Good morning. My name is Kelly. I will be your conference operator today. At this time, I would like to welcome everyone to the Teradyne third quarter 2015 earnings conference call. All lines have been placed on mute to prevent any background noise. After the speakers' 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 1 on your telephone keypad. If you would like to withdraw your question, press the pound key. Thank you. I would now like to turn the conference over to Mr. Andy Blanchard. Please go ahead, sir.
Thank you, Kelly. Good morning, everyone. Welcome to our discussion of Teradyne's most recent financial results. I'm joined this morning by our CEO, Mark Jagiela, and our Chief Financial Officer, Greg Beecher. Following our opening remarks, we'll provide details of our performance for the third quarter of 2015 and our outlook for the fourth quarter and provide some general market comments. 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. Those slides can be downloaded now or you can follow along live. In addition, replays of this call will be available via the same page about 24 hours after the call ends.
The matters that we discuss today will include forward-looking statements that involve risk factors that could cause Teradyne results to differ materially from management's current expectations. We encourage you to review the safe harbor statement contained in that earnings release, as well as our most recent SEC filings. Additionally, 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 those non-GAAP financial measures, including reconciliation to the most directly comparable GAAP financial measure where available on the investor page of our website. Also, between now and our next earnings call, Teradyne will be participating in investor conferences hosted by UBS and Credit Suisse. 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. Greg will offer more details on our quarterly financial results, along with our guidance for the fourth quarter. We'll answer your questions. This call is scheduled for one hour. Mark?
Thanks, Andy, and good morning, everyone. Today, I'll provide a brief summary of the highlights of our 2015 results to date. Then focus on how we're looking at 2016 and beyond for long-term growth at Teradyne. Greg will provide the details for the third quarter, the full year, and our Q4 guidance. In Teradyne's core semi test business, the 2015 market is playing out about as we expected at the beginning of the year as the annual SoC test market size is trending to about $2.1 billion, right in the $1.9 billion-$2.2 billion range we forecasted. The good news is that this is up from the previous down cycle odd year of 2013, where the market size was $1.9 billion, validating that increasing device complexity and the reduction in the impact of parallel test is starting to have a positive effect on the market.
We are pleased with our performance in this market as we've moved our market share from about 40%-50% over the past three years. A combination of tailoring our products for high-growth sub-markets and cultivating higher growth customers has been and will continue to be our playbook for success. The memory test market is also performing as expected at about a $500 million annual run rate. In memory, the technology story around test remains the increasing bus interface speeds that create demand for new testers. Our Magnum V continues to capture share in testing these higher speed, non-volatile devices. This trend will continue into 2016 as even higher speed UFS and PCIE versions of NAND will start to grow in the market.
For Teradyne, the good news is that our same Magnum V platform will be testing these higher speed devices as the architecture was set up from the beginning to track these trends. Our storage test business has seen a nice rebound in 2015, with sales more than doubling over 2014, resulting in a solid return to profitability. Buying for enterprise hard disk drive tests combined with new applications testing solid-state drives has led to the doubling. Also, as packaging of semiconductors moves to increasingly dense and complex system modules, we are finding applications in system-level test where long application-specific test suites can be run efficiently in our highly parallel, environmentally controlled tester. In industrial automation, Universal Robots had record sales of $16 million. Our recently introduced UR3 robot arm saw strong demand in North America and China in particular.
Early adopters of the new tabletop robot in the U.S. market are the automotive, electronics assembly, and high-precision metal fabrication industries. Common applications are screw driving, glue dispensing, assembly, and product testing. In China, we are experiencing demand from medium to large companies within the electronics industry, where the UR3 is typically used for pick-and-place tasks, machine loading, and light assembly. A good example that highlights the benefits of UR's collaborative robots in electronics assembly operations is Wistron. This Taiwanese manufacturer has quickly deployed over 60 UR robots to improve productivity in its laptop assembly operations. A video outlining Wistron's experience can be found in the case story section on the ur.com website. Turning to the current environment. Over the past few months, we've had many investor questions related to our overall business conditions, given the worries about slowing emerging market growth.
Far, we've seen little impact to our business. While we are in no way immune to swings in macroeconomic growth, the longer-term trends in semiconductors and automation remain solidly positive. Our third quarter unfolded much like those in the past. Semi test orders declined 47% sequentially, about in the middle of the five-year historical range of Q3 bookings decline. That normal seasonal order slowdown is reflected in our Q4 guidance. While it is too early to put a number on the SoC market size next year, we do expect historical SoC test buying patterns to continue and 2016 to be an up year for the industry. The drivers of test demand, unit growth, and increasing complexity remain intact. While smartphone unit growth is slowing, forecasters still expect over 140 million incremental units in 2016, similar to the 2015 increment.
