Ambarella Inc (AMBA) 2017 Q2 法說會逐字稿

Good day, ladies and gentlemen, and welcome to Ambarella's second FY17 earnings conference call. As a reminder, this conference call is being recorded. I would now like to introduce your host for today's conference, Ms. Deborah Stapleton of Investor Relations. Ms. Stapleton, you may begin.

Thank you, Christy. Good afternoon and welcome to Ambarella's second fiscal quarter 2017 financial results conference call. We appreciate you joining us today. Our speakers will be Dr. Fermi Wang, President and CEO; and George Laplante, CFO.

The primary purpose of today's call is to provide you with the information regarding our fiscal second quarter. The discussion today and the responses to your questions will contain forward-looking statements regarding our financial prospects, market growth, and demand for our solutions, among other things. These statements are subject to risks, uncertainties, and assumptions. Should any of these risks or uncertainties materialize or should our assumptions prove to be incorrect, our actual results could differ materially from these forward-looking statements. We are under no obligation to update these statements.

These risks, uncertainties and assumptions, as well as other information on potential risk factors that could affect our financial results, are more fully described in the documents that we file with the SEC, including the annual report on Form 10-K that we filed on March 25, 2016 for the FY16 and quarterly report on Form 10-Q for our first quarter FY17 that we filed on June 8, 2016.

Access to our second quarter results press release, historical results, SEC filings, and a replay of today's call can all be found on the Investor Relations portion of our website. I will now turn the call over to Dr. Fermi Wang.

Thank you, Deb, and good afternoon, everyone. First of all, I ought to point out that we released our official release at 1:05 PM today, but NASDAQ has failed to put our release on their website. If you haven't been able to locate one, please go to our official website, www.ambarella.com. You should be able to locate one there.

Now let me transition to our quarterly call. In Q2 2017, our revenue was $65.1 million, down as forecasted on our last call from $84.2 million of revenue in the same period over the prior year. The decline in revenue was primarily due to a decline in wearable sports camera market, as we have discussed. We have seen a near-term recovery in our China IP security camera business and are experiencing positive design win momentum in all of our current market categories.

During our last earning call, we discussed the possible impact of the Japan earthquake on the supply of Sony image sensor to our customers and to our revenue as a result. While some impact has been felt during Q2 and continuing into Q3, we believe that most supply issues will have been resolved by the end of the third quarter.

In the drone market, we are seeing significant design wins based on Ambarella's wide range of drone camera solutions. Our products span the high-risk quality professional models to consumer models that continue to hit new low price points. Our A12, A9SE and the H2 SoC families all offer 4K video resolution, high-quality photography, and advanced features, such as high dynamic range processing, and electronic image stabilization. Our ability to run many of these functions simultaneously provides us with a sustainable competitive advantage versus competing application processor solutions that are also failing to compete on video quality, features, power consumption, and the total solution costs. As a result, Ambarella is increasing its market share in all drone categories where HD or Ultra HD video is a requirement.

During the quarter, market leader DJI introduced its Zenmuse Z3 drone camera with integrated aerial zoom, featuring 12.4 megapixel stills and a 4K Ultra HD video. The camera includes a 3-access gimble and full 360-degree panning. The camera used Ambarella's A9 camera SoC.

Also during the quarter, China-based Yuneec began volume shipments of its Typhoon H drone. The Typhoon H won Engadget's Best of CES 2016 award and features a 360-degree view, gimble- stabilized 4K Ultra HD camera with 12.4 megapixels stills and 25 minutes of fly time. The camera is based on Ambarella's A9SE camera SoC.

Consumer drone manufacturers continue to drive price points down while offering intelligent fly control features, 4K video camera capabilities, and a smaller physical size. Sometimes referred to as selfie drones, these new models represent a significant opportunity for Ambarella, with our family of low power 4K camera SoCs. Many of the new consumer drones priced between $299 and $499 are now based on Ambarella solutions.

