Ambarella Inc (AMBA) 2018 Q1 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the Ambarella First Quarter Fiscal Year 2018 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.

Thank you, Andrew.

Good afternoon, and welcome to Ambarella's First Fiscal Quarter 2018 Financial Results Conference Call.

We thank you for joining us today.

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

The primary purpose of today's call is to provide you with information regarding our fiscal 2018 first quarter results.

The discussion today and the responses to your answers will contain forward-looking statements regarding our projected financial results, our financial prospects, market growth and demand for our solutions, among other things.

These statements are subject to risks, uncertainties and assumptions.

Should any of these risks or uncertainties materialize or should our assumptions prove to be incorrect, our actual results could differ materially from these forward-looking statements.

We are under no obligation to update these statements.

These risks, uncertainties and assumptions as well as other information on potential risk factors that could affect our financial results are more fully described in the documents that we filed with the SEC, including the annual report on Form 10-K that we filed on March 30, 2017, for the 2017 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 our website.

I will now turn the call over to Dr. Feng-Ming Wang.

Thank you, Deb, and good afternoon, everyone.

I'm pleased to report that our Q1 fiscal year 2018 revenue was $64.1 million, representing an increase of 12.2% from $57.2 million of revenue in the same period of the previous year.

Our non-GoPro revenue grew at 16.1%.

During the quarter, we enjoyed strong revenue growth from IP security cameras, including both professional and home monitoring markets as well as continued growth in body-worn police cameras and OEM automotive video recorders.

We also saw significant number of new 360-degree and VR camera introductions.

At the recent ISC West security show, Ambarella introduced the S5L and S5LM SoC families to enable a new generation of 4K IP security cameras.

The S5L is designed for professional IP cameras and includes 4K High Dynamic Range processing, multi-streaming, on-chip lens distortion correction and dual video inputs.

The S5LM targets the home monitoring market, including battery-powered IP cameras and video doorbell cameras, and uses a single 16-bit DDR memory for low-power operation.

Both chip families feature a quad-core ARM Cortex-A53 CPU for advanced analytics and are fabricated in low-power 14-nanometer process technology.

Also during the show, Ambarella demonstrated some of the first practical implementations of the neural network-based video analytics running on our current security chips.

These included battery-powered camera with person detection to avoid false alarms and extended battery life and advanced face detection and face recognition for intelligent home monitoring.

During the quarter, we successfully brought up our first computer vision chip, CV1.

The chip is on schedule to send to the customers in the second half of the year.

Ambarella is continuing to focus our efforts on the development of a complete family of advanced computer vision SoCs targeting current as well as new markets.

We expect to deliver between 2 to 3 new CV chips per year, covering multiple price and performance points.

We believe our unique approach to building a computer vision platform will give our customers the ability to build new and innovative solution to expand market opportunities.

In the professional IP security camera market, the adoption of the new HEVC video format continued to drive new design wins with our new low-cost 4K-capable S3L and S3LM camera SoCs in China.

During the quarter, HIKvision, the world's largest security product supplier, began volume shipments of an extensive new range of HEVC cameras based on S3LM and spanning from 2- to 8-megapixel resolutions.

Also during the quarter, Dahua began shipping multiple new S3L-based models, including 2-megapixel IR bullet domes, 4-megapixel dual-network cameras and 2-megapixel PTZ cameras with 12x optical zoom.

We expect continued design-win momentum in China for our HEVC IP security camera solution based on our S3L and the new S5L families.

In the home monitoring market, we enjoyed continued design win activity for our retail customers, including new battery-powered designs and HD baby monitors.

During the quarter, Motorola introduced its spherical, "Orbit" battery-powered indoor and outdoor waterproof camera.

Based on Ambarella's S2LM SoC, the camera features a full-HD video and a wide 150-degree field of view, while its built, its SD card supports continuous recording even if the Internet connection is lost.

Also during the quarter, baby monitors start-up, Nanit, introduced a novel baby monitor integrated with an LED nightlight and featuring temperature and humidity sensors.

Based on Ambarella's H3LM SoC, the Nanit's camera uses computer vision to learn how the baby moves, and to inform the parent if the baby is fussy, awake or sleeping.

In the drone market, market leader, DJI, introduced 2 new Ambarella-based drones during the quarter.

In March, DJI introduced the Phantom 3 SE in China, retailing for around $500, based on Ambarella's A9AC SoC.

The video offers a 4K video 12-megapixel still, 720p live view, vision positioning system and over 25-minute flight time.

