditized down the road?
Yeah. I think there's a view that the hardware is a commodity and the software is where all the money is going to be. I would just reiterate that the differentiation we have in hardware has been incredibly strong and powerful throughout the four and a half years we've owned UR, and continues to be. On the second point around software, isn't that sort of the pot of gold at the end of the rainbow? I think there is something to that over time, that more and more value can be captured through software. Not at the expense of hardware, however. I don't believe it's that, I do think it's another market that will grow in conjunction with the hardware. It's something we're looking at. We do have a good platform to develop some capability around that.
I wouldn't disabuse hardware for the sake of software. I think they're both attractive and interesting.
Got you. Thanks, Mark.
Your next question comes from the line of Rick Eastman with Baird. Go ahead, your line is open.
Yes, good morning. Thank you. Just around the IA business and in particular, UR. You made reference to Asia Pac strength both in China and Japan. Is there a different mix of end markets there, or have you added some distribution? What would you attribute that strength to? Again, some of the verticals certainly have slowed there in industrial automation as well. Curious as to where that strength come from. Maybe just a second question, similar thoughts. We now introduced this bin picking product. We had some enterprise agreements in the backlog there. As we move forward with UR, I'm not sure if you referenced this, but would you expect UR to maybe drop a little bit further here in the fourth quarter and in the first quarter? Seasonally it does that, but where do we start to inflect towards core growth at UR?
I mean, middle of next year, or what would your best guess be there?
Maybe on that last point first. I think UR is going to grow in fourth quarter over third because typically, fourth quarter is a big bump for us. It'll be single digits compared to year-over-year fourth quarter, but it'll likely be up a little bit. We're doing this in light of the fact that if you look at the macroeconomic indicators of industrial output in the U.S. and Europe, they're significantly weak and down. If you look at the traditional suppliers of industrial automation, for example, whether it's Japanese machine tool sales or some of the specific companies that have announced recent trends, you'll see that they're down anywhere from 10% to 30% year-over-year, where we're roughly low single-digit growth.
I think we're going to have that kind of a headroom over the traditional cyclical industrial automation market where we will oscillate with a 10, 15 point gap on top of those guys. The real question is when will the industrial manufacturing sector in North America and Europe return to an expansive mode? That one, I think, is hard for us to call. We think that we can grow next year, even in light of the current economic conditions we see in industrial automation. We think it'll be a growth year next year, but is it going to be single digits or back into that 30-40 range? It's a little tough for us to call that right now. On Japan specifically, it's interesting in Japan. Japan is a country that really sees the value of automation.
They've got a demographic issue that requires it for them to stay productive. We're seeing that our product is able to win a disproportionate amount of collaborative robot business there, despite the fact, as somebody mentioned earlier, Omron's home turf is there. They've decided to rep Techman there. The Universal Robot product and cobot is winning. That's all very encouraging to us. If you look at the PMI and other indicators in Japan, they're actually not quite down as much as they are elsewhere in the world. In China, it's a similar, believe it or not, demographic concern. There's an issue of finding staff to do a lot of this manual labor that has resulted in companies in China continuing to invest, even though there's a lot of economic uncertainty and in some industries, downturn there.
We have automotive suppliers in China and elsewhere, even though the automotive industry is down, and they're letting go of people, in some cases closing factories, they're still steady adding collaborative robots at a good clip because they see the long-term trend line. All of that to me is positive. I think we're learning, we're not immune. The ROI for collaborative robots is fantastic, whether it's an economic upturn or downturn. We'd like to believe we can fight through the downturn and still get growth. We are, but it's kind of single digit, and I expect we're always going to ride above the curve here throughout these cycles.
Okay. Thank you.
Your next question comes from the line of Sidney Ho with Deutsche Bank. Your line is open.
Thanks for taking my question. I wanted to clarify one thing. You said the 5G infrastructure added about $200 million to the TAM this year. Next year, there will be a pause on the infrastructure side, but maybe some pickup on the handset side. The real next big opportunity is until the mmWave, which is a couple of years away. I know it's early, but are you implying that the 5G-related incremental revenue for next year will kind of be flat? I guess that's net of any decline in 4G.
Yeah, it's Sanjay here. I think that, we believe it's roughly $200 million. I believe we said in the past that we believe all in all, 5G will have about a $300 million-$400 million TAM uplift to the market. Right now we're trying to calibrate, as Mark said earlier, utilization of testers is high, but we're trying to calibrate the build-out that is getting ready for the infrastructure rollout of sub-6 GHz in 2020, and have the customers created the capacity to handle the throughput required in 2020. We're kind of trying to triangulate that, and we'll provide an update of our best estimate in Q1. I think from a 5G handset perspective, I think you should think about sub-6 GHz is going to increase in 2020, and then also mmWave, first infrastructure, and then handsets in the following years.
There will be this wave of adoption first started with 5G infrastructure, then 5G handsets for sub-6, and then going to millimeter wave infrastructure more in 2021, and then handsets 2021, 2022, is the way we see it now.
Okay, great. As a follow-up, if you look back at the 4G opportunities, which obviously started early this decade, it may be still growing right now. If you look at the shape of that TAM, how it grew over time for you guys, as well as how long it took for the opportunity to plateau, and eventually start falling off. How do you think the shape of the 5G over the cycle, how is that going to be different than 4G?
