Ambarella Inc (AMBA) 2019 Q1 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the Ambarella First Quarter Fiscal Year 2019 Earnings Conference Call.

As a reminder, this call is being recorded.

I would now like to introduce your host for today's conference, Ms. Deborah Stapleton, Investor Relations.

Ms. Stapleton, you may begin.

Hello, and thank you.

Good afternoon to everyone, and welcome to Ambarella's first fiscal quarter 2019 financial results conference call.

Our speakers today will be Dr. Fermi Wang, President and CEO; and George Laplante, CFO.

The primary purpose of today's call is to provide you with information regarding our fiscal 2019 first quarter results.

The discussion today and the responses to your questions will contain forward-looking statements regarding our projected financial results, 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 filed with the SEC, including the annual report on Form 10-K that we filed on March 30, 2018, for the 2019 fiscal year.

Access to our first 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 the website.

I will now turn the call over to Dr. Fermi Wang.

Thank you, and good afternoon, everyone.

Our Q1 fiscal year 2019 revenue was at $56.9 million, representing a decrease of 11.2% from $64.1 million of revenue in the same period over prior year.

While revenue from our focus markets, IP security and automotive, grew year-over-year, revenue from drones, sports and VR cameras declined.

During the quarter, we continued to make steady progress in automotive and IP security markets with strong design-win momentum and revenue growth.

The development and the rollout of our CV solutions is also progressing well and in line with the plans shared with you at the recent Analyst Day event in March.

During the quarter, we delivered a working silicon on 2 new computer vision SoCs, CV22 and the CV2 based on our new CVflow architecture and offering significant increases in Convolutional Neural Network or CNN processing performance.

We have now successfully ported a large number of our public domain neural networks onto CV22 and have [major] the performance and the power consumption.

When we compare CV22 with CV1, we compare that to the CV22, the CNN performance is roughly 4x higher than CV1, as anticipated.

Additionally, when we compare the CV22 numbers to competing solutions based on GPU or DSP architectures, we have achieved results that are many times higher in terms of processing performance per watt.

In March, we successfully demonstrated to customer and investors our fully autonomous vehicle, EVA, or embedded vehicle autonomy on Silicon Valley Road.

EVA navigated various traffic scenarios presented by Silicon Valley's challenging urban environment.

The fully autonomous car combines software and algorithms based on over 20 years autonomous vehicle research with Ambarella's low-power CV1-embedded computer vision processors.

EVA's high-resolution SuperCam stereovision cameras delivered a 360-degree short- and long-distance viewing capability required for advanced perception and precise self-location.

They provide a perception range of over 150 meters for stereo obstacles detection and over 180 meters for monocular object classification.

Its features include automatic calibration, stereo generic obstacle detection, terrain modeling, traffic light detection, 3D free space detection, lane detection and the CNN classification for vehicle, pedestrians and bicycle, motorcycle.

EVA's sensor fusion software combines vision information with radar and map data to provide the information necessary for path planning and merging maneuvers.

During May, Ambarella's VisLab team commemorated the 20-year anniversary of Mille Miglia in Automatico, one of the world's first tests of autonomous vehicle on public roads.

With this test, VisLab, which was then part of the University of Parma, was first to use vision technology to autonomously drive a car utilizing only low-cost camera and the computing solutions.

To celebrate our anniversary, Ambarella began demonstration of EVA to key European customers on public roads in Italy.

As a result of the successful EVA demonstration in North America and Europe, Ambarella's SuperCam cameras, as used in EVA, has now been delivered for evaluation to automotive OEMs, Tier 1 suppliers and emerging autonomous vehicle companies.

Feedback from customers is that Ambarella solution are successfully addressing the inadequacies of a currently deployed system and meeting the requirements for next generation of automotive camera systems.

This include high accuracy perception, low-light imaging, recording as well as high-resolution long-distance stereovision and the monocular vision with advanced CNN processing.

Ambarella is seeking opportunities for its automotive computer vision SoCs in ADAS cameras, both stereo and monocular, with customers evaluating our solution for camera targeting Euro NCAP 5-star requirements.

It is estimated that by 2020, the available market for SoC solutions for such cameras will exceed $2 billion.

Additionally, we are seeing interest in camera systems, enabling higher level of autonomy, including level 2 through 3 and the supporting features, including lane keeping assist, advanced cruise control and parking assist.

We are also engaged with emerging leaders in the development of fully autonomous vehicles, leveraging the stereovision technology to develop with EVA.

And during the quarter, we continued with a working silicon on our new CV2 computer vision SoC based on our CV4 architecture.

The CV2 provides up to 20x the CNN performance of CV1 used in the current SuperCam cameras and represent a very significant advance in available performance processing Deep Neural Networks.