Those units will have faster, more complex processors, more communication links, sharper displays, more storage, and every device will be optimized to extend battery life, all made possible by the advanced semiconductor devices that we test. As customers strive for higher performance in smaller sizes with lower power consumption, advanced packaging technologies are becoming another area of differentiation for our customers. From a test perspective, we are well-aligned both technically and commercially with this trend. Advanced packaging puts a premium on known good die testing, as repair inside these packages is prohibitively expensive. More extensive testing at probe requires the precise accuracy and signal delivery technology built into our UltraFLEX. Longer term, we expect to see growth in semiconductor applications beyond just advanced smartphones. Continued unit and complexity growth in automotive, industrial, and consumer products should keep the market on a solid footing for the foreseeable future.
Two trends worth mentioning are 60 GHz applications and the continued proliferation of image sensors. A variety of applications are opening up around the 60 GHz and above frequency space. 802.11ad, automotive radar, and cellular picocells are all developing new silicon technology to exploit this space. This, in turn, drives a new wave of test equipment for semiconductor test and LitePoint. At LitePoint, we've already fielded our first 60 GHz tester, and early adopters are readying devices for commercial applications in late 2016. On the semi test side, not only do these RF devices need a much higher carrier frequency and information bandwidth, the downstream chips that modulate, demodulate, and process the information digitally also need much higher complexity, speed, and processing power to manage the higher bandwidth. This, in turn, requires faster memories and bus speeds to keep up with the movement of information.
All of this reinforces the complexity trend that drives the need for more test capability. Automotive radar is another emerging application in the 75 to 80 GHz spectrum, where enhanced driver safety and assistance packages will increasingly incorporate multiple transmitters and receivers per vehicle to monitor the surrounding environment. Our UltraFLEX tester is positioned directly in line with these trends. In the case of image sensors, the diversity and unit volume forecast is exciting. At Teradyne, we have about a 70% share of the world's image sensor test business on our J750 platform. The trend to higher resolution sensors means more test seconds. Improved sensor performance in low light conditions means the low-noise environment of the J750 is increasingly valued. Additionally, infrared, time of flight, and other new sensor technologies will show up in phones, automobiles, and security applications, which will continue to fuel unit growth and drive tester demand.
At LitePoint, our core connectivity and cellular test markets remain compressed, resulting in 2015 revenue coming in about flat with 2014. Looking forward, we see new markets like 802.11ad, protocol testing, and a few yet-to-be-announced new products opening up about a $200 million new set of markets over the next two years. Four months into our acquisition of Universal Robots, things look very promising. A strong $16 million quarter, a strong new product launch of the UR3, and numerous research reports validating the rapid development of this market are all positive. We expect to continue strong growth at 50% or more per year as the range of applications for UR's collaborative robots continues to expand and the attach rate for those applications increases. We are committed to maintaining our leadership in this rapidly developing cobot market and will continue to invest in enabling the growth of Universal Robots.
A capital allocation plan, including a balanced mix of dividends, buybacks, and M&A, will continue to be our strategy. The past two years are evidence of this commitment with the initiation of our first dividend, our largest buyback in over 10 years, and the promising acquisition of Universal Robots. In our January call, we plan to update you on our capital return plan for 2016. Here's Greg to review the financial details. Greg?
Thanks, Mark, and good morning, everyone. I'll start with the key accomplishments of 2015 as we near the finish line. I'll outline how we're positioning ourselves for earnings growth in 2016 and thereafter and wrap up with the Q3 details and the fourth quarter outlook. This year, we're tracking to just over $1.6 billion in revenue with a 20% operating profit rate and will generate about $300 million of free cash flow. As Mark noted, we're pleased with this performance considering the annual pattern of smaller steps in mobility complexity in odd years continued. This sets up the expected complexity jump in mobility next year. Looking at 2015, the year stands out for three key reasons. First, we're on track to have back-to-back years at just over $1.6 billion in sales, versus the typical odd year pattern seen in 2011 and 2013, where sales declined about $200 million.
Second, we have expanded into the fast-growing cobot space with the addition of the market leader, Universal Robots. Third, we're on track to return $350 million in capital to shareholders. I'll quickly comment on each of these three important milestones. First, on the strong back-to-back years. We described last quarter that the UltraFLEX leased tester buyout added just over $100 million of revenue in 2015. We've also seen continued strength in image sensor tests, up about threefold from last year, with the high growth in pixel count, multiple cameras per phone, and the increasing use of connected cameras for safety and security driving demand. Analog has also seen strength led by automotive applications, where semi content is growing for safety and entertainment.
Finally, we've commented on the start of the shift in the buy rate line with the slowing of parallel test efficiencies and the increase in device complexities. These factors have held up SemiTest in an odd year, and when combined with the strength in storage tests and Universal Robots, have kept us about $200 million above the recent downmarket years of 2011 and 2013. If you look more strategically at 2015 or the two-year 2014-2015 period, it actually reflects a very careful and targeted market segment and account focus. This is how, for many years, we've been able to invest far less OpEx than our largest SemiTest competitor, while at the same time gaining share and achieving top VLSIresearch customer satisfaction ratings.