And in April, DJI introduced its new Phantom 4 advanced model featuring a new camera based on Ambarella's H1 SoC with a 20-megapixel sensor, offering full 4K video recording at 60 frame per second and High Dynamic Range processing.

The drone offers over 30 minutes flight time and retail for $1,345.

Also during the quarter, drone maker, Hubsan, introduced its H501 Air Pro retailing at around $200.

Based on Ambarella's S2L SoC, the drone includes a 1080p video live streaming, Follow Me, waypoints and a 20-minute run time.

Although we are disappointed that the recently announced Spark drone from DJI does not use an Ambarella solution, we realized that as customer grows in new markets, they often desire to have multiple suppliers.

Although we expect DJI to continue to multisource in the future, we also expect that DJI and Ambarella will continue to engage in the development of future products, especially where image and video quality are of primary importance.

This is best demonstrated by the 2 new recently introduced Phantom models, as previously discussed.

In the automotive market, Ambarella has traditionally supplied SoC solutions for aftermarket dash cams as well as dealer-installed accessory models.

We are now experiencing increasing demand for camera video recorders being offered as original equipment by Chinese and Japanese automotive OEMs.

The Chinese automotive market is now the world's largest with 28 million cars sold in 2016 and therefore represents a significant opportunity for our automotive-grade solutions.

During the quarter, Dongfeng Motor Corporation, China's second-largest automaker, introduced its F37 car recorder, which is installed in its AX5 and the AX7 car models, based on Ambarella's A12A SoC.

The car recorder includes 1080p video HD wireless streaming and a wide 158-degree viewing.

And during the quarter, Chinese auto maker, Changan, introduced its CX70 SUV featuring an embedded camera design using Ambarella's A7LA SoC.

The camera includes a full HD recording and displays the video on the car's navigation and infotainment screen.

In the automotive retail market, Garmin introduced its 3 new models, the dash cam 45, 55 and 65W video recorders.

The Ambarella A12 camera feature advanced driver alerts such as forward collision and lane departure warning and can notify drivers when traffic ahead begins to move.

The dash cam 55, 65W allows video recording of up to quad-HD resolution while the dash cam 65W also leverages the A12 dewarping capability to offer a unique 180-degree field of view.

In the wearable market, we are seeing increased adoption of our low-powered HD SoC solution in body-worn police cameras.

At the recent ISC West Show, Panasonic introduced its "Arbitrator" Body Worn Camera.

The robust camera includes wireless connectivity and is integrated with Panasonic's Evidence Management Software.

Based on Ambarella's A12 SoC, it includes support for full HD 1080p video electronic image stabilization and provides more than 12 hours of battery life.

Also during the quarter, Hikvision introduced its "Hai Kang" body-worn police camera.

The new camera is designed to meet the Chinese government's requirement to use the new HEVC encoding standard to provide higher encoding efficiency than existing H.264 standard.

The small form factor camera used Ambarella's S3 SoC with highly efficient HEVC encoding to provide 16 hours of 1080p recording in just 16 gigabytes of storage space.

And during this quarter, we saw the introduction of a number of new 360-degree or VR cameras spanning professional, stand-alone consumer and smartphone-attached accessory models.

The cameras typically employ dual lenses to capture a full 360-degree view and a single Ambarella H2 or A12 SoC to capture, encode and stitch the dual HD or ultra-HD video streams.

While the VR camera market remains nascent, it's too early to forecast the size of the market, the new cameras satisfy the increasing demand for immersive video content driven by Facebook and YouTube support for 360-degree video and VR headsets.

At the NAB conference held in April, Yi Technology introduced its Yi HALO camera professional and was developed in partnership with Google.

Based on Ambarella's A9SE SoCs, the Yi HALO is a 17-camera design capable of shooting stereoscopic video at up to 8K resolution and is built to work with Google's Jump VR creation platform.

In the high-end 360-degree camera category, Ambarella's H2V VR SoC empowers 2 new models introduced by Garmin and Yi Technology.

Garmin's new rugged and waterproof VIRB 360 model captures spherical video in up to 5.7K resolution with a 360-degree audio and includes GPS and sensors for augmented reality overlays.

Videos and photos are automatically stitched in the camera, enabling fast sharing as well as instant live streaming.

Yi Technology's 360 VR camera delivers 360-degree images or videos offers support for live streaming to Facebook and YouTube and is priced at $399.

In the entry level 360-degree camera category, Ambarella's A12 SoC empowers 2 new models introduced by Xiaomi and Ion.