I think I want to break that into two different test markets. There's the LitePoint business we have that saw a tremendous spike when 4G rolled out, of handsets shifting over to 4G, which happened quite rapidly. LitePoint is fueled by those standards shifts. When you think of LitePoint's business, that market grew to like $1 billion in size. LitePoint's business grew upwards of $a few hundred million. It was a huge boom and then a bust. 5G for LitePoint, I think, is going to go slower because under the 5G umbrella, there's more sub-standards and it's a slower rollout. In SemiTest, I think it's a different issue. SemiTest has been benefiting less from sort of the standards change to 4G than the complexity growth associated with the standard, its evolution, and all of the sort of other features in phones.
We've had a decade almost now of strong mobility SemiTest demand. Some of that you can attribute to the 4G standard, much more of it you attribute to the complexity growth in the phone and other perhaps correlated, but other areas like the apps processor, the AI getting built in, the image sensors, et cetera. When we look at the 5G world for SemiTest, there's a difference. One subtle difference, but I think the net result's going to be the same. The subtle difference is the 5G millimeter wave step-in, is a much bigger technological impact than the 4G LTE step for SemiTest. The silicon to go at millimeter wave frequencies is going to be more test intensive and require higher test intensity for SemiTest. That's a balloon. And it'll be metered out over years. The complexity growth is a constant. That one, I think, continues.
You've got a little bit more of a bump on top of it with millimeter wave being more complex. Looking forward, marginally more test intensive than looking backwards is how I think about it.
Great. Last question from me. A couple of years ago, your SoC Test share was in the mid-50s. It dropped off last year and may be coming back a little bit, this year. I guess next year it probably should be up as well because of better end market customer mix and whatnot. What would it take for you guys to get your share back to the 50% level?
I think overall, we've got to get a kind of rollout of mmWave. That's a key part. That drives a lot of the RF test intensity that we're solid and good at. The memory story has to play out as it's playing out, are sort of two key elements of that. Less important, but also in there is picking up a little more what we would look at as edge AI type test business. We have to get more devices in automotive and in consumer products that are digitally intensive and executing these inference algorithms that have been developed in the cloud. We're strong there. That growth is sort of the third leg of the stool that gets us there.
Great. Thank you very much.
Your next question comes from the line of David Duley with Steelhead Your line is open.
Yeah, thanks for taking my question. A couple things on 5G handsets. Could you take a stab at how much more test time is involved in a 5G handset versus a 4G handset? I realize there's more RF and power management content and more complexity in some of the parts, but if you could take a stab at estimating how much more test time intensity there is with a handset with 5G versus 4G, that'd be great. The second question is, have you seen an acceleration in customers' plans to roll out 5G phones next year?
On the second point, yes. There's an acceleration of 5G rollout in general that's happened this year. If you went back to the beginning of the year and said, "What's the pace of 5G rollout to today?" It's quite accelerated in both base stations and in handsets. However, the difference, a significant point in all that is that most of the acceleration is in sub-6 GHz. It's the base stations and the handsets that are really starting to roll fast are in Asia, where sub-6 GHz is rolling out ahead of the rest of the world. Next year, we'll probably get a little bit of broadening of that. That's what's expected, and I would expect that, too. The infrastructure in the United States and Europe really isn't going to be in place to take advantage of a lot of that.
That's going to be a bit of an anchor. Thank God Asia's running, because it's been great for us. The test time question, there's two different ways. I'm not sure which question you're asking. There's the testing of handsets that LitePoint does in a 5G phone, and then there's the testing of all the silicon that rolls into the 5G phone. Which one are you?
I was referring to all of the silicon in the phone.
Okay. On the silicon side, that one's a tough one to call right now because people are still developing early test cases for the 5G silicon. The 5G-specific silicon tends to be, there's transceivers, there's antennas, and there's a modem. All of those are in early pilot phase production. All of them have, today, much higher test times, which, if you extrapolated, would come up with stupid numbers for test markets. We expect they'll get optimized, as always, over the next year as they go to production. I think in the end, we're going to see small, maybe handful, 10%-plus or minus kind of test time increases when people finally sort through all this. That's sort of a rough number we use.
I think I'd just add one quick point on the 5G acceleration. I think on the mmWave, we're seeing a lot of engineering efforts and testers to support that endeavor. Where the handsets, as I said earlier, would come out, we believe would ramp in a year-plus. We are seeing an uptick in the engineering efforts on 5G mmWave.
Okay, great. Thank you.
Okay. operator, we have time to sneak in just one final question, please.
Certainly. Your next question comes from the line of Tom Diffely with D.A. Davidson. Your line is open.
Yes, good morning. I was hoping you could add a little bit more on the automotive market. I know it's down quite a bit this year. What is your long-term view there? Do you view that more as a capacity play, or is it the technology drivers as well?
Yeah, it's primarily a technology driver. The automotive unit volume in the world's not going to move much over the next decade, probably, or at least for the next five years. The complexity growth of the silicon in automotive is what's important. Much of that relates to more autonomous driving, driver assistance technologies flowing into the cars. From the design point of view, a lot of tooling was put in place for that last year. This year, I think because unit volumes are down and such, the manufacturers are kind of holding off. I expect, like I said before, that'll return to growth late next year, and that trend line for automotive test should be above the trend line of the rest of the industry in terms of growth.
We look at that, it's going to grow probably greater than 10% on average, whereas the rest of the test market we've modeled in that sort of 4%-5% range.
Great. Very helpful. I appreciate it.
Great. Thanks, everyone, and we look forward to talking with you in the days and weeks ahead. Bye-bye.
Ladies and gentlemen, this concludes today's conference call. You may now disconnect.