This week, Yuneec announced its new Breeze Selfie drone based on Ambarella's A9SE SoC priced at $499. The Breeze can take 4K video and 13-megapixel images and easily share them with social media accounts through its smartphone app. The Breeze comes with five automated flying modes, an indoor positioning system, weighs less than a pound, and has a flight time of 12 minutes.

And in August, China-based Wingsland announced its new S6 selfie drone, which is similar in size to an Apple iPhone S6 Plus smartphone. Priced at only around $300 and weighing only 230 grams, the H6 drone is based on Ambarella's A12 camera SoC. It includes 4K video, three-axis electronic image stabilization, wide-angle viewing, object avoidance, automatic follow me, and offers 10 minute flight time.

And in the sports camera market, customers are increasingly demanding 4K Ultra HD models with image stabilization, high megapixel photography, and extended battery life. This week, US-based Garmin introduced its Virb Ultra 30 action camera. Based on Ambarella's A9SE SoC, the camera supports 4K ultra HD video, 12 megapixel photos, three-axis image stabilization, and Wi-Fi connectivity. The Virb Ultra 30 also integrates arrays of sensors to support the capture of speed, pace, altitude, G4s, and other data and overlay them with the captured video.

In the professional IP camera security market, we are enjoying significant design win success for our S3L and S3Lm4K HDV camera SoCs, introduced at the China Security show at the end of last year. During the quarter, China sexurity giant Dauhua introduced its HF8431 camera series, utilizing Ambarella's S3L and offering 4 megapixel video resolution, 120-dB high dynamic range processing, 3-D noise reduction, and the support for the HEVC video standard.

We expect that our design win momentum based on our HEVC SoC solutions will continue in China as customers adopt the latest video standard and recognize our encoding efficiency advantages versus our competitors.

In the home monitoring market, we see continued design win activity, both from retail and the service provider customers, and an increasing focus on integration with smart home ecosystems. In June, Ambarella announced its support for new Apple HomeKit camera and doorbell services. A key benefit of HomeKit ecosystem is that all cameras in the home, including entrance video doorbells, home monitoring cameras, outdoor cameras, and baby monitors, can be viewed using the new Home app on an iPhone, iPad or Apple Watch. By offering both of these services on the widely deployed S2LM IP camera SoC, Ambarella accelerates adoption of the HomeKit video monitoring products into Apple's growing ecosystem.

At the Apple Worldwide Developers Conference in June, Apple highlighted a number of leading products based on Ambarella's platform and HomeKit services, including the Ring Video Doorbell Pro, FLIR Secure, Canary Plus, Kuna Smart Security Light, and August Doorbell Cam.

Also during the quarter, Comcast announced its new xCam indoor/outdoor home security camera, based on Ambarella's S2LM camera SoC. The xCam features HD video, enhanced Wi-Fi connectivity, wide-angle viewing, night vision, and integration with Comcast's Xfinity Home platform. The Xfinity Home app allows consumer to create rules that trigger other small home devices, for example, to capture video whenever a door sensor has been activated. Comcast has also added voice control to its smart home platform, allowing customer to see xCam video feed on their television, controlled by their X1 voice remote control.

In August, China-based Inomi introduced its smart home security camera as part of a Xiaomi ecosystem. Powered by Ambarella's S2LM camera SoC, the camera includes smart motion detection, full 1080P HD video, wide-angle viewing, and magic zoom capability. Its voice recognition allows the camera to turn 360 degrees to face the speaker and supports a voice assistant service.

And also during the quarter, Dahua released its new BM1 HDV baby monitor, under its LeChange consumer IP camera brand. Based on Ambarella's S2LM, the baby monitor offers tiny HD video, ultra-wide field of view, baby motion, and baby crying detection, in addition to other standard baby monitor functions.

In the automotive market, Ambarella has traditionally provided camera SoC solutions for video camera recorders or dash cams for aftermarket sales. We are now seeing revenue and the new design wins for video camera recorders that have been offered either as a dealer fit option or an integrated option in new OEM car models. Most of these new opportunities are in China, but also we began shipping SoCs for new European OEM models during the quarter.