In short, we're well-aligned to the higher growth segments and accounts in mobility, analog, microcontrollers, and image sensor with our fast time to market software, high accuracy specs, and high test system throughput. We have equally avoided some closely adjacent SemiTest segments and offerings which have declining or razor-thin profit pools. This includes microprocessor testing, commodity DRAM testing, and a slew of test cell offerings such as probe cards, handlers, and simple DIBs. Moving to the second important 2015 milestone, the acquisition of Universal Robots. UR has expanded the definition of a cobot beyond just a safe, force-limited robot, that is, one that doesn't need protective caging as it will stop when it feels any pressure in a handful of milliseconds, to also being a robot that is smart and easy to train and redeploy.
UR's unique programming model shatters the past model and enables UR cobots to be deployed by the very smallest of manufacturers. It is the ultimate equalizer for small and medium-sized enterprises, as you don't need skilled engineering resources on staff. This simple-to-train capability is one of the key reasons UR is the clear leader with SMEs. These customers are able to get a very fast payback, often about six months, and they also know they can easily redeploy their UR cobot when their demand invariably changes. Universal Robots is on track to achieve about $60 million in sales in calendar 2015, and we expect 50%-type growth next year. At this early stage, cobots have barely penetrated the available market, and some third parties have pegged the TAM at over $3 billion by 2020.
UR's product lead has attracted many of the top distributors and integrators, there's a growing army of third parties developing solutions in many different verticals on Universal's platform, which should further extend their formidable market lead. It won't solely be about having the best product. It will also be about having a developed and robust ecosystem, which will be increasingly hard for followers to replicate on their platform. Shifting to the third important milestone. We're on track to return $350 million in capital this year. This sizable level of capital return signals the strong confidence we have in our future earnings power. We also stepped up the pace of buybacks in the third quarter given the financial market weakness.
With our steady performance of solid free cash flow averaging about $250 million a year since 2011, we've had two fundamental strategic choices at the extreme, to immediately return all of our profits to shareholders or to use these profits to grow the top line through M&A. We've elected neither extreme. Rather, we use a very balanced approach, where at any one time period, we may favor highly selective M&A or higher levels of capital return, simply depending on what is the best financial return at the time, considering both our stock price and the small number of attractive opportunities in our M&A funnel. Now shifting to 2016 and beyond, as Mark described, we see favorable trends. These include rising complexity, the slowing of parallel tests, and the need for tighter accuracy and higher quality testing.
I'll take a quick moment to comment on the tighter performance specs required for some of the new packaging used for mobile products. In advanced packaging for mobility markets, high-performance devices are being placed much closer than in the past, and they have very stringent power requirements, which require more accurate tester instruments, as found in the UltraFLEX. You may recall the UltraFLEX hit the market with the most efficient parallel test, and then it became recognized for superior programming and debug tools critical for the vertical mobility ramps, and now it's earning its stripes yet again with its performance for testing advanced packages. We're pipe planning some inventory in the fourth quarter for 2016 demand to ensure we maintain attractive lead times. We did the same back in 2013, which positioned us well for the 2014 ramp.
Shifting to the third quarter results, we closed the quarter with total cash and marketable securities of $1.077 billion, up $48 million from the prior quarter after returning $111 million through buybacks and dividends. Third quarter free cash flow of $159 million was due to strong profits and the sale of previously leased systems. On the buyback front, we've repurchased 13.4 million shares totaling $254 million at an average price of $19.02 through yesterday. We stepped that buying up in the third quarter to take advantage of market volatility. Against our $500 million authorization, we have a remaining balance of $246 million.
As a quick reminder, we plan on buying back $300 million in 2015, which when combined with our quarterly dividend totaling about $50 million for the year, will lower our U.S. cash and marketable securities to a level much closer to a minimum U.S. cash operating balance of $400 million. As we've done in the past, we'll update you in January on our 2016 capital return plan. Moving to the details of the third quarter. Semi test bookings were $211 million, driven principally by the seasonal patterns, and revenue was $326 million. Recall that our third quarter company bookings typically fall 30% to over 60% from the second quarter levels. This drop off is due to the timing of annual mobility launches and the overall high consumer consumption connected with back to school and end of the year holiday buying.
SoC test orders were $191 million, memory test orders were $20 million. Semi test service orders were $34 million of the total. Shifting to wireless tests, we booked $40 million and recorded $55 million sales for the third quarter. LitePoint is on track to be about flat with last year, with solid margins above our 15% industry target, and has funded a series of new products that should benefit us in the years ahead. The wild card for LitePoint in the foreseeable future remains our large customers' buying patterns, which have declined considerably from 2012 and 2013. It's too early to get a picture of 2016, but we expect further share gains with our production-optimized solutions. Moving to system test. Orders were $47 million in the quarter, and shipments were $69 million.
