Ambarella Inc (AMBA) 2019 Q3 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the Ambarella Third 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, Mr. Louis Gerhardy, Director of Corporate Development and Investor Relations.

Mr. Gerhardy, you may begin.

Thank you, Sonia.

Good afternoon, and welcome to Ambarella's Third Fiscal Quarter 2019 Financial Results Conference Call.

Thank you for joining us today.

Our speakers will be Dr. Fermi Wang, President and CEO; and Casey Eichler, CFO.

The primary purpose of today's call is to provide you with information regarding our fiscal 2019 third 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 2018 fiscal year, and the Form 10-Q filed on June 8, 2018, and September 7, 2018, for the fiscal year ending January 31, 2019.

Access to our third quarter results press release, historical results, SEC filings, and a replay of today's call can be found on Investor Relations portion of our website.

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

Thank you, Louis.

Good afternoon, everyone.

Our Q3 fiscal year 2019 revenue of $57.3 million was down 8% from the prior quarter.

The surveillance increased 9% sequentially and the automotive declined sequentially as anticipated.

Revenue from consumer and other applications was down more than 40% sequentially and represents less than 15% of the total revenue.

Our resource allocation demand is heavily focused on computer vision opportunities in surveillance, automotive and other computer vision-related applications.

In the surveillance camera market, we see strong interest for single-chip solutions with integrated computer vision processing to support the latest deep neural network-based algorithms.

Ambarella's SoC solutions based on our CVflow architecture offer industry-leading CV performance with extremely low power consumption, combined with advanced image processing capabilities required to operate in very low light and high contrast scenarios.

Additionally, the CVflow family addresses the need for on-chip highway security features to avoid hacking and other forms of network attacks.

Ambarella's CV22 and CVflow SoCs have Tier 1 designs or under evaluations at each of the leading top 10 professional security camera suppliers.

We now expect to make our first production volume shipment by the end of our fiscal fourth quarter this year, with customer product introductions expected mid-next year.

In October, Ambarella participated in 2018 Security China Exhibition on Public Safety and Security in Beijing.

At the event, we demonstrated our new CV2 and CV22 SoCs running multiple neural networks as well as in multi sensor and stereo applications.

We also demonstrated home monitoring cameras, including solutions addressing the fast-growing battery-powered and doorbell camera segments with third-party partner demonstrations of features including intelligent person detection and face recognition.

We are seeing non-China security camera makers increasingly reluctant to design in SoCs from our main competitor HiSilicon, which is part of the large Chinese network company Huawei.

Huawei is banned by the U.S. government from supplying network equipment for government contracts, and the camera makers are therefore, looking to find alternatives to HiSilicon for cameras to be sold into U.S. and for government projects.

Together with our technical advantages, this has helped us win new designs and major customers that had previously used HiSilicon, providing opportunities to expand our market share for both our existing video SoC as well as our new CVflow chip families outside China.

In consumer IoT markets, we are seeing cameras increasingly being offered as a part of the home automation ecosystems, including a broader set of smart products and services.

In Japan during October, Sony announced it entered the smart home business with the introduction of its MANOMA service.

MANOMA includes security, automation and lifestyle services and includes a home gateway indoor communication cameras, security sensors and smart locks.

The camera, based on Ambarella's S2L SoC, also integrates Amazon Alexa audio systems for voice control of home appliances.

In China, Xiaomi introduced its latest Mijia smartphone camera based on Ambarella's S3LM HEVC camera SoC.

The full HD camera supports 360-degree operations pan, tilt and zoom operations includes Xiaomi's AI assistant to enable device command to take pictures, record video message or control other smart home products.

In Europe, Hive, a brand of UK service company Centrica, introduced its Hive View Outdoor camera, part of a portfolio of connected products also including connecting sensors, thermostats and lighting.

Based on Ambarella's S2E SoC, the camera features night vision, HD live streaming, person detection, smartphone notifications, and 2-way audio.

And in the U.S., home security service provider SimpliSafe further extends its product line with the introduction of Video Doorbell Pro, based on Ambarella's S2LM SoC, the Doorbell includes wide angle full-HD video, HDR imaging, motion alerts, and a 2-way audio.

In automotive markets, we continue to experience strong interest and designing activity for both our human vision and the computer vision solutions.