During the second quarter, China car maker Geely began shipping its popular Emgrand EC7 sedan with a built-in dash camera based on Ambarella's A7LA automotive camera SoC. The camera features emergency video recording and still picture capture, while integrating with the car's in-dash GPS display system to provide touch screen digital video recorder control and video playback.

Also in Q2, Dongfeng Renault introduced its Kadjar model with dealer- fitted camera and digital video recorder capability to the Chinese market. Based on Ambarella's A12a automotive camera SoC, the camera is integrated into the rear view mirror and feeds video into the car's navigation and entertainment display systems.

So in summary, while Q2 revenue was impacted by the wearable sports camera market and Sony sensor supply issues, we expect to return to growth in Q3, as customer build for the holiday season. With strong design win momentum in security IP cameras, home monitoring cameras, and drones, and with new opportunities in OEM automotive and the virtual reality applications, we believe we are well positioned for continued growth driven by multiple markets.

Now I will hand it over to George for more details on the financials and our Q3 guidance.

Thank you, Fermi, and good afternoon, everyone. Today I will start with a review of the financial highlights for the second quarter of FY17, ending July 31, 2016, and then move on to the financial outlook for Q3 of FY17 that ends on October 31, 2016. During the call, I will discuss non-GAAP results and ask that you refer to today's press release for a detailed reconciliation of GAAP to non-GAAP results. For non-GAAP reporting for Q2, we have eliminated stock-based compensation expense adjusted for the impact of taxes.

Our Q2 FY17 revenue of $65.1 million represents a decrease of 22.6% from the $84.2 million of revenue in the same period in the prior year. In the second quarter, we had strong year-over-year growth in drones and home security revenues, offset by a substantial decline in the wearable sports market. On a year-over-year basis, overall revenues in the IP security market grew modestly, with professional security experiencing a small decline, offset by strong growth in the home security market. On a sequential basis, the IP security market grew modestly, with growth in both professional and home security revenues.

As forecasted on our Q4 and Q1 earnings calls, Q2 revenues from the wearable market were down substantially from the previous year and flat sequentially, due primarily to a significant decline in revenues from GoPro. The Q2 auto aftermarket revenues were flat year over year, primarily as a result of some softness in China.

The impact of the Sony sensor shortage was in line with our expectations, with the most significant impact in the quarter on new product launches. As most new product launches have limited on hand sensor inventory, purchases of other components were often delayed. We estimate the loss or delay in revenue was between $2 million and $4 million for the quarter.

Non-GAAP gross margin for Q2 of FY17 was 67.1%, compared to 64.6% in the immediate preceding quarter and 65.3% in the second quarter of the prior year. Gross margin in the quarter was higher than anticipated, primarily due to the benefit received of approximately $955,000 from the recovery and sale of inventory previously written down as a result of yield loss in the manufacturing process. In addition, a strong mix of higher end products in the drone and auto markets and less than expected revenues from the low end of the China professional security market contributed to a higher margin mix of revenues in the quarter.

Non-GAAP operating expenses for the second quarter were $23 million, compared to $24.3 million for Q1 of 2017 and $22.7 million for Q2 of the prior year. This year's Q2 OpEx declined from the previous quarter, due primarily to lower chip development costs and employment taxes related to our annual bonus payments in Q1. During the quarter, we also continued to monitor and control costs until we better understood the potential impacts to revenues from the GoPro inventory and Sony sensor issues.

Non-GAAP net income for Q2 2017 was $18.5 million, or $0.54 per diluted ordinary share, compared with non-GAAP net income of $29.7 million, or $0.88 per diluted ordinary share for the same period in the previous year. The non-GAAP effective tax rate in Q2 2017 was 11.5%. In the second quarter, the non-GAAP earnings per diluted ordinary share are based on 34.2 million diluted shares, as compared to 33.9 million diluted shares for Q2 of FY16.