The performance it delivered was extremely low-powered, allowing deployment by our customers in demanding automotive security and the robotic application.

In summary, we are pleased with our progress in delivering advanced computer vision solution from automotive markets based on our execution and the customer feedback.

And in our existing automotive business, we continue to see video recording be more widely adopted, primarily in Asia.

During the first quarter, we enjoyed record sales for our SoC for using OEM automotive car recorders, driven primarily by major Japanese carmakers, including Toyota, Honda and Nissan.

In China, automaker, Geely, introduced its new Lynk & Co 02 model which includes a full HD car recorder based on Ambarella's A12A automotive camera SoC.

And recently, we have noted with interest that Chinese government is considering legislation to mandate the inclusion of car recorders in new cars, a move which could significantly increase the size of our market opportunity.

Additionally, German courts have now approved submission of video evidence from car recorders, clearing the way for wider adoption of such cameras.

And in China, we have been working with Tier 1 automotive supplier, O-Film technology, a leading supplier of automotive electronics and the camera modules on the development of both single channel and the triple channel electronic mirror designs.

Based on this relationship, we are pleased to report that during the quarter, Chinese automaker, BAIC, introduced its BJ40 SUV, the first vehicle to offer an electronic mirror based on Ambarella's A9AQ automotive camera SoC.

Electronic mirror provides a wide and unrestricted rearview and leverages the A9AQ advanced image processing, low latency, HDR and LED flicker mitigation capabilities.

It provides excellent image quality even in lowlight and high contrast situations.

I will now give an update on our progress in IT security markets for both professional IP cameras and the consumer home monitoring applications.

At the ISC West Security show in April, we demonstrated our CV22 camera SoC based on our CVflow architecture.

The CV22 combines significant advancements in image processing HDR, industry-leading HEVC encoding efficiency and advanced computer vision processing in a very low power SoC.

It provides 4x the Deep Neural Network processing performance of our first generation CV1 chip.

We have now sampled CV22 and our CVflow software development tools to key security customers.

Additionally, with the addition of the new CV2 SoC 12 CVflow family, customers can now develop multiple camera designs with different price and the performance point while leveraging software developments across a common hardware, neural network processing and software platform.

During the ISC West show, U.S.-based thermal imaging specialist, FLIR, introduced its Saros line of IP cameras, combining 2 thermal sensors with a 4K video and video analytics.

Thermal sensors provide the ability to deliver wide area monitoring in adverse weather conditions, while high-resolution optical cameras realize accurate video identification of intruders.

The Saros is the first camera to be released based on Ambarella's new S5L camera SoC.

In our IoT home monitoring market, we are continuing to see new opportunities in a wide range of products, including video doorbells, battery-powered cameras, HD baby monitors and indoor/outdoor cameras.

During the quarter, service provider, Cox Communications, introduced its iCAM 2, home monitoring camera, as part of its Homelife advance security and automation service.

The iCAM 2 HD camera is based on Ambarella's S2LM SoC and offer indoor/outdoor operation, motion sensing capability and live video streaming.

Also, during the quarter, Safety 1st, a leading retailer of baby products, introduced its HD Wi-Fi baby monitors.

Based on Ambarella's S2LM, the baby monitor provides HD wireless streaming 2-way audio and a free cloud storage.

In the battery-powered camera category, Shenzhen-based camera maker, Foscam, introduced its E1 cameras based on Ambarella's ultra-low power S2LM EX camera processor shipped with a base station that supports local video storage.

The camera features extended battery life, full HD video and smart motion detection.

We have also been working with French drone maker, Parrot, the leading European drone group and the world #2 in the consumer segment on a new drone design that will combine outstanding video quality with ease of use.

Since the announcement is tomorrow morning, I will not steal their thunder but point you in their direction for tomorrow's announcement.

Also, during the quarter, Snap introduced Spectacles 2 which has a number of enhancement over the first generation, including high-resolution videos and photos as well as being waterproof.

The new Spectacles 2 slim form factor presents the state-of-the-art in terms of integrating an HD camera into an extremely small, body-worn design.

In summary, while we see headwinds in sports and the VR camera market, we believe that automotive and IP security markets represent a large and extending opportunity for Ambarella.

We have made great progress in delivering advanced computer vision SoC to key customers in both of these markets and are very encouraged by the feedback we received.

We will continue to focus on this market while identifying new markets such as robotics that will require advanced computer vision.

With that, I will now hand back to George.

Thank you, Fermi, and good afternoon, everyone.

Today, I will review the financial highlights for the first quarter of fiscal 2019 ended April 30, 2018, as well as the financial outlook for Q2 of fiscal year 2019 and update the outlook for the full fiscal year.