A substantial portion of those orders are annual service contracts and products with long lead times, which will convert to revenue over the next year. In storage test, the three-and-a-half-inch cloud demand should bring our annual sales to about $85 million, versus only about $40 million last year. As drive capacity continues to grow, so does test time, we're well-positioned to capitalize on this demand with our new Saturn tester, which has up to 13,000 asynchronous slots. In industrial automation, we have bookings and sales of $16 million with strong demand across the product line. At the company level, our sales were $466 million. Non-GAAP gross margin was 56%, down from last quarter due to mix shift, including increased storage test shipments. The non-GAAP operating profit rate was 23%, and non-GAAP EPS was $0.40. We had one 10% customer in the quarter.
You'll see our non-GAAP operating expenses of $152 million were down about $1 million compared to the second quarter due to lower variable compensation expense. Shifting to our outlook for the fourth quarter, sales are expected to be between $295 million and $320 million, the non-GAAP EPS range is $0.07-$0.12 on a 208 million diluted shares. Q4 guidance excludes the amortization of acquired intangibles and the related tax impact. The operating profit rate at the midpoint of our fourth quarter guidance is about 8%. Our 2015 tax rate remains at 23%. If the R&D tax credit is reinstated for 2015, that rate will drop to 21%. In summary, 2015 will go down as our first back-to-back year with sales just over $1.6 billion versus the typical $200 million fall off.
We've added Universal Robots, the clear cobot market leader, to the Teradyne fold, we're returning significant capital, a strong signal of our confidence in our future earnings power ahead. 2016 is an even year where mobility complexity is our friend once again as it, along with device unit growth, drives more incremental tester demand. With that, I'll turn the call back to Andy.
Thanks, Greg. Kelly would now like to take some questions. As a reminder, please limit yourself to one question and a follow-up.
As a reminder, if you would like to ask a question, simply press star, then the number one on your telephone keypad. We'll pause for just a moment to compile the Q&A roster. Your first question will come from the line of Atif Malik with Citigroup.
Hi. Thanks for taking my question. Congratulations on a good quarter and decent guide. Question on integrated fan-out packaging, a new type of packaging for mobile devices next year. Can you talk about the impact on the test equipment demand next year? Maybe you can compare the test times versus package-on-package type testing.
Sure. I think there's general trends and there's specific applications. Overall, these packages require, as I mentioned in my remarks, much longer test times at probe in order to guarantee high yields coming out of a probe situation before they're mounted on that substrate. What we expect to see is that probe test times will increase. For the integrated package itself, that will vary depending on what's integrated into the package. In general, if you look across probe and final insertion, the number of test seconds applied to the total part when it ships to, let's say, a phone manufacturer, will be proportionally up as compared to the prior generation. It may be up 10% or so. That's how we've modeled it at the moment. As it plays out next year, we'll get a better read on it.
Okay. The second question on system test sales. I mean, system test sales were up 50% in the reported quarter. How should we think about the long-term revenue run rate? If you can also comment on Western Digital SanDisk merger. I assume SanDisk is a big customer of yours on SSD side and Western Digital on the HDD side. What impact the merger could have on your sales?
Overall system test was strong this year because of storage test. As we look forward, storage test is positioned reasonably well now with enterprise drive and SSD application. The wild card for us is what I described as the system-level test applications, which are less storage-related and more long cycle time burn-in applications. On a baseline, I'd say we'd look at next year in that segment, roughly flat with this year, with the upside, it's the system-level test could provide a kicker. It's really too early to tell if the combination between Western Digital and SanDisk will matter in any significant way to us. That's how we're looking at the overall business.
Thanks.
Your next question will come from the line of Mehdi Hosseini with Susquehanna Financial Group.
Yes. Thank you. Thanks for taking my question. I'm looking at the market, if I just were to go and look at your market share, it seems like turning the lease into an outright purchase has actually been instrumental in helping you to maintain the market share, especially, when I look at the market share trend over the past four years. As the lease program ends or turning the lease into an outright purchase end, how should we think about your ability to maintain and exceed market share? Market share, I'm referring to memory and SoC combined. I have a follow-up.
Mehdi, this is Greg. If you look at market share over a multi-year period, you can see that we've been steadily gaining quite considerably. Your point that the $100 million or so this year, really probably belonged last year. Our last year's SoC market share is actually higher last year, a little bit lower this year. If you put this all on a trend line, you'd see the trend line's very clear that we consistently are gaining market share, and it's from a couple of reasons. One is we're aligned to faster-growing segments, the healthier segments that have growth. We've optimized our product solution set around the requirements that those growing segments have. That's what's gotten us this steady market share trajectory growth.