Total headcount at the end of Q2 was 654, compared to 638 at the end of the previous quarter, with about 82% of the employees dedicated to engineering. Approximately 73% of our total headcount is located in Asia, primarily in Taiwan and in China.

We ended Q2 with cash and marketable securities of $327 million, adding $22.5 million of cash from operations in the quarter. Total accounts receivable at the end of Q2 were $33.6 million, or about 47 days sales outstanding. This compares to accounts receivable of $30.5 million, or 48 days sales outstanding, in the prior quarter. Net inventory at the end of Q2 was $20.9 million, or about 82 days of inventory, compared to $17.9 million, or 81 days of inventory, the end of Q1. Accounts receivable and inventory remain in line with Company targets.

WT Microelectronics, our Asia logistics supplier, represented approximately 72% of our revenue in the quarter, compared to 71% for the same period in the previous year. Chicony Electronics Company, a manufacturer of camera products for multiple OEM customers, as well as for their own distribution, represented approximately 12% of revenue in Q2 of FY2017, compared to 20% for the same period in the prior year. These customers were the Company's only 10% customers for the quarter.

I would now like to discuss the outlook for Q3 of FY17. We expect revenues for the third quarter of FY17, ending October 31, 2016, to be between $95 million and $99 million, representing an increase of between 2% and 6%, respectively, from Q3 of last year. For the full year revenues, we reiterate our full-year FY17 guidance of flat to down 5%.

We now believe our customers have a better understanding of the Sony sensor recovery plan. For more established companies' customers, we expect they will be receiving the majority of sensor deliveries by the end of our Q3, and therefore, we would expect a more normal seasonal pattern of deliveries in the second half. Overall in the quarter, the impact of the Sony sensor shortage should be in line with what we incurred in Q2.

In the third quarter, we are also expecting strong growth in wearable sports and professional and consumer IP security cameras. In the wearable sports and consumer security markets, we are seeing a strong build of product for the holiday season. In the professional security market, an expected recovery in China, at least in the near term, should help drive year over year as well as strong sequential growth. We see the drone and auto markets sequentially flat, coming off strong Q2 production builds.

We estimate Q3 non-GAAP gross margins to be between 63.5% and 65%, compared to 67.1% in Q2 of this fiscal year and 65.9% in Q3 of the prior year. We expect non-GAAP OpEx in the third quarter to be between $25 million and $26.5 million, with the increase from Q2 primarily coming from new chip development cost. We expect the non-GAAP effective tax rate to be approximately 12% in the quarter and estimate our diluted share count for Q3 to be approximately 34.9 million shares.

On the last earnings call, we announced that the Board of Directors had authorized the purchase of up to $75 million of ordinary shares over a six-month period. During the open period from June 6 to June 30, 2016, we repurchased a total of 405,089 shares in the open market for a total cash consideration of approximately $20.2 million, or $49.80 per share. The repurchase of shares was funded from the Company's working capital. Although the repurchase program does not obligate the Company to acquire any particular amount of shares, we will continue to evaluate the opportunity to repurchase shares for the remainder of the Board approved period.

I would like to thank everyone for joining our call today, and now we'll turn it back to the operator to manage the Q&A session.

(Operator Instructions)

Kevin Cassidy, Stifel Nicolaus.

Thank you for taking my question and congratulations on the great results. I wonder if you could give us an idea of the percentage of total revenue that home security, it seems like that market is starting to gain adoption. Can you give us an idea of what the percentage of revenue would be?

Yes, I think the home security side is probably in the high single digits right now and it'll increase a bit in the second half as a percent of revenue.

Okay.

Kevin, it's Fermi. I just want to add one more comment, because, for example, our big customer, like (Inaudible), they took chips for both professional consumer use. So it's very hard for us to get accurate estimates, so we just have to give you some rough guess that we can come out.

I see. And maybe to try to get another rough guess of, is this higher gross margin business or is it in line with what the IP professional business would be for gross margin?