During the call, I'll 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 Q1 through fiscal '19, we have eliminated stock-based compensation expense adjusted for the impact of taxes.

Our Q1 2019 revenue of $56.9 million represents a decrease of 11.2% from the $64.1 million of revenue in the same period in the prior year.

Non-GoPro revenue declined approximately 12.5% in the quarter.

In Q1, we saw year-over-year growth in IP security and auto OEM offset by a substantial decline in year-over-year -- in drone revenues.

In IP security, year-over-year growth was led by professional security revenues, primarily from expansion of revenues from the China customers.

Solid revenue increases from OEM automotive video recorders in Japan and China resulted in year-over-year growth for the auto market.

Consumer drone revenues in the quarter were down substantially from last year with the majority of our revenue coming from high-end drone market, primarily at DJI.

Other wearables declined from last year as weak sales in the VR and non-GoPro sports camera markets offset growth in police cameras.

Although an administratively complicated and expensive, the change to average selling price model under the new revenue recognition rules did not materially impact the revenue in the quarter.

Non-GAAP gross margin for Q1 was 61.8% compared to 64.7% in the preceding quarter and 64.3% in the first quarter of the prior year.

The lower gross margin both from previous quarter and prior year was the result of a significant mix change as lower margin China security revenues increased and higher margin drone revenues decreased as a percent of sales.

Non-GAAP operating expenses for the first quarter were $31 million compared to $29.5 million for the previous quarter and $25.7 million for Q1 of the prior year.

OpEx increased from the previous quarter primarily as a result of increased headcount cost, including payroll taxes, as we continue to grow development and support headcount for the computer vision opportunities.

Non-GAAP net income for Q1 was $4.5 million or $0.13 per diluted ordinary share compared with adjusted non-GAAP net income of $15.3 million or $0.44 per diluted ordinary share for the same period in the previous year.

Non-GAAP effective tax rate for Q1 of 2019 was approximately 10%.

The non-GAAP effective tax rate for fiscal '19 includes a change in the allocation of stock-based compensation expense across the company's tax jurisdictions to improve alignment of the non-GAAP tax rate to the GAAP tax rate.

For Q1 2018, non-GAAP net income was adjusted from $13.4 million to $15.3 million and EPS was adjusted from $0.39 to $0.44 per share for comparative purposes.

The change has no impact on GAAP net income or tax rates.

In the first quarter, the non-GAAP earnings per diluted ordinary share were based on 34.5 million diluted shares as compared to 34.7 million diluted shares for Q1 of fiscal year 2018.

Total headcount at the end of Q1 was 733 compared to 706 at the end of the previous quarter, with about 81% of employees dedicated to engineering.

Approximately 72% of our total headcount is located in Asia.

We ended Q1 with cash and marketable securities of $414.1 million, adding $64,000 of cash from operations in the quarter.

During the quarter, the company repurchased 437,448 shares at an average price of $47.13 for total consideration of approximately $20.6 million.

The total number of shares purchased since the inception of the repurchase program in June of 2016 is 1,937,332 shares at an average price of $49.32 for total consideration of approximately $95.6 million.

As of May 1, 2018, there was approximately $11.1 million available for repurchase under the repurchase program through June 30, 2018.

In addition, Ambarella Board of Directors has authorized the repurchase of up to an additional $100 million of its ordinary shares over the 12-month period commencing June 5, 2018.

Repurchases under the new program may be made from time to time through open market purchases or through privately negotiated transactions subject to market conditions, applicable legal requirements and other relevant factors.

The repurchase program does not obligate the company to acquire any particular amount of ordinary shares and it may be suspended at any time at the company's discretion.

The repurchase program will be funded using Ambarella's working cap.

Total accounts receivable at the end of Q1 was $25.7 million or 40 days sales outstanding.

This compares to accounts receivable of $31.3 million or 41 days sales outstanding at the end of the prior quarter.

Net inventory at the end of Q1 was $23 million or about 96 days of inventory compared to $23.4 million or 80 days of inventory at the end of Q4.

We do not anticipate any valuation issues as a result of the higher inventory levels as the inventory consists mostly of high-volume chips for our primary markets.

We had 3 10%-plus revenue customers in Q1.

WT Microelectronics came in at 58.3% of revenue compared to 72.3% in the first quarter of the prior year.

Hakuto came in at 12.3% compared to 4% in the first quarter of the prior year.

Hakuto is our Japanese distributor selling primarily to the local auto OEM market.

Chicony came in at 11.2% compared to 7.1% in the first quarter of the prior year.

I would now like to discuss the outlook for Q2 and update you on the fiscal year 2019 outlook.

We expect total revenues for our second quarter ending July 31, 2018, to be between $60 million and $64 million, representing a decrease of between 16.2% and 10.6%, respectively, from Q2 of last year.