Can you increase the share from here, assuming that there won't be any impact from the lease program?
Yes.
Should we assume that the share remains unchanged?
No, I think for a long time now, in fact, almost seven or eight years, we've had a model of about a point or so of market share gain per year in SoC. It never happens linearly, and the market noise amplifies and shrinks that in any given year. We still see the ability to move about a point of real market share a year, up to at least into that 60% range.
Got it. Moving on to the next topic, InFO. There's a lot of noise and commentary out there. So far, and I think InFO or the advanced packaging has been discussed for the past 15 years, and I still see a limited application. Do we have any signs out there that the key fruit company has decided to adopt this technology? Because if they don't adopt it, I see its application limited, and especially with tests, because you're going to be involved at a wafer level. In that context, the question is, are you seeing any signs that the system company is actually going to step up and adopt this?
If not, how do you see the existing infrastructure that has been added over the past year is going to be utilized over the next one or two years until the system company adopts this technology?
Two things. I can't comment at all on what any particular customer's plans might be. In general, I'll say two things. First, Teradyne, as it relates to test, is reasonably indifferent in terms of the impact to our business, whether the technology's adopted or not. It shifts more test seconds to probe, as I described earlier, but net-net, it's not a big swing factor for us. We think our product, the UltraFLEX.
Has some advantages in advanced packaging test, but not a big swing. Second, do I think even though it's been around for 10 years, are we at an inflection point or an adoption point, in general in the industry? I think the answer to that is probably yes. The economics are improving quite dramatically, it seems, and the miniaturization of benefits are also compelling. I would expect if we look over two to three years, that you'll see a much faster adoption of advanced packaging than we've seen in the past three or four.
Got it. Thank you.
Your next question will come from the line of Farhan Ahmad with Credit Suisse.
Thanks for taking my question. Just a quick question on the odd year, even year cyclicality. For last 10 years, obviously, there's been a trend where your revenues grow every even year and odd year they decline. However, one thing that's been also going on for last 10 and even much longer has been that we have been on a two-year cadence on Moore's Law, which seems to be taking a break this year, and we are basically on 20 nanometers/16 nanometer node, instead of two years to three years, and 10 nanometer most likely is not going to be introduced next year into most products.
I wanted to ask you if you think that could have an impact in any way for next year, and if the Moore's Law, in any way, had anything to do with the cyclicality of odd year, even year that you saw in your business.
I think it's a reasonable point that usually when a new process node is introduced, there is a little bit extra buying of capacity for yield issues surrounding that new node. Next year, the absence of any dramatic step in lithography might dampen somewhat the peaking. What happens when it peaks in the past is the following year, it gets amplified in terms of the absorption of that extra capacity as yields improve. If it does dampen slightly the peak next year, it will serve to raise the following year's buying. Long-term trend, the node issue doesn't inflate the long-term market. It just amplifies the peaks and valleys.
Got it. Thank you. One question on your storage business. Could you just give some clarity into how much of your business is coming from hard drives versus SSDs? Also, in terms of the SSD test market, how's your market share in that particular market, and how big is the overall market? We are seeing pretty healthy growth in SSD, and I wanted to understand if that could be a big driver for you, going forward as you introduce new products, and can increase market share in that area.
Right. Okay, I'll take that one. Last year, a good part of our or most of our demand was SSD. This year, it's more magnetic or hard drive demand with some SSD demand. First, on the hard drive side, as you probably know, we're the only freestanding merchant supplier, we have a very strong position with the three buyers that are out there of our tooling. The only other competitor is captive now in Seagate. In terms of SSD, we have multiple applications at a strategic account, and we're speaking to other customers to see if their test times and their volumes make sense to use our solution. It's early in that process.
We do believe that market is going to grow for us, it needs some time to play out, as, again, units and volumes grow and test times get longer, I think that moves them towards our solution.
Thank you. That's all I have.
Your next question will come from the line of Jagadish Iyer with Redstone.
Hi. Thanks for taking my question. Two questions, Greg, Mark. First, how should we think about LitePoint going forward in 2016 in the backdrop of your comments saying that slowing smartphone growth next year? What are some things that you can do to buck the trend for LitePoint? I have a follow-up.
Okay. I'll take the first part of that. LitePoint, I'll start with, is very healthy on the bottom line, profit rate-wise, and has been able to gain market share year after year as they have a very production-optimized solution. The competitors tend to come at it from a different DNA, whether it's an R&D DNA or other type solution, and it's not optimized for the problem at hand. I expect LitePoint will continue to do quite well in production testing, albeit that market's been compressed. What has LitePoint done to try to improve their lot in life? A year ago, they got designed into a bunch of the Asian up-and-coming cellular players. Now, unfortunately, a number of them have not done terribly well this year, so we have initial design-ins, but there hasn't been much activity for anybody over there.