In particular, our CVflow SoCs are being evaluated and designed by traditional Tier 1 and OEM customers as well as Auto 2.0 companies for ADAS applications and for vehicle offering higher levels of autonomy.

We now have our first design win at a Tier 1 for front ADAS cameras, with the customer choosing our solution because of its ability to deliver cost-effective low-power implementation with advanced AI features.

Our CV22 SOC addresses and exceeds in the performance specifications of a monocular Euro New Car Assessment Program, or NCAP, ADAS cameras.

While it's low-power consumption allows its use in the small form factors of windshield-mounted camera designs, the size and thermal limits of this camera typically preclude the use of high-powered GPU-based solutions.

Customer benchmarks also confirm that our performance is significantly higher than competing solutions from legacy SoC chip suppliers, enabling customers to offer Ambarella-based cameras with more advanced ADAS features.

In vehicles addressing autonomous Levels 3 and 4, cameras need to capture full 360-degree video.

Ambarella has previously demonstrated its capability in its EVA, or Embedded Autonomy Vehicle, using multiple short and long-range stereo cameras based on its CVflow processors.

Our new CV2 CVflow processor offers a combination of extremely high computer vision processing performance, advanced image processing and stereo processing.

CV2 is currently in design with a leading autonomous vehicle OEM to enable multiple short-range stereo cameras in the level 4 design.

In this application, the CV2's stereo processing provides the necessary depth information for the detection and the classification of objects along the vehicle.

Its ability to work in challenge like conditions enable it to be used in vision-based assistance and offer greater resolution, accuracy and lower costs than alternative technologies.

Our human vision devices also continue to see strong design activity in both drive recorders and electronic mirrors, with this product stealing the market and allowing customers to extend their product lines to add ADAS features.

During the quarter, ZF a global leader in driveline and chassis technology as well as active in the passive safeties technology, held its Vision Zero event in Michigan.

The event focused on presenting ZF's innovative solution to automotive OEM customers and included electronic mirror and the surround-view monitoring system, based on Ambarella A9AQ SoC.

The demonstrations highlighted excellent image quality under various weather conditions and LED flicker mitigation, both enabled by A9AQ powerful image processing.

In electronic mirror applications, we now have design wins with 2 leading Asian Tier 1s for our CV22 SoCs, which support both human viewing as well as computer vision based blind spot detection.

Vision based blind spot detection increases the accuracy of the detection and potentially lower costs by eliminating the need for dedicated radars on each side of the car.

The CV22 wins are secured as a result of our traditional strength in HDR, lowlight performance and low latency as well as the ability of our CVflow processor to advanced run ADAS algorithms.

In China, we are continuing to win new designs for car recorders as OEM manufacturers offer this as preinstalled options.

During the quarter, Chinese car OEMs BYD, Dongfeng and Changan all introduced new car and SUV models with car recorders based on Ambarella's A12A camera SoCs.

Features include full HD resolution, Ethernet connectivity and Lane Departure Warning systems.

The need to run complex AI or computer vision algorithm to monitor drivers, passengers and the entire cabin of vehicle also represents an excellent new opportunity for Ambarella.

During the quarter, China-based Roadefend introduced combination car recorder and the driver monitor cameras based on Ambarella's quad-core S5L SoCs.

Targeting commercial vehicles, the aftermarket cameras include ADAS features such as floor collision and Lane Departure Warning, as well as drive monitor features including distracted driver, driver fatigue and cell phone use warnings.

In summary, our traditional low-power and high-resolution human vision solutions continue to exhibit strong momentum in our focus markets: surveillance and automotive.

Our existing presence and the credibility in this market are helping seed the market for our new computer vision products.

The CV product come in significantly higher average selling prices than our traditional devices, while our automotive CV application also offer lifetime revenue opportunities substantially higher than Ambarella has experienced in the past.

We are encouraged by the numerous design wins we have secured during the last quarter and remain excited about the future growth opportunities in automotive, surveillance and other computer vision applications.

We hope to see you at CES in January, where we will be demonstrating our latest computer-vision solutions, including new product introductions for both surveillance and automotive markets.

With that, I'll transition to Casey for the financial report.

Thank you, Fermi, and good afternoon, everyone.

Today, I'll review the financial highlights for the third quarter of fiscal 2019, ended October 31, 2018, as well as financial outlook for Q4.