In the quarter, we expect GoPro revenues to be resulting in non-GoPro revenues at the midpoint of guidance, declining approximately 8.6% from last year.

In the quarter, we expect year-over-year revenue increases in IP security and auto OEM markets, offset by declines in drones and other wearables markets, primarily virtual reality and non-GoPro sports cameras.

We now believe revenues from customer -- consumer wearable market such as virtual reality and non-GoPro sports cameras will decline this year particularly in the second half as compared to the prior year.

For example, sales in virtual reality and non-GoPro sports camera markets so far this year are significantly below our expectations, and based on customer feedback, we see this continuing through the normally strong holiday selling season.

This, combined with previously discussed decline in drones, will negatively impact the remainder of the year.

As a result of these changes, we now see the non-GoPro revenues being plus or minus 3% as compared to fiscal '18 non-GoPro revenues of $258 million.

In addition, we expect GoPro revenues to be immaterial to the total revenues for the remainder of the year as compared to approximately $37 million for the full fiscal year 2018.

The increase in security as a percent of revenue for the year will bring our gross margins into the low end of our target range of 59% to 62% for the year.

The current uncertainty arising from both the U.S. China trade discussions and the U.S. House of Representatives' proposed ban on security cameras from Dahua and Hikvision has not been considered in our current forecast as the impact is not predictable at this time.

We continue to monitor the potential risk and associated impact that may result from trade discussions such as Chinese companies shifting to sourcing more from domestic Chinese semiconductor suppliers.

We do supply Dahua and Hikvision chips that are used in security cameras for both their domestic and export business but do not have data on what is sold into the U.S. government or commercial markets.

As we supply security chips to the major U.S. and European security camera manufacturers, we would expect to be able to offset some of the potential loss from China customers if the U.S. government banned the use of their products.

We estimate Q2 non-GAAP gross margins to be between 59% and 61% compared to 61.8% in Q1 of 2019 and 63% in Q2 of the prior year.

As forecasted, we are seeing the increase in China professional security revenues and a decline in drone revenues as having the largest negative impact to margins in the quarter.

We expect non-GAAP OpEx in the second quarter to be between $32 million and $33.5 million with the increase from Q1 primarily coming from increased chip take-out fees and engineering headcount increases.

We estimated our diluted share count for Q2 to be approximately 34.2 million shares.

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

Operator?

(Operator Instructions) Our first question comes from Matt Ramsay with Cowen.

Fermi, you guys had some pretty impressive displays of the autonomous driving stack at the Analyst Day.

And I understand the intention from Alberto, yourself and Les was to get those demonstrations in front of customers for initial feedback pretty quickly after the Analyst Day.

Maybe you could give us a little bit of commentary about how those demonstrations went with customers, what particular feedback was, and if there's any paying points that they're dealing with their current camera suppliers that your solutions might address.

Thank you.

So first of all, I think the customer feedback is very positive and I think we -- I think everybody understand we bring some special value in terms our perception system for different market.

For example, we talked about 3 different, let's say, category of a customer.

One is ADAS.

The second one is level 2, level 3. The third one is level 4 above autonomous driving.

For ADAS customer, they view us as a SoC supplier to really provide a silicon solution for Euro NCAP type of 5-star requirements, and this kind of requirement really become very challenging as well as become high demand in terms of performance requirements as well as CM performance as well as perception quality.

So I think that market definitely is an area which we're going to focus on because it really gives us the fastest way to go to a revenue in terms of automotive computer vision revenue.

And the feedback was that we definitely proved that we have quality, not only on the video quality side but also on the computer vision side, now processing performance, lower power consumption and also accuracy to deliver a solution for ADAS.

So that's one category.

In terms of level 2, level 3, we view as a perception subsystem that can really go in their current system to improve their perception system, accuracy and quality, particularly not only of the image processing side but also on the high level lane keeping assistance or advanced cruise control and parking assist, those kind of high-level functions that we definitely add value there.

And then the third category is really on the autonomous car, which is really long-term development but I think that's strengthen our VisLab and we will continue to engage customer there.

And in there, we're also being viewed as a perception subsystem to augment what they have right now.

So I think this is where the feedback from the customer and really that -- or the single SoC for ADAS or perception system for level 2 above really is where our customer are looking at us right now and where we're going to continue to provide solution.

Got it.

That commentary is helpful.

George, just a couple of questions for you.

One is around the comments that you made about U.S.-China relations with respect to where chipsets can go for both China domestic and, I guess, China exported security cameras.

Maybe you could give us a little color at where we should look and watch for comments and how quickly you expect a decision to potentially be made there.

And then second, you talked about operating expenses continuing to ramp maybe in the face of a little bit more challenging revenue environment.