Xiaomi's MiJia, 360-degree panoramic camera records 360-degree video with a 3.5K resolution, and it includes image stabilization, WiFi connectivity and 90-minute battery life.

Ion's ION360 U smartphone camera introduced at the Billboard Music Awards in May includes a phone charging battery case and detachable dual-lens camera.

It captures 8-megapixel 360-degree image and can shoot up to 2 hours of a 4K 360-degree video.

In summary, in addition to continuing growth opportunities in our existing market, we see the combination of video with computer vision technology are the primary drivers of new opportunities in both current and emerging markets, including OEM automotive and robotics.

Our investment in the development of new family of computer vision SoCs to enable our customers next generation of advanced video cameras will be the foundation for the future expansion of our business.

Now I will hand it over to George for more details on the financials and our guidance for Q2.

Thank you, Fermi, and good afternoon, everyone.

Today, I will focus my review on the financial highlights for the first quarter of fiscal 2018 ended April 30, 2017, and then review the financial outlook for Q2 of fiscal year 2018 that ends on July 31, 2017.

During the call, I'll discuss non-GAAP results and, as such, refer to today's press release for a detailed reconciliation of GAAP to non-GAAP results for Q1 2018.

For non-GAAP reporting for Q1, we have eliminated stock-based compensation expense adjusted for the impact of taxes.

Our Q1 2018 revenue of $64.1 million represents an increase of 12.2% from the $57.2 million of revenue in the same period for the prior year.

Revenues in Q1, excluding sales to GoPro and the ODMs, grew 16.1% from the previous year, finishing the quarter at $62.5 million.

In Q1, we saw solid year-over-year growth in the IP security, auto and non-GoPro wearable markets, along with the initial shipments into the virtual reality market, partially offset by a decline in drone revenues.

In IP security, both professional and consumer revenues demonstrated strong growth, led by a solid performance from China's security market and the home monitoring market in the U.S. Initial ramps in the OEM automotive video recorders plus in-line aftermarket shipments resulted in good year-over-year growth for the auto market.

Drone revenues in the quarter were down from last year, primarily due to weakness from Tier 2 drone customers.

Non-GAAP gross margin for Q1 was 64.3% compared to 67.2% in the preceding quarter and 64.6% in the first quarter of the prior year.

The decline in gross margin from the previous quarter reflects the increase in lower-margin China security and U.S. home monitoring revenues as well as the decline in higher-margin sports camera revenues.

In addition, in Q1, we did not have a material impact from the recovery of previously written-down inventory as incurred in the previous 3 quarters.

Non-GAAP operating expenses for the first quarter were $25.7 million compared to $25.4 million for Q4 of 2017 and $24.3 million for Q1 of the prior year.

Headcount and new chip development costs continued to be the major drivers of the operating expenses, representing 67% and 16%, respectively, of total OpEx in the quarter.

Non-GAAP net income for Q1 was $13.4 million, or $0.39 per diluted ordinary share, compared with non-GAAP net income of $11.4 million, or $0.34 per diluted ordinary share in the same period in the previous year.

The non-GAAP effective tax rate in Q1 2018 was 14.9%.

In the first quarter, the non-GAAP earnings per diluted ordinary share are based on 34.7 million diluted shares as compared to 34 million diluted shares in Q1 of 2017.

The total headcount at the end of Q1 was 672 compared to 669 at the end of the previous quarter, with about 82% of the employees dedicated to engineering.

Approximately 72% of our total headcount is located in Asia, primarily in Taiwan and China.

We ended Q1 with cash and marketable securities of $420.2 million, adding $23 million of cash from operations in the quarter.

Total accounts receivable at the end of Q1 was $22.9 million or 32 days sales outstanding.

This compares to accounts receivable of $38.6 million or 41 days sales outstanding at the end of the prior quarter.

Net inventory at the end of Q1 was $19.1 million or 78 days of inventory compared to $20.1 million or 68 days of inventory at the end of Q4.

Accounts receivable and inventory remain in line with company targets.

WT Microelectronics, our Asia logistics supplier, represented approximately 72% of our revenue in the quarter compared to 74% from the same period in the previous year.

WT was the company's only 10% customer in Q1 for both fiscal years 2017 and 2018.

I would now like to discuss the outlook for Q2 of fiscal year 2018.

We expect revenues for our second quarter ending July 31, 2017, to be between $69 million and $72 million, representing an increase of between 6% and 10.6%, respectively, from Q2 of last year.