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, we have eliminated stock-based compensation expense adjusted for the impact of taxes.

In line with our previous guidance, our Q3 revenue totaled $57.3 million, which represents a decrease of 8% from Q2 and 36% when compared to the same quarter of the prior year.

GoPro revenue continued to be immaterial in the current quarter.

In Q3, we saw growth in both professional and consumer surveillance markets when compared to the prior quarter.

The increase in surveillance was led by our consumer security, although professional markets also increased.

Auto revenue declined when compared to Q2 as a result of some customer promotional programs ending, as we discussed last quarter and the overall tone of the auto market, which was more cautious than expected.

Consumer and other product revenue continued to substantially decline during the quarter.

Non-GAAP gross margin for Q3 was $60.9 million compared to 61.4 -- excuse me, 60.9% compared to 61.4% in the preceding quarter.

Gross margins primarily reflect a shift in market mix to lower margin surveillance revenue in China and a drop in higher-margin consumer revenue.

Non-GAAP operating expenses for the third quarter were $28.7 million compared to $29.9 million for the previous quarter.

OpEx decreased when compared to the prior quarter, primarily due to decrease in employee related expenses and general cost control as a result of lower revenue.

Non-GAAP net income for Q3 was $7 million or $0.21 per share compared to non-GAAP net income of $8.5 million or $0.25 per share in Q2.

The non-GAAP effective tax rate in Q3 was approximately 3.2%.

The non-GAAP effective tax rate for fiscal year 2019 includes a change in the allocation of stock-based compensation across the company's tax jurisdictions to improve alignment of the non-GAAP tax rate to the GAAP tax rate.

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

In the third quarter, the non-GAAP earnings per share were based on 32.9 million shares as compared to 34.2 million shares in the prior quarter.

Total headcount at the end of the third quarter was 747 compared to 741 at the end of the previous quarter, with about 82% of our employees dedicated to engineering.

Approximately, 71% of our total headcount is located in Asia.

We generated operating cash flow of $3.5 million, and we ended Q3 with cash and marketable securities of $349 million, compared to $376 million at the end of Q2.

During the quarter, the company repurchased approximately 825,000 shares at an average price of $37.31 for total consideration of approximately $30.8 million.

As of October 31, 2018, approximately $35.4 million remained available for repurchase under the $100 million repurchase program authorized through June 4, 2019.

Total accounts receivable at the end of Q3 was $32.4 million or 51 days sales outstanding.

This compares to accounts receivable of $29 million or 43 days at the end of the prior quarter.

Net inventory at the end of the third quarter was $23.3 million compared to $30.8 million at the end of the previous quarter.

Q3 days of inventory increased to 109 days in Q3 from 100 days in Q2.

As discussed last quarter, inventory was higher-than-normal and was reduced in the current quarter to a more appropriate level.

We had 2 10% plus revenue customers in Q3: WT Microelectronics, a fulfillment partner in Taiwan, came in at 49% of revenue; and Chicony, a Taiwanese ODM, who manufactures for multiple customers, came in at 23%.

I will now discuss the outlook for the fourth quarter.

We expect total revenue for the fourth quarter ending January 31, 2019, to be $51 million, plus or minus 3%.

In the quarter, we anticipate revenue will continue to be negatively impacted by the decline in demand for consumer and other products.

Macroeconomic factors are also making it somewhat more challenging to forecast.

Global tariff issues, like the potential expansion of products subject to tariff, potential increase in the U.S. tariff to 25% at the beginning of the year and the potential for new export restrictions are issues that we will continue to monitor during the quarter.

In addition, slower-than-expected light vehicle production and some deleveraging in China could also be a factor.

We estimate Q4 non-GAAP gross margins to be between 59% and 60.5% compared to 60.9% in Q3.

Change in revenue mix will continue to have a slightly negative impact to margins in the quarter.

We expect non-GAAP OpEx in the fourth quarter to be between $29 million and $31 million, with an increase from Q3 primarily coming from increased engineering headcount.

The annual tax rate should be modeled at 7%.

We estimate our diluted shares for Q4 to be approximately 32.8 million.

As previously discussed, the substantial weakness in our consumer electronics business, particularly sports cameras, along with the decline in consumer drones, is expected to continue through the remainder of the year and into fiscal 2020.