Do you have an OpEx forecast for us for the full year just to sort of bound where spending can go over the next few quarters?

I'll answer the second one first.

On the OpEx for the year, I had given previous guidance of $120 million to $130 million non-GAAP.

That's excluding stock-based compensation.

I think that range is still a proper range.

We probably right now are targeting or be successfully hitting the high end of that range.

On the China situation, I think there's a couple moving parts there.

The first one is particularly around the security space with Hikvision and Dahua potentially being excluded from government sales.

In that case, I think it's really going to be a function of: number one, whether the House of Representatives bill passes the Senate; and then we would have to judge the transition in the U.S. and Europe from China-based products to U.S. and European-based security products.

I think we would participate still in that.

We still have a very high market share outside of China security.

So I feel we could retain some of that business.

When it comes to the second issue, is the basic trade negotiations going on between the U.S. and China, obviously, that is very difficult to project what the impact is.

I think we have to watch over the next month or so how China and the U.S. resolve the ZTE issue and resolve the tariff issue to determine if this escalates into something that could impact our sales to China customers.

I don't think there's anything specific other than watch what's in the press on a daily basis as to what changes take place there.

Our next question comes from Ross Seymore with Deutsche Bank.

First question is on some of the new CV chips, the 22 and the 2. Maybe more specifically on the 2, Fermi, you talked about you got the first silicon back and that's working well.

Can you give us an update on the chronology of when that will be sampling to customers and if the customers agree that it's a great chip and want to put it into their vehicles, when the revenue ramp would occur?

Right.

So in fact, the CV2 is being evaluated, of course, for security camera and auto space, right?

That's if you were on the top of short-term revenue, potential, I think the security camera space is the one we should focus on, and let's talk about that first.

In terms of -- because we already sampled CV22 to security camera customer, they already evaluated the CV22 -- the CM performance, the power consumption like I described in our script, which is a really impressive number and also that we showed so far (inaudible) platform is very similar to how our H2LM type of software platform.

They're very similar so they can leverage their software development easily.

So with all of that, I think that the CV22 has -- we expect there's a good chance they'll get the design in.

And also, we're -- considering CV2 is really a high-end chip than CV22 because we have almost all similar function but the CM performance is 4 to 5x better than CV22.

So for our security camera customer, particular Hikvision, Dahua, they are very aggressive in trying to push out a CV-based security camera.

They want to have a road map on the high end to the mainstream to the low end.

And we are providing really enough road map and the coverage to put in CV22 and CV2 for them to use.

So in terms of -- we are working hard and we expect that Hikvision, Dahua to design those in.

But I think there are 3 drivers.

Like I said, we have performance together with Hikvision, Dahua.

And also, internally, Hikvision, Dahua really have high desire to introduce CV based cost-effective for high performance solution.

The third thing is really interesting is ISC West.

Huawei and HiSilicon come out and claim that they're going to become a systems supplier for security camera in addition to being just a silicon supplier.

So basically -- suddenly that Hikvision, Dahua can view HiSilicon as their competitor.

So I think that the combination with these 3 issues is really helping us to continue to promote CV2 and CV22 to our key customers.

So that's really on the security camera side.

On the auto side, although the revenue is probably still several years away, but I think, like I said, we made progress in multiple different places.

I think, first of all, the customer feedback, like I said, is very positive and also because we really showed that Ambarella solution can address the inadequacy of some -- of the currently deployed system and also, more importantly, how CV2 and CV22 meet the specification that we present to our key customer in the past and also, that we can double CV22 up.

We're going to sample that, CV2.

I will say in early summer, I will definitely try to get out in our July-August time frame so we can sample CV2 and SDK to our key customers.

And also, as I said before, there are 3 buckets of auto customer we're approaching and they're approaching us, too.

One is ADAS.

We sold CV2 and CV22 software as SoC for the computer vision solution.

And for the level 2, level 3 or above for autonomous driving, we have served as a perception subsystem to augment what they have.

And all of them, for all of categories we are engaging with customer in terms of technology as well as business discussion with multiple customer at this point.

Just to add one thing.

Just in the non-CV auto space, we're still on track with our electronic mirrors and our surround view projects.

As you noticed today, we announced an electronic mirror deal with BAIC through one of our Tier 1s in China.

So I think we're on target to continue to expand that business and hit our first half of next year revenue targets that we had previously discussed.

Perfect.

And a couple of quick follow-ups for you, George.

On the gross margin side, I think somewhat similar to the prior question on the OpEx side.

Given all the puts and takes on the mix side of the equation, any color you can provide in the back half of the year there?

And then 2 quick clarifications.

Tax rate for the rest of the year, you talked about a bunch of moving parts in the pro forma tax rate.