Well, it's at the low end of the overall security business gross margin. And that's along with the China portion of the gross margin profile. There are other areas of security at the high end, where we do get margins up near our corporate margins.

Okay. Great. Thanks.

Ross Seymore, Deutsche Bank.

Good afternoon, guys. Good results here. This is Matt Diamond on Ross's behalf. The first question, are you still -- GoPro has mentioned that they're on track with their new products for the second half. Are you seeing anything to sway them or anything to either endorse or refute that?

(Laughing) As always, we don't comment on our customer product launches.

Understood. And on the gross margin side of things, it looks like you got enough gross margin pretty nicely. What is behind that? Is that purely a function of mix or is there something else to call out there?

As I said in the earlier comments, we did have the recovery of almost $1 million of inventory that was previously written down through yield loss and the manufacturing process. And that happens when the chips go through tests and don't pass at a higher specification, they go into a yield loss bucket. If we can find opportunities to use those chips at lower specifications, we run them back through tests and we recover those and sell them. So in Q2, we had an unusual high amount of previously written off inventory which we were able to retest and sell. So normally, we have that on a regular basis, but it's usually a much smaller number. We just got, I think, a couple really good opportunities to reuse some of these chips and we reprocessed them. We will still have a little bit of that in Q3, but it will go back to normal levels after that.

And then on the product side -- I was going to continue -- on the product side, our China security business was lower than we originally anticipated, which is our lower margin business, so therefore the mix itself was more positive than we had anticipated.

Great. Thanks so much.

Matt Ramsay, Canaccord Genuity.

Thank you very much. Good afternoon, guys. Fermi, I thought it was interesting in your prepared remarks that you talked about the momentum that you guys are seeing in design wins in the professional security space on 4K and HEVC cameras. I know that there's been some debate and consternation within the investor ranks about how applicable 4K resolution and HEVC compression are in different tiers of that professional security market. I guess the question is, how big are the different tiers? For what tiers are 1080 P good enough and for what tiers do you need 4K? Maybe you could give a little color on what you're seeing in that market. Thanks.

I think we need to separate 4K and the HEVC to two different buckets and talk about it. Let's talk about HEVC first. I think outside China, HEVC is not popular at all, due to the royalty issues. But in China, we start seeing that HEVC standard being widely adopted and a lot of people are using them for the Chinese internal consumption. So I think that HEVC definitely is a China market only.

In terms of the 4K, while in fact, what I say was a 4 megapixel. So even 4 megapixel is much higher than traditional full HD video. We do see that the trend for the professional IP security camera is to get a higher pixel, because you want to get the most resolution, so you can do video analytics on it. And the resolution definitely matters for that purpose. So we start seeing the momentum moving toward to higher pixel. We start seeing 4 megapixel become very popular. I think 4K is the next trend. So we haven't seen 4K in the mainstream yet, but I do believe 4 megapixel for next year could be a mainstream product.

No, that's really helpful. Thank you. I guess my follow-up question is on the automotive market. It was very interesting in your prepared remarks to hear about design wins, and not just design wins, but shipments into automotive OEM cameras. Maybe you could talk a little bit about the different applications there, what kind of ASPs our margins you expect out of that versus the business that you've had in automotive aftermarket cameras up to this point and when that can be material to the overall P&L? Thank you.

So for the automotive OEM business, there are multiple possible application for us. In the near-term application, the most, the one that we talk about is being a video recorders function for the OEM business and we announced several customer products in our release today. And we also have a lot of momentum. In terms of ASP, I think it's a little higher than our aftermarket product, because they ask for higher spec, like higher temperature range, and some of it require (Inaudible) type of auto grade chip. So we do believe that ASP and gross margins is higher than our typical aftermarket dash cam.