We believe our revenue from the automotive market will be relatively flat sequentially in Q4.

As a result of the U.S. ban on the purchase of surveillance cameras for government facilities from Hikvision and Dahua, we are continuing to see some near-term reduction in their orders, particularly in the high end of the product range normally associated with their export business.

We are starting to see an offset increase in orders from non-China customers.

As it relates to other trade disputes, we have not currently been impacted by any tariffs that have been put in place by non-U.

S. governments but continue to monitor the situation.

Other impacts from the trade dispute remain at risk, such as potential inventory bills in anticipation of an increase in the tariff rate.

But at this time, we have not seen this affect our business.

In addition, we have not seen the dispute impact the development effort of new surveillance cameras based on our CB chips at our China customers.

We will continue to control OpEx in Q4 to reflect lower revenue expectation in the quarter.

We continue to focus our headcount resources on CB product development for the surveillance, automotive and other CB markets, thereby decreasing the need for additional headcount.

We hope to see you at CES January 8 to 11.

Please contact Louis to reserve a time to meet and see our demos.

Thank you for joining us today, and with that, I will turn the call over to the operator for questions.

(Operator Instructions) Our first question comes from Kevin Cassidy of Stifel.

Congratulations on the design wins for the CB chips.

I just wonder if you could give us a little more detail on the design wins, say, first on the automotive or the ADAS wins.

Is that -- to an OEM, is it North America based?

Or maybe, just give a little more description and say who the company is?

First of all, I think that the Tier 1 that we're working with is -- has a very strong active safety, presence as well as they have very strong ADAS background.

And what we have established is demoing our product, our CV22 product, and helping them to import their algorithm on our platform and -- so that they can demo this platform to the several OEMs and also, this Tier 1 has an established relationship with a number of global auto OEMs, and we saw -- and also with those OEMs, they are demoing the solution at this point.

This is a non-Chinese Tier 1 that we're talking about, and we hope that we can continue to offer you more updates in the near future.

But let me say one more thing.

I think the reason that CV22 is getting this design win is because our performance on CV and also that we are running all those algorithms in a power consumption around 2 watts, and more important -- and also as important that we have been using our video processing technology, to help to get better CV result.

I think the combination of this is a reason we won the design.

Okay, great.

And maybe, similar type of detail on the surveillance camera.

You say first going, one design goes into production at the end of this quarter, fourth quarter.

Can you give us an idea of what the ramp looks like?

Does it accelerate through the year or...

First of all, I think 's that the 1 customer can start taking shipment in this -- Q4 of this year, and we believe they will -- they told us they are going to start shipping probably mid next year and they are going to be a leading customer.

This is a non-Chinese customers, and we expect that the balance, the rest of other professional surveillance customers is going to gradually ramp up first half of next year, and I hope that most of them will be in production by the end of next year.

And our next question comes from Quinn Bolton of Needham & Company.

Was -- hoping to follow-up on Kevin's question around sort of the timing.

It sounds like the Tier 1 OEMs you have for the ADAS system is more of a parts -- Tier 1 parts manufacturer that has relationships with multiple Tier 1 OEMs, to the extent they are successful in getting their system designed into the automotive OEMs.

When do you think you would see revenue in the model productions?

Is that still a, kind, of a 2-plus year window or could it happen sooner?

And then, I've got a follow-up.

Well, so this particular I think, we're looking at 2021, 2022 OEM basically we're looking at those kind of design wins opportunity with OEMs in those -- in that kind of time frames.

So I think that we are talking about at least 2 years out from ADAS revenue point of view.

Okay.

And the follow-up question was, I think you mentioned that it's the customer software that they imported to the CV22 processor, but just wanted to confirm that your relationships with a lot of the ADAS and level 4 wins that you mentioned, it's going to be running customer developed models on your CVflow architecture.

Is that right?

On the -- the level 4 design win we talked about is using stereo solution to do a short-range object detection and classification and those software will belong to us.

So we are providing a software and hardware in this particular design.

And our next question comes from Ross Seymore of Deutsche Bank.

Just wanted to follow-up on the auto side as well and maybe, even from a higher level on the CV side.

Fermi, as you're seeing the chronology of how CV plays out can you just walk us through it doesn't have to be quarterly but maybe, even on an annual basis.