Could you just give us what your expectations are there?

And then finally, a clarification on the full year non-GoPro revenue guide.

Did you say plus or minus 3%?

So flat plus or minus?

Or was it 3% plus or minus?

Just wanted to clarify that.

Okay, we'll start with that one and it's flat plus or minus 3% on the $258 million.

So on the tax rate, I think the tax rate should be somewhere between, on average, somewhere around the 12% to 13% range for the year.

We're still looking at all the impacts later in the year.

The gross margin question, I think, is very similar to what we had said in the past.

The higher mix of China security will bring some of the margin down in the second half.

That's why I changed the range to the lower end of my 59% to 62%.

So as we lose the higher margin virtual reality portion and the non-GoPro cameras and China security increases as a percent of revenue, we will see some degradation in the margin compared to what we had previously thought.

Our next question comes from Kevin Cassidy with Stifel.

Maybe just give a little more details on the conversations you've had with the automotive companies.

What are the key features that they're looking for with your device, especially for ADAS that maybe they're not getting from what's in the market today or any other competitors?

Well, first of all, I think that for ADAS function, basically, we talked both about stereo processing as well as monocular.

And the CM performance we showed in the power consumption, for example, I quote that compared to current GPU or ASP solution, we have seen performance with CV22 is multiple times better in terms of processing -- CM processing performance per watt.

I think really that we're providing good -- complete SoC solution with very good quality on the video processing side as well as a very strong CV solution that the people can really improve the CM performance in a big way.

So I think that is something we view that as a strength, our CV2 and CV22.

And on top of CV22, in terms of the performance, CV2 improved the CM performance by another 4 to 5x.

So really I think a lot of our customers view that as a very powerful low power consumption solution and plus, our current video quality.

So all the combination really help people to think that we have a unique offering in the ADAS function.

I think for ADAS, really majority of CV function moving to CM base and that was (inaudible).

But I think that with the very powerful CM performance at a lower power consumption is sort of key.

Okay.

And I think there in the Analyst Day, you had mentioned you'd be selling the SuperCam.

Have you had any revenue from your customers yet or potential customers?

Yes.

We just started shipping those, and obviously, some of those units are going out as development units but some of them are being sold.

So yes, that's starting.

We will have a second version later in the year which takes the SuperCam and incorporates our CV2 chip which will be higher performance.

So I would expect those to continue to be sold through the year.

Again, I don't think the revenue is going to move the needle that much but I think there's going to be quite a few customers who take the product, both the short range and the long-range camera systems for development.

Right.

And it could be interesting for investors just for a leading indicator of future designs perhaps.

Yes.

I'll take that under advisement.

Our next question comes from Quinn Bolton with Needham & Company.

Just wondered, George, on the updated annual outlook, it certainly sounds like the VR and the non-wearable, non-GoPro wearable cameras are down, drones continues to be weak.

But just wanted to ask was there any change in your outlook in either the automotive or IP security relative to prior expectations?

Not a significant change there.

What we noticed just recently is a lot of our VR and non-GoPro sports camera customers started pushing out orders.

So we took it upon ourselves to go out and do a detailed review there and concluded that we don't see the market expanding.

And I think the general feeling there is, is the second half would be down quite a bit from last year.

A lot of that probably being market-driven especially in the VR section which were their new products.

On the security side, I think forecast came in very similar.

And same with the auto, I think there are still uncertainty around what's going on between the U.S. and China.

We just could not make any miracle determination of what the impact was, so we held off on that type of adjustment.

And this is Fermi.

One more comment about security camera.

I really think that the thing -- what move the needle on the security camera revenue is when we successfully introduced product into Hikvision, Dahua.

We expect if that happened, the ASP jump will be significant with a total of high teens of ASP instead of current ASP in mid-teen -- mid-single digit to high single digit.

So we're seeing a significant ASP jump for CV solution being introduced.

Got it.

And just sort of a follow-up on the potential denial of -- or U.S. government not purchasing China-based camera -- security cameras.

If that is enacted, I guess, I'd think about your share in China where you compete pretty heavily with HiSilicon relative to your share outside of China.

Would you say that your share in the non-Chinese OEMs is higher than your share within the Chinese OEMs?

And if the government doesn't buy China-based cameras, could that actually end up being a slight benefit?

I think it potentially could.

We do have a higher share in major customers in North America and Europe.

So I think potentially, there may be an ASP improvement.

It's difficult to say on the units but I think you're right.

There could be potential upside there.

But obviously, it just sounds like until we know what the government enacts it, that it's just -- there's some uncertainty in that segment of the market.

Yes.

I think there's a lot of uncertainty because there's also the potential of China forcing the China customers to use more China source semiconductors.