Outside of the video recorders, there are other video applications available for the OEM business. We talk about AVM, basically surround view. We talk about electronic mirror in the past. Those two application, I would say it's a medium-term opportunity for us. We demo'ed technology at CES this year and we are continuing to demo to our customer, and we got very good feedback from the customer on these two applications. So I think we talk about the end of next year, we're going to start seeing some revenue from those two markets, and we're going to stick to that.

In terms of ASP and gross margin, I think it's still too early to talk about. The longest term opportunity for us is really become ADAS and autonomous car, which I think is still four or five years away.

All right. Thank you very much for taking my questions. I'll get back in the queue.

Quinn Bolton, Needham and Company.

Hello, guys. Let me offer my congratulations, as well. Fermi, just wanted to follow up on those comments on ADAS and autonomous car opportunities, and knowing that it's four or five years out, you guys have previously talked about taping out your first machine vision chip later this calendar year. Just wondering if you could give us any updates on where you are with that project, and perhaps more importantly, any hints you can give us on what sort of features you're going to be including in that chip? Is this going to be a machine learning or deep learning solution, or is it really going to focus on video preprocessing and trying to get the best video into the autonomous vehicle or drone? Thank you.

First of all, the target, you know that we acquired VisLab, which is a company that is very famous to video analytics for autonomous car. So our intention to convert their complete software stack implementation for autonomous car onto our first chip. But application for that chip is beyond just car. We believe the autonomous drone and other application will require similar functions that you can detect objects, you can do [prescient] learning on that. All those function will be part of our chip. And our goal is that our chip will be able to implement all of the VisLab functions when the chip is available. We talk about late this year or early next year when (Inaudible) chip was still on the schedule to do that. And we hope for that some time next year we'll be able to demo the chip.

Great. And then, George, I know that the gross margins here benefited from some of the previously sold, previously written down inventory. But even with the China security business coming back in the October quarter, plus the ramp of the consumer business, your gross margins are still well above your long-term corporate average. Can you just give us a sense, do you think that you are doing a better job with the operations team to keep those margins higher or is there something you see, say, next year that will bring the gross margins back down into the corporate average range that you project?

Well, I haven't done a lot of work on next year yet. There's a lot of variables still being looked at. For this year, though, we are going to benefit a little bit from recovery of some written down parts in Q3, which helps. We also won't see the really big ramp on the low end of the China security business until the exit to the quarter and into Q4. So we will see, I think, the margins float down a little bit this year.

I would expect to have a better picture of next year. I think right now, the biggest wild card for next year is where is China security headed? We have a recovery. We're forecasting some recovery in the second half. But at this time, I don't think we're willing to commit that this is a permanent recovery in the China market. I think we'll know more towards year end. So if that market expands and continues to recover next year, we would see our margins flow down probably into the high end of our target model in the early part of the year.

Great. Thank you, George.

Joe Moore, Morgan Stanley.

Great. Thank you. I guess following up on that gross margin point. Can you talk about, as you roll 14-nanometer out into more end markets, what does that do for your like-for-like gross margin? Is there any big change there?

I think it's going to depend on the market and where we are. Because we will be rolling out lower end 14-nanometer chips, as well as high end. So it will be dependent on how those markets mature and at what rate. Obviously, we think at the higher end, our margins still should be very strong. But we will continue to compete at the lower end of these markets with 14 nanometers. And we would expect those margins probably be in line with what we're selling at 28 and 32. So I think 14 nanometer's probably not going to have that much impact, just being 14 nanometers, it's going to be more the mix of how those chips roll out.

And Joe, this is Fermi. Let me add one more comment. I think when we look at our how we differentiate ourselves against our competitors, we talk about before it's a continue to improve performance with our inquiries in the power budget. And 14 nanometer plays a sweet spot on that. And just like George said, we are going to use 14 nanometer for low end [fet]. We just tape our second 14 nanometer chip just this month, and we're going to continue to work on that. And we're probably going to start talking about 10 nanometer chip very soon. So you can see that we're going to continue to invest highly on this process node, so that we will be continue to be competitive on both performance, as well as on the power consumption.