How do you think the CV benefits will start to flow through the income statement at your company, either by end segments or kind of the ASPs versus the unit side of the equation?

Any color you can give there would be helpful.

Right.

First of all, I think our CV revenue on the surveillance camera, now I'm pretty confident that we're going to start seeing revenue next year.

I think in the fiscal year '21 or kind of the year 2020, we should see a material revenue from those professionals surveillance cameras.

For the consumer surveillance camera, I think that will lag behind it, maybe by 6 to 12 months because the design win just started, and we probably will give you more updates later on.

On the auto side, in fact, we talk about 3 new design wins, in fact, 4 design wins in 3 different auto markets.

We talk about when design win in ADAS, one design in level 4, and 2 design wins in e-mirrors for CV.

And I believe those e-mirror application has a chance to get into the revenue earlier than the other 2 but still, we are looking at 2 to 3 years out from that point of view.

So from a revenue point of view, that's how the CV revenue will work out in in my opinion.

However, let me add one more thing we didn't talk about is there's a lot of CV opportunity in China.

While it's also applied to ADAS Level 3, level 4 as well as e-mirrors, I -- while we haven't talked about anything in terms of design win in China, but I believe that the revenue opportunities in China is earlier -- somehow earlier than the outside China.

So does that comment, the last time that you just said, assuming you sort of resolution of the tariff side, I would assume that the Chinese customers that are staying within China don't really care about the tariffs, but does the adoption of CV into the IP security get impacted in any way due to the macro and kind of governmental issues right now?

I think our customer in China are watching this progress as much as we do.

However, I don't -- we still continue to have a design win engagement and discussion.

So while I do believe they all worry about the outcome of those situation, but I think at this point, the project is still ongoing and moving forward.

And then one final quick question for Casey on the OpEx side.

I know you mentioned that you're keep running it tightly, and you came in below your guidance in the quarter and below expectation for the guide.

How are you guys thinking about the OpEx side as we look into next fiscal year?

I know you're not going to guide the full year, but are you expecting to keep a tight?

Or is the opportunity so large in CV that you would expect to continue to add headcount in strategic areas?

Yes, I think from an OpEx standpoint, obviously, we've got sales and marketing G&A in there as well engineering.

And from the sales and marketing G&A perspective, I think we're going to try to keep it pretty tight and be pretty -- as prudent as we can in those areas.

In R&D, we're not going to let the opportunity get away from us.

So we're going to continue to invest and more resources over from our traditional products into the CV and also continue to hire.

So as we see more and more interest like we've seen this quarter, I think you're going to see us make sure that we have the right amount of resource applied to that so that we don't let the opportunity get away from us.

Because as you know, you have to make the investments today for the revenue, as Fermi's been talking about, 2, 3 and 4 years out.

And our next question comes from Adam Gonzalez of Bank of America, Merrill Lynch.

Congrats on the first CV design win in autos.

Just wondering, Mobileye has become a lot more vocal recently about opening up their platform, perhaps allowing customers to port their own algorithms over onto their platform, especially with the next EyeQ chip.

Just wondering what your thoughts are and what this -- how this could impact your long-term ability to gain share and penetrate the market?

Right.

First of all, I think what Mobileye is really focused on EyeQ 5. So EyeQ 3, EyeQ 4, which are mainly designed for ADAS are not opened up for any Tier 1 OEMs.

And the EyeQ 5 is pretty much focused on, in my opinion, Level 3 or above autonomous car.

So that is definitely something we need to continue to watch.

But from ADAS markets, I think that their strategy doesn't change.

While the -- I really think that by opening up a platform versus a closed platform, the support structure and the software requirement are quite different.

And while they ask you [to play], we'll continue watch how they ask you to it.

The -- when we try to open up the platform, the amount of resources we threw out to try to make sure that our customer can program easily and making sure that they can -- that they have enough visibility on our hardware and software so that they can do differentiation.

Not only the efforts is nontrivial, but at a certain point, you have to disclose not only just your software structure, you have the hardware structure.

And I would like to see how Mobileye is going to do that.

Got it.

And just a follow-up.

If I were look at your business over the next, say, 4 quarters or so, I think prior to this quarter, you had 15% or 20% of your business still in legacy consumer applications.

If I were to fast forward, what percentage of revenues do you think that will be at the end of next year?

Yes, we're not likely going to guide out, but I think it'll definitely be lower.