Non-GoPro revenues in Q2 of fiscal year '18 are expected to grow between 11.8% and 17%.

Non-GoPro revenues exclude estimated revenues to GoPro and ODMs of $6.8 million and $3.8 million in Q2 of fiscal years 2017 and 2018, respectively.

We expect solid year-over-year revenue increases in the IP security, auto and non-GoPro wearable markets as well as continued growth in new virtual reality market, offset by a decline in the drone market.

We expect the drone market to be down from previous year primarily as a result of continued weakness from Tier 2 drone customers as well as the difficult year-over-year comparable due to timing of product launches.

Although the recent introduction of the DJI Spark drone could potentially have a negative impact on the sales of higher-performance drones, we believe our revenues with DJI will grow this year.

For fiscal year 2018, we remain comfortable with our revenue forecast of plus-or-minus 3% growth from last year despite potential weakness in the drone market, particularly from Tier 2 customers.

At this point, we believe our revenues from other markets will be sufficient to offset any potential drone revenue shortfalls.

We estimate Q2 non-GAAP gross margins to be between 62% and 63.5% compared to 64.3% in Q1 of 2018 and 67.1% in Q2 of the prior year.

As expected, we are anticipating the increase in China's security revenues and year-over-year decline in drone revenues to be dilutive to margin in the quarter.

We continue to see gross margins for the year moving into the high end of our target margins of 59% to 62%.

We expect non-GAAP OpEx in Q2 -- in the second quarter to be between $26 million and $27.5 million, with the increase from Q1 primarily coming from increased R&D, headcount and new chip development costs.

We expect the non-GAAP effective tax rate to be approximately 15% in the quarter.

Based on changes in the geographic distribution profits from the company, we expect the non-GAAP tax rate for fiscal year 2018 to be approximately 15%.

We estimate our diluted share count for Q2 to be approximately 34.9 million shares.

In Q1 of fiscal year 2018, the company repurchased a total of 162,738 shares at an average price of $53.91 for a total consideration of approximately $8.77 million under the $75 million share repurchase program that we initially announced in June of 2016.

Approximately $46 million remain available for purchase of shares under the program through June 30, 2017.

While the previous authorized program will continue through June 30, Ambarella's Board of Directors has authorized the repurchase of up to an additional $50 million of its ordinary shares over a 12-month period commencing July 1, 2017.

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 particularly managed 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 capital.

I would like to thank everyone for joining the call today.

And now I will turn it back to the operator to manage the Q&A session.

(Operator Instructions) And our first question comes from Joe Moore with Morgan Stanley.

So you reiterated the growth rate for the year, but I guess, you didn't reiterate the comment about your ex GoPro.

I think it was a 20% growth rate ex GoPro this year.

Is that still the thinking even though you're below 20% in the first half?

Yes.

I think we're still comfortable with the previous numbers.

There may be a little bit higher percentage for GoPro, but I think 20% to 30% is still an appropriate range at this point for non-GoPro growth.

Got it.

And then with regards to the Spark situation.

I mean, I guess, can you just talk about -- that part looks like it only supports 1080p30.

How do you think about those types of markets?

It seems like it's always been competitive.

Is there something new there in that level of product?

Is it more competitive?

Or how do you think about whether you would pursue something with sort of the specs at the lower end of your portfolio?

Right.

So first of all, I think, like we said before, we start -- starting last year, we started seeing a lot of lower cost that are being introduced by customers, and that trend continues.

For example, in addition to the DJI Spark announcement, we also announced a customer called Hubsan introducing a 1080p30 video drone at $200.

So the trend is there.

So I think what DJI, the Spark, is that they have some CV function in there and also that they're very cost-sensitive.

And we believe that in that -- although we haven't seen the final product, Spark, but we do believe that in there is a 2-chip solution, one chip doing CV; the other chip is doing 1080p -- a 40-level video encoding.

So our goal is we're going to continue to support our existing customer with our current product road map for the -- anything below $500 for the low-end drones.

But I think the key is when our CV family of chips are available, we won't be able to provide single-chip solution for the low-end drones.

I think that's going to be the [all] go-to work out in the next couple of years.

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

Maybe I'll continue on the topic of drones here.

You mentioned some weakness with Tier 2 drone manufacturers out there.

Maybe you can comment on the environment there a little bit more?

And are you seeing maybe a shake-up between kind of larger Tier 1s and Tier 2s?

Or if you can help us understand that when you might expect the drone market to come back later this year.