Okay. Great. That's very helpful. Thank you. And then separately, the loss, the $2 million to $4 million of lost revenue due to the image sensor problem for, I guess, the two consecutive quarters, what happens to that revenue? Is there some of that that's regained as the supply chain catches up, or is that driving business to other image sensor suppliers and therefore, other video processing suppliers?

It's a combination of both. In certain markets, it would've driven the sales to other sensors and potentially other suppliers, particularly in China, in the security market. But in a lot of the cases, the new rollouts will just be delayed. So we will see some pick up and roll through, potentially some of that will come back in Q3 and Q4.

Great. Thank you very much.

Charlie Anderson, Dougherty and Company.

Great. Thank you for taking my questions. I wanted to start with the drone market. You did talk about how that was very strong in the quarter and you are expecting it to be down, take a breather in the October quarter. I wondered, is that what's going to be typical here? Was the spike in July abnormal? And then big picture, with the drones coming down to $300 to $500 level, the selfie drones that you mentioned. What does that suggest about unit volumes going forward from here and the momentum of that category?

I'll take, I guess, the first part of that question, where the timing of revenues in the drone market. The consumer portion of the drone market we eventually will expect to see that probably go to the normal holiday season build. I think what you're seeing now in the drone market is companies attempting to get their products out as fast as they can, regardless of the timing during the year. So as companies are building out new products, they're launching immediately. I would guess by next year, you'll start seeing more seasonality to that marketplace.

So this is Fermi. Let me (Inaudible) about that pricing point. I think we definitely continue to experience a fast growth in drone market. We see a lot of design win activity. We see our customer start ramping up volumes. But also, I think the fast growth is driven by two things. One is definitely the customer demand. The other one is the price erosion. The ASP of 4K drone was above $1,000 just last year, when DGI introduced their first drone. And quickly, early this year at CES time, DGI took down the price, and so that the 4K dropped way below $1,000. And now with the selfie drone that we just talked about, the ASP dropped to $299, (Inaudible) that range. I really think with that price erosion, the volume ramping up is mainly associated with that. And we expect the $299 to $499 range of selfie drone will be the biggest market for drone in the next 12 to 18 months.

If I could sneak in one follow-up, just going back to some of the questions around autonomous vehicles. There's certainly been a lot of announcements over the past few months of people in that market. I wonder, Fermi, if you could address in terms of where you guys fit, do you guys feel like there's a need for partnerships with other people with more processors in the car, tier 1s? As you start to point toward that market, what's going to be required for you guys to be a player in that market?

Our partner will be tier 1. And also, we need to work with OEM together. So for example, if you look at the software we acquired from VisLab, we have a complete software stack. Because VisLab software can drive the car already. And it's our intention to drive the complete software stack on our chip. So I don't think we need to partner with another processor company. But we do need to partner with a tier 1 who is willing to use our solution when we are ready. So I think that's our intention. But Iike I said, for autonomous car, we put a lot of effort to enable us, but I do believe our near-term revenue opportunity is other market. For example, I do believe autonomous drones is going to be a very important market for us in the near term and our first CV chip will be addressing that market also.

Great. Thanks so much.

Jorge Rivas, Craig Hallum.

Thank you for taking my question. It's Jorge on behalf of Richard today. Our first question, on the last conference call, it was assumed that Sony was going to get back to full utilization at the end of August. But it seems like they have actually reached that level at the end of July, so they are a month ahead of schedule. So I'm wondering, what's the reason why you haven't changed your assumptions on the supply chain availability of the sensors? Whether it's conservatism on your customers end, or if you can provide some color on that, it would be great.

Yes. I think first, we didn't say they would be full production by the end of August from a standpoint of creating or fixing all the backlog situation. They said they would have the lines up and running full capacity by the end of August, which means it would take numerous months after that to catch up on the backlog. So it's a little bit different interpretation. So it's really kind of rolls out as we expected there.