That market will continue to decline for us.

We're putting any of our investment resources in it and any of our focus in it.

So I think the tail on that can go for several years, but it definitely will be declining over that period of time.

And our next question comes from Joe Moore of Morgan Stanley.

You talked about the ADAS win on the car side.

Can you talk about autonomous?

And I guess, I know it's going to take a lot longer to implement those solutions, but what's the time frame for potential design wins kind of on the more autonomous side of equation?

So for that particular design, we talk about our CV2 chip is used only for the short-range stereo.

I do believe that for autonomous driving, there is -- there are multiple cameras being used, and right now, people are trying to using camera to solve different problems.

And I think that in this particular design, CV2 proved that they have a unique position to solve this particular problem.

So -- and the revenue we are talking about is also 2022, in that time frame when that level car is ready to go into production.

So I think, for any Level 3, Level 4 car, that, it will take a while to getting there, and we are patient.

And in fact, we're going to continue develop technology not only on long-range to short-range stereo but also on the chip side.

We knew that we need to continue to evolve and produce a more powerful and [more better] chip for that kind of a market.

Okay, great.

And in terms of the trade potential issues, there's been, I guess, some conversation about AI-specific limitations of Chinese customers to procure AI-centric product.

Do you -- and I don't think -- there's not a lot of detail provided on that, but do you guys have a sense for whether that could be an issue down the road or just if you could fit under that umbrella?

Right.

So first of all, the list is very vague.

But Joe, if you look at the list, in 12 items, AI is one of them.

If somebody asks you to that 12 items, basically, the whole semiconductor industry, probably majority of them will be barred from selling to China.

So I would like to see how that thing develops and how that's going to impact us.

But from an AI-particular point of view, I want to point out that we don't see our AI algorithm particularly.

For example, when we sell our AI chip, we're selling an inference engine.

We don't do training, but we don't sell our algorithms to our customers.

So we really need to find out how the regulation to participate in different AI functions and markets, and that all the details will impact us, whether that would be a part that limitation.

And our next question comes from Suji Desilva of Roth Capital.

I'll echo my congratulations on the design win here.

In the auto market, you have 4 design wins.

You have the CV win right now.

What's the right pace of additional wins we should be thinking about in the next 12 to 24 months?

Does it take a lot of effort for the first one or a lot of them in a waterfall on the back?

I just want to level set the expectation here.

I think we are talking about the design wins that we are allowed to talk about.

We definitely are working on more design win potentials.

And I definitely, if -- as soon as there are more that we can talk about, we'll continue to provide you updates.

Okay, that's fair.

And just real quick follow-up there, where -- how does the headcount scale as you add on more engagements?

Do you have to add headcount for each of these guys?

Or do you scale the headcount you have now across multiple efforts?

Well, first of all, as Casey just said, we are aggressively moving resource from our traditional video product line to CV product line.

So when we scale customer, we probably going to continue to use our resource on the video side to support the opportunities.

And gradually, that will be one way to staff up.

Of course, that, if you don't have enough people, we're going to continue to hire to staff the new opportunities.

But I think at this point, our current resources plus a little bit more will be enough to support the opportunity we are seeing right now.

That's helpful.

And if I could just make one last question.

Can you just update us on the content per car for things like e-mirror versus ADAS versus L4 just so we can get some numbers as to dollars per auto we can think about?

Yes, so every OEM has -- they all have ideas.

So let me be clear, I think for e-mirror, it for us is always one single chip to design.

So for all of the -- cover 3 mirrors.

So it's one powerful chip to do all the video processing plus the blind spot detection.

We haven't -- the ASP, I would say, is in the CV22 range, in the range of $15, $20, in that range.

And for the ADAS, because there's an [ASO] chip, the ASP were a little higher, but that's going to be also -- I think that's going to be the biggest market in 2021, 2022 for us.

And people talking about 15 million units car will use ADAS technology at that point, and that is a single-chip solution.

For PMS and CMS for [company] monitoring, it's also single chip with multiple camera input, and we handle all the video processing and the CV application.

I think the autonomous car for Level 3 or above is probably the most difficult to answer because every OEM or Tier 1 have their own idea.

But everybody's talking about 10, 12 cameras.

Some of them try to use [quite a huge] chip.