So there's so many moving parts right now.

And I think the view is, is that things are changing so rapidly that the final outcome and impact is extremely difficult to project.

Great.

And then just a last question, kind of longer term as you look at the ADAS market and as well as the perception subsystems, where you said you were improving the performance of existing systems.

Just kind of curious whether you're seeing customers looking to develop their own CNNs, where they're running on the CV22 or CV2 chips or are you starting to deliver some of your own CNNs along with the hardware to those auto and ADAS suppliers?

In fact, I think that -- absolutely talk to a customer, majority customer, Tier 1 OEMs.

Particularly, the large Tier 1s, they prefer to use their own CNN network.

And our job is to going to help them to port their CNN network in to CV2 or CV22 and deliver the performance.

That's why it's so important that we can port public domain neural network quickly after we got CV22 and [establish] that we can exclude the performance of power consumption number we talked about.

That will give our customer confidence that we can port their neural network that is already available in different platform and put it over and use it there.

I think majority -- in fact, even surprising, I think, even in China, most of customer want to use their own neural network versus using hours, which I think is definitely a trend in the market people spending so much money in investment to develop neural network so that they can differentiate.

I think that's definitely [displayed on my trend].

Our next question comes from Suji Desilva with Roth Capital.

I know there's a lot of uncertainty around the China situation, and it's a difficult question.

But is there any distinction between selling your base video chips versus the upgrade to the computer vision chip into the China IP security market?

Or do you think this is more of a blanket sort of use local semiconductor versus imported semiconductors?

No, I think, you're right.

There's no differentiation between video-only or CV-based chip for this particular problem.

So I think they're just trying to identify whether it's American-based supplier or a Chinese-based supplier.

I think that's where the differentiation is.

Okay.

And then on the Japan auto market, I'm trying to understand how penetrated you are into the surround view and other auto OEM opportunities.

How much further do you have an opportunity in terms of incremental customers?

And what does the pipeline look like?

Is it just a few customers?

Or are you already widely penetrated in Japan?

So let's take it one at a time.

I think for a weak order market, we really are doing very well in both China and Japan and Korea that we talked about in the past, and we continue to having new design wins in this category.

So I think that on the weak order side, we are confident that we are going to continue to perform and are capable of revenue upside from that market.

On the e-mirror side, like we said, in the past, we talked about later this year we're going to start seeing earlier revenue from e-mirror and we just delivered our first e-mirror design from China, which we also believe that's probably most aggressive market in terms of using this e-mirror technology at this point.

And we believe that the current Tier 1, [Ophium], will use this design to go to multiple OEMs and also that we have other design wins with a Tier 1 in China, Japan, Korea and the U.S. that will use this kind of e-mirror design was A9AQ.

So there's a pipeline there but we just cannot give you enough detail about when they're going into production.

But I think the key is that Ophium was the -- is in production as we give you indication A9AQ solution is mature and ready going to production right now.

Yes, I was just going to say, Suji, I think we're still on plan to hopefully have some material revenue in the first half of next year in those product classes.

Okay, great.

One last question for George.

I know this may be difficult but in terms of the seasonality, is the profile changing with the mix of products as you have some headwinds in some of the other categories emerging?

And if you can even answer if the second half of this year looks like a typical seasonal pattern or if it would deviate from, that would be helpful.

Yes.

I think it's going to deviate this year because some of the consumer markets like sports cameras and virtual reality cameras are probably not going to ramp like that we would normally see in a volume market.

I still think we will have some impact from home security cameras but that will be because it's a smaller market.

Probably we will see a flattening in the second half.

Our next question comes from Joe Moore with Morgan Stanley.

Wondering if the non-GoPro business being flat for the year, plus or minus, when you're down kind of high single digits for the first half of the year, can you just kind of -- I understand the headwinds in the back half.

What are the tailwinds that can kind of get you back up to that flat level?

Yes.

It hasn't really changed as security being up on last year and also, in the automotive area we would see both the aftermarket and the OEM business, I think, holding up the numbers to get to breakeven for the year.

That's our target.

Okay.

And then within China surveillance, there's been some focus on custom parts from some of your customers there in both the video processing in the traditional markets as well as AI.

Can you talk a little bit about that?

I know you've been aware of those efforts.

Can you just kind of give us an update on how you sit competitively within those customers?

Right.

So that's video-only because that's become very obvious that, for example, Hikvision has its own, I will say, affiliation to low-end chip.

The company is called [Fu Han], they're doing the low-end chip which is lower end than HiSilicon chip so that I think from video point of view, the Hikvision is using that solution for the lowest end solution for now.

And -- but nonetheless, it's definitely an indication that our big Chinese customer has the ambition to come to -- to become a silicon supplier.