Okay. And then one last question for me. On your development of 14 nanometers products, can you give us an idea when you expect to have some, when do you expect to have those products on the market, whether that's a near-term driver for your business or would it be more of a calendar 2017 story?

Well, I think the first class will use our 14 nanometers chip will be in production probably early next year. But I think because of the benefit of power consumption, I think there will be some customer will quickly transition to 14 nanometer products because of the power consumptions and most critical demands for that. And also that we think that because the performance that we can upgrade for our first generation 14 nanometer chip, we see a lot of momentum for virtual reality cameras, so that we are going to see customers who are really willing to pay for the power consumption, as well as the customer who wants to do virtual reality. Most (Inaudible) will be the first ones to adopt 14 nanometer chips.

Okay. That's excellent. Thanks, guys. That's all for me.

Matt Ramsay, Canaccord Genuity.

Hello, guys. Thank you for letting me get back in and follow up. Fermi, I wanted to follow up on the drone market for a second. It seems to me it's bifurcating a little bit. At the really high end, you're seeing drones with two chip solutions, sometimes with your chips alongside product from Intel and others at the high end of the market, and then in the lower tiers of the market and selfie drones, et cetera, that you're talking about ramping in volume, I imagine it's a one chip solution. Maybe you could talk about those two markets, how they differ and maybe how your competitive position is different between the two? Thanks.

Right, thank you. For the high-end drone, you are absolutely right. For example, the TGI Phantom 4, they use a multiple chip solution for their full implementation. They use our A9 chip for the 4K video capture and they use some other chip for the computer vision and also flight control. So I think the high-end drones will stay that way, because they have much less pressure on the price, but they have a lot of, because of their price, they have to prove to be the best quality in terms of the product quality and flight control quality and video quality. So I think they will probably continue to use multiple chips for their solution.

On the selfie drone, it's a little different, because when you drop to $299 or $499, the numbers can definitely become an important portion of your bond. In fact, there was a first drone based on Qualcomm solution was introduced just several weeks ago. So we have a chance to compare the Qualcomm based drone with drones based on our low-end solution, A12, and for the first time. And we believe that A12-based solution has advantages on video resolution, EIS, power consumption and total system costs. And for example, the Qualcomm-based solution can run only 10 kp 30, compared to A12's 4K video, while A12's power consumption is significantly lower than Qualcomm's. And also, that integration that our customer prove that they can use our low-end chip to go down to that price point with a very good quality on the video, as well as on the flight control. So I really think that the selfie drone that's being introduced with Yuneec and Wingspan, and we believe there are several more coming up in the near future with (Inaudible) solution, will prove that we are the leader in the selfie drone market, too. Thank you very much. Appreciate it.

Brad Erickson, Pacific Crest Securities.

Hello, guys. Thanks for taking my questions. Just a couple of clarifications. First, around the drones and the action cameras, obviously seeing, looks like significant builds ahead of the holiday. Relative to your guidance, can you talk about your confidence level and visibility around any potential for inventory correction and maybe any measures you've taken within the forecast to hedge against that?

Yes, I think we're pretty confident of the near-term forecast for those two markets. Obviously, we'll have a lot of PO coverage and visibility of forecast from customers. But I think with that said, we are, you notice we're having a reasonably strong Q3, but we have not changed our guidance for the year. And I think until we see sell through in most of these consumer markets, we're going to stay conservative.

Got it. And then just a quick clarification on the $2 million to $4 million impact from Sony. Was that, it wasn't quite clear, was that in line with what your forecasting assumptions were or was that at all incremental or better than expected?

Q2 is pretty much in line, with Q3 being better than our expectations.

Got it. That's helpful. Thank you.

Thank you. And that does conclude our Q&A session for today. I would like to turn the call back over to Mr. Fermi Wong for any further remarks.

Thank you. I would like to thank all our employees for their continued support and performance, and most important, thank you all to joining our conference today. See you next time.

Ladies and gentlemen, thank you for participating in today's conference. This does conclude today's program. You may all disconnect. Everyone, have a great day.