Some people try to use multiple smaller chip to solve the problem.

But I think that's -- definitely marks the higher ASP opportunity for semiconductor companies.

And our next question comes from Charlie Anderson of Dougherty & Company.

My congrats on the CV wins as well.

I wanted to focus on Level 3, Level 4 and the win there.

I think you said it was an autonomous vehicle, and we have some maybe nontraditional carmaker or someone who's more of a startup.

I wonder if you could give us a little more detail there and maybe the region.

And then beyond that, it sounded like they're using it for short range.

Are they adding a sensor to what's sort of the traditional regular multiple cameras plus radar and lidar?

Or are knocking out one of the existing sensors we would be used to there?

And then I've got a follow-up.

Right, so for this particular design win, it's a non-Chinese design win.

I think that the usage model is very clear because we are in a very congested traffic condition.

It will need to know that, that information among the -- for all the cars around you.

And cars use lidar to address the issue.

And that means you to put multiple lidar around the car, become very expensive proposition.

So then when we demo and show our capability as short-range stereo, even at the very challenging light condition, we show people that we can correctly identify objects as well as determine the depth among all the cars around the car you are driving.

So I think that really show people we have a cost-effective and easy enough solution to solve the problem.

So you can imagine that you have a short-range stereo camera around your car, multiple pairs of them.

In this case, there's 4 pair of them.

And those 4 pairs, information go to one CV2 and would process all of the information, including the stereo processing as well as the classification that require for to perform this function.

Great.

And then, for a follow-up.

It was great to hear about the potential take some market share with HiSilicon.

So I wonder if maybe you could quantify to any degree in terms of from a total market share of the industry or unit perspective.

Any additional color there would be helpful.

It's hard for us to give you a number as mentioned, but I can give you high-level assessment.

In the past, outside China, although we are market leader, HiSilicon still have design wins because every customer wants to use to 2 source -- 2 supplier as well as that HiSilicon always try to offer a much cheaper solution than us.

And there are some customers that are using them as the first supplier, and we are second.

But I think recently, after we demo CV and after the current trade war situation, I think that we start seeing a lot of design wins moving to us, not only from the CV side but also on the video product side that used to use HiSilicon.

So I would imagine that our market share outside China will increase.

And also, that Hikvision, Dahua because we got banned from the U.S. government, most new opportunity available to non-Chinese surveillance camera system vendors, I think that a portion of that will become ours too.

And our next question comes from Matt Ramsay of Cowen.

Fermi, I think the team and Les have talked for a while about a number of angles of differentiation on your CV road map.

You mentioned power envelope around 2 watts.

The fact that you do stereo, the fact that your video is human readable, that you are doing -- or human viewable that you're doing the analytics on and then some of the video stabilization and things that you have learned from the action camera market, maybe you can talk about like where your real -- do you feel like the core of the differentiation is.

And what has been the competitive response that you've seen from other folks with similar architectures?

Right.

So in fact, among all of the advantage you talked about, every different application enjoy different advantage from us.

For example, for ADAS camera, this front-facing ADAS camera, the most important features are, one, because of limited form factor, your power consumption had to be small; two, with the limited power budget, you have to deliver as much CV performance as possible.

That really plays to our strength.

In the Level 3, Level 4 design win we just mentioned, power is important but most important feature enable us to win the design is our stereo processing and our video processing capability like process a video at low light and associate it with the stereo processing to make us win this design win.

In the e-mirror design, it's is almost like combination of that because the power consumption is a concern, CV performance is a concern and also the video processing because now with CV processing, you also still need to continue to display the video while you can the processing.

So I think that the market require all 3 of them.

So for a different market, they have different requirements.

And fortunately, we have multiple advantages that can help different customer from different markets for their requirements.

No, thanks, that's really helpful.

Maybe, Casey, a question for you, the -- we talked about over the -- all the Q&A on the call here about some of the really exciting long-term CV opportunities in the automotive space.

But you guys have a number of design wins for data-recording cameras and 360 surround view and even some e-mirror projects that are not CV-based, that are maybe nearer term in revenue.

Can you maybe, I don't know, in the next 6 to 8 quarters, while we're waiting for some of the CV stuff to hit the P&L, can you maybe talk us through how you're thinking about the automotive business over that time frame?

So Matt, this is Fermi.