So I think that for us, sometimes we still have a great relationship with Hikvision in terms of talking about the potential partnership in terms of a CV and a video-only solution.

We have a very open dialogue in terms of what they are trying to do and what they can -- what's our position in their supply chain.

So I think that Hikvision definitely trying to prepare some low-end solution just in case HiSilicon become their competitors because Huawei is coming to compete also in that.

It's definitely a worry for us.

But moving forward, we think that with our CV2 and CV22, we have a chance there to become -- come here to be main partners or one of the main partners in their silicon supplier.

Our next question comes from Charlie Anderson with Dougherty & Company.

Fermi, you mentioned in your script about the legislation in China that could maybe help the dash camera market.

I think at Analyst Day, you talked about the SAM there being about 20 million units and then OEM unit 6 million, 7 million.

So if that legislation passes in China, can you maybe help us understand how much larger that market can potentially get as a result of that?

Right.

So first of all, although we are not directly involved in the legislation but I think you can imagine that a lot of autonomous in China are involving in the government legislation trying to finalize in the near future.

And based on the legislation, basically, you're going through -- if that pass, mandate cars have a record system for different purpose, for the insurance, for the accident or any kind of arbitration on the court cases.

So I think, among other reasons, China -- we believe China will pass the legislation based on the current momentum, and the impact is huge.

I think basically that means all the 20 million to 25 million new cars every year in China need to have this and also that basically will push the whole after market transition into the OEM market and -- which is important for us because for OEM market, the demand on the quality, I mean, silicon quality AEC-Q, wide temperature range, they are all our strength that we are being -- we are a proven supplier on that in China.

So definitely that help us to get to a 20 million to 25 million units market which is much bigger than the current service market that we have in China.

At the same time, give us a chance to continue to compete with our advantages.

At the same time, I want to also mention that there is a trend that we are seeing that more and more people are starting to talk about autonomous car.

There will be more and more people start talking about auto OEM, start talking about they need to record video just because there's many accidents happens with ADAS-based or autonomous car happening in the field and the people want to know exactly what happened.

So a [light] box type approach is definitely being discussed.

And I think the trend we are seeing is that more and more people will consider maybe even legislation in different country, we'll start talking about putting light box into those autonomous car.

So I think that trend is definitely something we're watching.

Hopefully, that will help us to increase our total available TAM in the future.

Great.

That sounds good.

So second question for me is it sounds like on the auto side, you got a lot of evaluation tips out there, a lot of good customer response.

I wonder, Fermi, just how you're defining success this year in that endeavor.

What are your goals for this year in terms of what you want to accomplish as far as any progress with customers is concerned.

So in fact -- in terms of auto side, I think there are several important internal milestones we have set up together, right, we look at the design we can secure.

We are looking at having some customer announce with us.

We are looking at continuing to increase the design win in the recorders and any type of solutions and also trying to get some design win in ADAS in level 2, level 3 cars.

Those kind of things are major milestones that we're looking at as evaluation, how well we are doing internally in terms of automotive market.

Our next question is from Ross Seymore with Deutsche Bank.

George, I just wanted to get back to the seasonality question that you mentioned before, not to parse it out in too fine a way.

But when you talked about a flattening out, it looks like the implied second half of the year has to be up somewhere in the kind of non-GoPro $140 million range given your guidance in the first half of the year to get you to the flat midpoint.

So when you talked about the seasonality, were you more talking about the third quarter will still probably go up more and the second quarter down but not as much as in the past?

Any sort of just shaping color in the near term.

And then as we think probably equally important, as we go forward, does seasonality kind of just go out the door as the design win you guys have in places like the automotive market and then, obviously, continuation in the IP security really doesn't have as much of a seasonal pattern as your former business did?

I think the answer to the first part of your question is yes.

I think what we're -- I didn't mean that would be flat on Q2.

I just felt the -- right now, you'll see the increase in Q3 and 4 be less probably than we've seen in the past.

Just a function of the lower consumer portion of that business.

Going forward, I would expect that to continue to be less of an impact as we get more into auto OEMs and that becomes a bigger part of our business.

There is some seasonality to security.

The second half in professional security and consumer security is normally higher than the first half.

That's been part of seasonality, and I think that part will continue.

But I think the step-up that we used to see with the heavy concentration of consumer products is probably going to come down from previous years.

I'm not showing any further questions at this time.

I would now like to turn the call back over to Dr. Fermi Wang for any further remarks.

Yes.

I would like to thank everyone for joining us today, and a very special thanks to all our employees for their continued dedication and hard work.

Thank you, and goodbye.

Ladies and gentlemen, thank you for participating in today's conference this does conclude today's program and you may all disconnect.

Everyone, have a great day.