Well, I think if you look at the overall drone market, I think it's still continuing to grow.

It's just that DJI is dominating this market.

I think everybody is seeing that.

Some of the second-tier suppliers have introduced products, but they are having difficulty getting traction in the market.

I think that's also obvious in assessing the second-tier guys introducing a product but cannot really sell in a volume they expected.

And we are still seeing -- continue to see customer introducing new products, especially on the low-cost drones.

So I expect that the trend will continue, that people are going to continue to try with the different featured differentiation on both the low end and high end.

But we're going to continue to see low-end, low-cost drones coming up and -- but we'll continue to watch the momentum.

Today, we still believe that the market is still growing.

Who knows how big the market is?

But we believe that both the high-end drones and low-end drones are continuing to enjoy the market growth.

One other comment there, Richard, is I think potentially, as computer vision, it becomes more capable, the functionality becomes more capable and more cost-effective.

I think that might open up other opportunities, in particular maybe on the commercial side.

Right.

Okay.

I agree, and I look forward to hearing more about that maybe offline.

My second question, on the China IP security market.

I know you've talked about this being pretty good for the last couple of quarters, I think mostly from introduction of your HEVC-enabled product.

How long do you expect to see improvement in China from that?

And is there a stronger cycle going on there?

Or was it just primarily coming from the introduction of your HEVC product there?

Yes.

I think the primary driver this year is from us penetrating the HEVC low-end market with our S3L and S3LM.

I think that's probably going to continue through the remainder of the year.

There's a lot of product introductions that we look forward to over the year, so I think that will continue.

We also see a pretty stable market for the rest of the security market globally.

Other customers seem to be having reasonably good years.

So I think overall, we probably see overall professional security grow a little more than we thought at the beginning of the year.

And our next version comes from Kevin Cassidy with Stifel.

The China OEM automotive designs that you've won, when would those ramp into production?

Well, Kevin, we are talking about multiple different product categories.

For the Chinese video recorder OEM products, we already -- introduction with some of the -- we already started seeing early revenue.

We continue to see the ramp-up in the Chinese OEM for the video recorders.

And in terms of the other markup video product, like [EMEO] and the surround views, we believe it's still second half next year, so around that time.

But we continue to see very high activity on the product design win cycles.

And on the home security camera market, did you see growth in that quarter-over-quarter?

In the home market, we did.

Yes.

Okay.

And there was a new product announced from Nest.

Were you involved with that design?

Or -- that seems to be the first new design they've had in a while.

Can you say what your involvement is still with Nest?

Well, I think you're referring to the new product they announced, which is not ours.

I think that's using our processor design in that product.

And your next question comes from Suji Desilva with Roth Capital.

So on the computer vision chip there.

Can you talk about when you get to 1-chip solution versus 2-chip, do you save the customer relative dollars versus the 2-chip solution?

How are your computer vision chip be priced relative to your current portfolio?

Well, when you go to a 2-chip to single-chip, you -- the biggest saving is the components around the chips because every chip comes with DRAM -- so flash, it's a power controller.

You really need a lot of other components associated with those 2 chips.

So just by 2.2 single-chip design, you're saving huge on the bottom side.

So from the -- our ASP saving versus a 2-chip, I expect we're going to be also providing some savings.

But I think our ASP will determine how competitive the market is at a time.

Okay.

I was just going to add.

I think one thing you will get is a lot of increase in performance from what they get now.

And a lot of what's being rolled out now is software running out, let's say, an apps processor.

Our architecture, I think, will give you much more performance and the ability to expand the functionality in the device.

Okay.

And when the computer vision chip rolls, should we expect a price bump-up relative to the current portfolio?

Or is it kind of at the same price point when it comes to...

I think I expect it to go up.

Yes.

You'll have -- I believe your ASPs will go up, which is normal for us when we introduce new features and functions.

Fair enough.

And then a follow-up question on the drone market.

Can you give us a sense of what percent of revenue is roughly coming in that now and how concentrated it is around DJI to sort of stand the exposure there?

Well, we haven't talked about the concentration.

They probably have 70% to 80% market share.

Last year was about 17% of our revenue, and we expect now it to be probably flat this year, total drone market.

And our next question is from Matt Ramsay with Canaccord Genuity.

I think I just wanted to follow up on a bit of the guidance question.

I think Joe asked at the beginning the non-GoPro piece, George.

If there is a reacceleration in the back half, maybe you can walk us through the drivers of that reacceleration by market.