We'll continue to monitor that situation and hopefully improve our position when buying does eventually return. We've also have introduced some new products, some we'll talk to you more about in the future, but these new products will take some time to latch. They will open up new markets, and we expect the old LitePoint high differentiation, production-optimized DNA will show up in these new products that you'll hear more about as we get into next year. I do think just to color that a little bit more, the $200 million of new TAM we're talking about is really how we expect LitePoint to return to growth over the next couple of years. The traditional cellular and connectivity space, we expect the market size there will maintain its level. It's not expected to grow, and it's expected to remain very competitive.
The emergence of this AD standard at 60 gigahertz is really different than the typical progression in Wi-Fi and connectivity of old, where you went from G to N to AC. This is a very different application and a different test insertion in mobile devices and other devices that'll emerge. That's one. Protocol testing is another area where LitePoint has not participated in, and we're now moving products into that set of applications. It's really TAM expansion with new products is the strategy at LitePoint.
Okay. Fair enough. A big picture question. Your shares have been somewhat range-bound for a while, and given your cash position and what choices you might have in terms of potentially raising additional capital, what is management strategy in terms of crossing over the hurdle of your stock price? Would love to hear your thoughts, including your views on accelerated buyback.
This year, as you know, we're returning significant capital, $350 million, about that amount, and we're lowering our available U.S. cash to about $400 million, a little bit above that. That's what we want to maintain as a buffer for some possible severe downturn, which we don't see, but every 10 years, you can have an event. In terms of how we get the shares out of this range bound, we have plans to grow each of our businesses quite considerably that could get us in the midterm to a $2 EPS stock price. The organic plan is job one, and that includes Universal Robots is going to be a big contributor.
Semi Test is going to continue what they've been doing, that staying in that trend line of share gain with some of the improvements in the market based upon parallel test slowing down and intensity increasing, I think will give some help on the market size. Then we'll also return, it'll be over time, we'll return a bit more capital, and that could take some shares out. We have a plan that we're looking to formalize in the fourth quarter that will tighten up how we think we can get to $2 four years out, and that very well could include another acquisition. It doesn't have to, but that certainly could be a part of the plan as well.
Excellent. Thanks so much.
Your next question will come from the line of Stephen Chin with UBS.
Hi. Thanks for taking the question. You mentioned in the press release that you would be making some collective inventory investments in Q4. Just wanted to get some color on what that relates to, what the magnitude of the inventory build might be, and then what gives you confidence that 2016 will see this big pickup in demand?
The rough numbers are at the end of the fourth quarter, if you looked at our balance sheet carefully, you'd probably see $50 million more of inventory. Some of it will be in inventory, some will be in a prepaid account, depending on how the accounting works with our contract manufacturer. About $50 million of more inventory in the fourth quarter. We have high confidence with this demand based upon extensive, very close discussions all along the way. We've been in a similar situation a couple of years ago, too, in the last odd year going into an even year. We've been down this path before, and we feel good about it, and this is also inventory that can be used across a broad set of customers as well.
This whole even-odd year has played out for many years because there's a bigger step in some of these new phones that come out in the even year, and that's much more test time, greater complexity. We see that continuing, and that drives ultimate demand.
Got it. We've heard some commentary that Q4 might be seasonally weak for some of the OSAT customers. Just curious to hear whether you're seeing that weakness and that's baked into your Q4 guidance and sort of whether you're expecting a pickup after Q4 into early 2016.
We think we factored everything in that's knowable into the fourth quarter guidance. Some of the segments that remain strong are image sensor and mobility. They're holding up. We think next year should start off healthy. We're feeling good about the first quarter, even though it's some time away. Yeah, I think that if you look at. There's a broad spectrum of utilization depending on the segment. Some segments are incredibly tight and capacity is being added and pulled on quite rapidly here in the fourth quarter, and others are slack, as happens throughout the season. All in all, to us, it doesn't seem to be that different from prior Q4s that we've seen.
Great. Thank you.
Your next question will come from the line of CJ Muse with Evercore.
Yeah, good morning. Thank you for taking my question. I guess first quick question is the $2.1 billion for SoC, does that include the $100 million in leases?
Yes, it does.
Okay, great. Given your confidence for next year, I think one of the concerns that the market has is that the 810 chip is not shrinking and therefore there will be less demand there. At the same time, though, you've talked about lower benefit of parallelism, you talked about a 10% uptick in terms of packaging and test times there. Would love if you could maybe dig a little bit deeper holistically for the whole year. What's driving the confidence. What's mitigating the lack of a shrink in the mobility side? Also, are you expecting within that image sensors, baseband, RF, PMIC, et cetera, growing as well? Or are there certain pockets that's giving you the confidence that overall mobility and overall SoC can grow year-on-year?