I think that in next 6 quarters, majority of our auto revenue will come from recorders because they've become so big and so quickly.

Both, we see a huge opportunity in China.

We start seeing big and big.

And also, we start seeing Korea having this kind of demand.

So in China, we mentioned that Chinese government is putting a regulation together for data -- for the recorders.

And in Japan, we see opportunity.

They're called data loggers.

So I think that because we have been working on the recorder business for 4 years with OEMs and Tier 1s, this is going to be continue to be dominant auto revenue for the next 4 to 6 quarters.

And we believe we are going to continue to see revenue from the e-mirrors, AVMs, but they will be much less than the recorders in the near term.

But I do believe that when e-mirror's opportunity, CV will happens, the market size can grow quite dramatically, then that's what enable the opportunity for us.

But there's one thing I want to point out.

We believe in the short term, the CMS -- the PMS and CMS market is real, and it can be a short-term revenue for us.

We are just start seeing it, and we're putting a product together.

We are demoing some there at CES.

So please join us at CES.

We'll give you complete road map display at the time.

Yes.

The only thing I would add to that is, clearly, there is a lot of interest around the auto market and the surveillance market, and in particular, a lot of emphasis has been in the promise in the long term over the CV product line.

That isn't to deemphasize the fact that we have a lot of products in the marketplace today and a lot of demand for that, that's going to continue.

So to your point, we will continue, I think, to see good activity on our traditional products and our traditional product lines.

But clearly, the focus for all of you and the focus for us is to continue to demonstrate the promise of the CV architecture that we've talked about.

So that's just all I would add to it.

And our next question comes from Richard Shannon of Craig-Hallum.

I'll add congratulations on top of everyone else's.

Quick question on professional security.

Obviously, you started -- you talked about first shipments of CV into the -- into that market with one customer.

I think you mentioned a few others coming later this year.

How should we think about that professional security market in the context of that?

Is there any substitution effect of other design wins you've been in with these customers?

And do you think that market is definitely a grower next year even with the geopolitical and tariff issues that could be lurking out there?

Yes.

So first of all, let me qualify that all my -- what I have said about surveillance market is assuming that there is no more tariff or other impact because I cannot forecast it.

So I am assuming that the business will continue running under the current situation.

That's the first one.

Second one, I think that what we -- what I just said is we're going to have one customer taking volume shipment in this -- in Q4 this year.

But I expect majority while other professional surveillance customer while going into production by the end of next year, not this year.

I just want to be clear on that.

And we are seeing design wins with almost all of the top 10 customers.

And like I said in our script, that we have design wins or under evaluation with each one of that at this point.

And we are continue having design wins in the last 3 months, and I hope that will continue.

With that momentum, I'm confident that we will continue to be a major player in this market.

In terms of size of market, first of all, I still think professional surveillance is a growing market, but the growth rate is definitely not as high as before.

In fact, Hikivision and Dahua came out to say they expect their growth to start slowing down.

However, my belief is this: although while is the growth rate on the whole market might not be as high as before, but I think the CV technology will trigger a new wave of a refreshing cycle because of the power -- the performance and the functions that are enabled by CV really will trigger everybody to quickly adopt this at the right pricing.

So I think that throughout the years, I think the CV function will trigger the whole surveillance market goes to the refreshing cycle just like when the IP security camera comes in and create a hugely refreshing cycle versus, again -- against the traditional analog cameras.

Okay, great.

Great for all that detail.

Fermi, one follow-up question on the automotive CV win in ADAS.

I guess, it wasn't clear to me based on your commentary and answers to the questions whether this design is with an automaker itself or a Tier 1 OEM supplying that.

And so if it's the latter, is there a design specifically with an automaker that you are predicting revenue starting in 2021 or 2022?

We are.

Look, the design win was a Tier 1, but we know that they are presenting a solution to multiple OEMs.

Most of the OEM are -- is around 2021, 2022 at this point.

Okay.

Do you have an expectation of timing of when that customer would have a win with an automaker then?

I hope it's as soon as possible.

I'll try to offer you more updates in the near future.

And ladies and gentlemen, this does conclude our question-and-answer session.

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

Thank you.

And in summary, I think we are very happy with the progress we have made with our CV products in our target markets.

I'd really like to thank all of our colleagues to make this happen.

And thank you for joining us today.

I'll 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.