Well, again, without being specific around any particular part or customer, I think that what we see at this point in the year are early samples from our customers where test program development's occurring for parts that should ramp next year. We get a little bit of visibility into the test times and the intensity of those parts that becomes much clearer as we get into January. As we look at it, and we look at those early signs, we see that the trends we've talked about are continuing, and therefore we would expect to get an even year-type impact from a broad set of devices. Not just any one particular part, but even on the cellular side with RF transceivers and such, you see the proliferation of more frequency bands in phones. You see the adoption of more antennas in MIMO.
All of those things contribute to incremental test intensity.
That's very helpful. If I could just follow up, I guess, on the non-mobility parts of SoC. I believe this year, linear was a great year, the catalog business, auto down. Curious, standing here today, what kind of visibility you have there into next year?
Yeah, that's a harder one to read. Linear started out very strong. It's cooled off a bit here. It's very episodic, is the history there, and it doesn't follow a good pattern. That piece of the markets, I think, standing here today, is harder to prognosticate for next year. On automotive, I think what automotive is going to show, that it's historically shown, is a very steady, slow growth and the safety-related enhancements to cars that will roll out over the next three to five years, whether it's autonomous driving or driver assistance, will really, I think, impact the test side of life a couple of years down the road. When we're in 2017 and 2018, I think we'll really see the automotive market take off for test. Right now, the devices are somewhat in early development.
I'd expect automotive next year to be about the same, and a kicker coming in 2017 and 2018.
Great. Very helpful. Thank you.
Your next question will come from the line of Timothy Arcuri with Cowen and Company.
Thanks. I had two questions. I guess the first question is on the fourth quarter. Usually, what your revenue in Q4 is somewhere between what you book in Q3 and what you book in Q4. For your revenue midpoint of 308, roughly, it would imply that bookings have to be down, and that seems a little odd given that systems should be up because it's typically up seasonally in the fourth quarter. That would say that SemiTest orders ought to be down. Yes, things are bad for sure. We know that. SemiTest orders in the fourth quarter are almost never down. I'm just wondering if you can bridge the gap between your revenue guidance and maybe what you expect from a bookings perspective.
Tim, the fourth quarter guidance is, at this moment, our best view of what will happen. As you know, we tend to set the guidance somewhat cautiously given our history of at least meeting or beating. That's what we see now, and there are some customers that are going to likely get testers in early Q1, not late Q4, given lead times that we have now. Some of the orders seem to be coming in late this quarter, but we expect it to be a good quarter. It's just how the slot plan all shakes out, and maybe there's a tad of conservatism added to it.
Okay. I guess, Greg, just on that point, so that would argue, it would seem to say that Q1 maybe sets up a little better than a typical Q1. Q1's typically up, but it would seem like it might be up a little more this year than it typically is?
Could be, Tim.
Okay, thanks. then, just last big-picture question. Have you guys gone through and of the $1.1 billion that you're getting from SoC, have you tried to parse this by end market and focus in on how much of that's coming from the industrial and the automotive markets?
We have, although we don't spend a lot of time trying to do that because some of it is hard to bucket if it's a wire, if it's a LAN device, is it going into a PC application or some other application? But in general, as we look at our business, roughly half of our business, and slightly more than half, is related to mobility in some way, shape, or form in SemiTest, in the SoC side. figure somewhere around 55%, 60% of our revenue comes from mobility-related devices. But that includes image sensors are in there, power management's in there, baseband, apps processors, there's some linear in there, touch screen controllers, microcontrollers. There's a lot of things in there that roll up to that number. Automotive and linear are roughly about the same at maybe somewhere between 15%-20% each.
we tend to be pretty low than in PCs. those are kind of the four buckets we try to drop things in and watch.
Awesome, Mark. Thanks so much.
Next question will come from the line of Tom Diffely with Davidson.
Yes, good morning. First, I want to dive into the comments you make about the decreasing efficiency of the parallel test going forward. How do those trends vary between your memory and logic side of the business?
I think that logic is more where we see the inflection points, where efficiency is starting to diminish. The memory side is really a harder area to characterize in any general way because, for example, as NAND flash devices go from toggle NAND to the serial interfaces, the connection to the device during test will become perhaps lower pin count or pumping more data through a narrower pipe. The tester resources diminish per device. Let's say CapEx per device may go down a bit. The capability per device goes up. I think the way you net it all out is, in memory, what memory's been able to do is grow bit growth quite dramatically and have a market that's flat at 500 million through a variety of test optimization techniques.
I expect that will continue. I do not expect the memory market will show some form of growth, whereas on the logic side, we do see these inflection points happening.
Okay. Do you see a wall for some of these tricks a few years down the road? is it a situation where, as far as you can see, it looks like efficiencies are going to continue to work in?
Well, that's really hard. We're not planning on a wall. For quite some time, people have thought there might be a wall, but I think the memory device has been. The one thing that's somewhat encouraging in memory that I will put out there as a carrot, but I'm not then going to change my flat market, is that the diversity of the types of DRAMs and the diversity of the types of non-volatile memories is growing. One of the things that has helped get efficiency in test is that design for test optimization was possible because the proliferation of device types was relatively few, and that design money and silicon area could be amortized over a lot of devices. As memories get a little more niche and custom, perhaps that will be less affordable, and it may be more economical to do test versus do design.
Okay. No, that's helpful. Then looking at your bookings, obviously the Asian bookings were down, normal seasonality there, but what drove the increase in Japan?
As we talked about, Japan's been very good market for us. One of the segments there that's strong is image sensors. Image sensors is a large part of that story.
Great. Okay. Thank you.
Your next question will come from the line of David Duley with Steelhead.
Yeah. Thanks for taking my question. Just a couple of clarifications. When you talk about the SoC market and the impact of this odd even cadence from the big apps processor customer, what generally is the delta in that segment of the market on an annual basis between the even and the odd years? Is it a $200 million increase or less? Could you give us some sort of idea what the swing factor here is?
You could actually look at our financials, and you could see we go from 1.4 to 1.6. We go back and forth. it can be roughly in that neighborhood. Now, this year, we didn't fall back for the reasons we outlined before, but if this was the historic pattern, we would've fallen back close to $200 million. You can also see in those even years, the margin percent goes down, but the dollars go up. that's the pattern we were largely referring to.
If we would naturally apply that pattern to this upcoming year, so we might naturally think that the SoC test business for you would be up at least $200 million.
It certainly could be up. You got to keep in mind there was $100 million of leased testers that hit this year that really were put in place last year, and that was more of a financing settlement this year, but the capacity was put in place last year. You might want to just think about that as well.
Okay. I think when you bought the Universal Robots business, you talked about 50% kind of annual growth, and that was what was already in place with the backlog and the customers that the company already had. I think it was your plan to add some larger electronics customers. Could you talk about what you're seeing thus far on that initiative and what kind of growth rate that science distribution could add to that initial growth rate?
Well, first of all, that initial growth rate wasn't predicated on any backlog. It's a high turns business. Sustaining 50% over multiple years really presumes penetration in a wide variety of markets. Now, in electronics test area, one of the things we describe there is that the sales cycle for our customers for these large deployments tends to be longer than the typical Universal Robots sales cycle to a small to medium size enterprise. A sales cycle there could be on the order of a handful of weeks, whereas with these large OEMs, it could be six to nine months, because they're just looking at different scales and evaluated at different levels. We're in the middle of several evaluations at our customers. Those started in the third quarter, and we would expect to see results coming there sometime early next year.
At the same time, Universal Robots, as I mentioned, the example of Wistron. Wistron's a large ODM. They manufacture a variety of things, laptops being one, laptop computers. They're in electronics assembly. They've put 60 robots in doing a variety of tasks around that piece. Universal Robots, on their own, has been penetrating electronics assembly. What we would expect to see is around tests, some incremental business in early next year.
Said another way, is the growth rate should accelerate sometime mid-next year based on you bringing more customers to the table.
It very well may. But remember, next year, we would expect them to be running at a 50% rate, close to 90 million. What we would add on to that is speculative next year, but it's probably in the first year in the order of a handful of five to 10 million kind of number.
Okay. Thank you.
Operator, we have time for just one more question, please.
You have a question from Krish Sankar with BofA.
Yeah, hi. Thanks for squeezing me in. Just a couple of quick questions, not to beat a dead horse on this mobile SoC increase next year. Now, I do remember a couple of years ago when the demand went up and you guys were fully sold out, lost some share to Advantest. I understand that's why you're building up inventory. Are you getting any indications from your customers to do it, or are you just being prudent in anticipation of a sharp ramp for next year?
Yeah, we don't want to talk about a specific customer, but all the signals we can read would suggest we should be well-positioned. We feel very good about this mobility complexity jump. We think it will happen. Every way we look at it, looks very good to us, but I can't get any more specific than that.
Got it. then just as a follow-up, with Universal Robots, looks like you guys have raised the profile for cobots in general, and I'm wondering, have you seen any increase in competition, especially from the Japanese robotic makers in the last few months? do you think it's still early that there's enough room for everybody to have their lunch and not try to grab the other ones?
Yeah, I would say we've seen no impact or slowdown or anything related to what is and will continue to be a lot of players trying to participate in the market. the market is growing rapidly. We're still a leading supplier, intend to stay there. Nothing from the competitors at this point has resulted in any impacts that we see.
Got it. Thanks Mark . Thanks, Greg.
Okay, thank you.
Great. Thank you everyone for joining us this morning, and those in the queue, I'll get back to you immediately after this call. This concludes the call, and we look forward to talking to you in the days and weeks ahead.
Thank you.
This does conclude today's conference call. You may now disconnect.