I think a lot of us had maybe assumed or hoped that the business would be significantly less seasonal now that GoPro is maybe out of the model and going out of the model.

But it looks like from what you're implying, the seasonality might not be that different for this year.

So if could just walk us through that, that would be helpful.

Yes.

I think the seasonality still will be there.

You'll still see step-up this year in drones.

We have a wearable component now, both non-GoPro sports and non-sports wearable, if that makes sense.

Those will expand in Q3.

And particularly, the auto market does not necessarily expand that much.

But this year, we have a rollout of the security market, which is -- that's 2 components.

One of them is seasonal when you have your home security.

So we expect to step up in the home security market in the second half, particularly on the retail component.

And then this year, because we're ramping really a new market on the professional side on the HEVC market in China, we expect that to continue to grow through the second half.

So our year-over-year growth rate now really is across both the professional and the home.

Got it.

That's really helpful.

For me, I wanted to ask a couple of questions on the new CV1 chip.

You talked about sampling this year.

And if I'm not mistaken, I think the first target revenue would be in the drone space.

Maybe you could just walk us through how things have gone so far since you've gotten the chip back and what the milestones are for customer samples and then, of course, revenue.

Right.

So from a CV1 point of view, when we received our CV1 chip from the foundry early May and we successfully boarded up, and in fact, we are writing software and bringing up software at this point and we are -- sit back and we said, on target to sample the hardware and the software to our key customer in the second half of the year.

And we believe that now it's a working chip, we are verifying that our approach to finalize a CV1 was a first.

Then we find the best implementation of a hardware and software combination for those algorithms.

And with that approach, we're now seeing some firsthand results, and we are happy to say that we continue to see a validation of our -- those statements.

And -- but -- and more importantly, we saw CV1 being in the lab, and then we start working on it.

We knew that we couldn't continue to develop our CV road map.

And we believe we'll introduce 2 to 3 new chips per year with the CV capability in the future.

And with also 2 to 3 new chips, we will try to address overall current market and the emerging new market.

The current market, including drones, including autos, including security camera and also for the emerging market, I really think that for the auto OEMs as well as in robotics, it's going to be the key area we're going to focus on.

And also with that 2 to 3 new chips per year with a CV capability, we will try to cover the low to high price point.

So our goal is try to cut, provide a complete coverage on our CV road map in the future so that we can address not only just market but also price points.

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

I want to go back to the drone market, if I could, for a little bit.

With some of the competition emerging in that and then you're talking about a little bit of this weakness on the second-tier or the Tier 2 guys, how do you expect pricing to be affected at the high end and even across your stack?

If competition's coming in and they're choosing to go elsewhere at the low end, do you compete via price to try to keep those sockets?

Or do we kind of a hole we have to deal with until the CV chip can address it?

Well, in fact, for a 1080p30 video that DJI Spark used, it's really old (inaudible) point of view.

We started selling 1080p video solution 10 years ago.

So from the price point, we talked about there probably multiple Taiwanese and Chinese vendors have chips that's capable doing 1080p video at this point.

So the price point is very challenging, but we do have a solution and an ambition to continue to compete with that on any kind of price point.

On the -- any kind of price point.

On the high-end side, of course, because the competition continue to try to go up and we will continue to provide very competitive pricing points as well as the technical solutions.

So I don't think that we won't reduce any design win because of price.

But at 1080p, there's just so many low-end guys out there, sometimes you just don't know who are you competing with.

And I guess, as my follow-up, I wanted to talk about the OpEx side of things, 2 aspects to it.

George, is the fiscal year still going to be up 12% to 14%?

And then I think, Fermi, you talked a couple of times about 2 to 3 new CV chips coming per year.

Can you just talk us through the investment phase of that?

Are we 2/3 of the way through it, and then we're closer to the revenue coming in towards the end of this year and next year?

Or is that an elevated investment going to stay that way as you perhaps even move to 10-nanometer in that 2 to 3 per year cadence?

Okay.

There's a couple of questions there.

I think yes, we're still comfortable with 12% to 14% at this point.

And that includes the rollout of our chips this year of 2 to 3 chips.

Remember, the 2 to 3 chips is 2 to -- total chips every year, which is what we have been doing historically.

So really, the OpEx impact from a chip development standpoint right now is the increased cost from process nodes, which is obviously several million dollars per chip.

But it's not potentially as large as maybe you're thinking.

I think next year, we will continue to on that road map and probably, as we talked about, move to 10 nanometer.

So that investment will continue through next year, particularly when we move towards auto wall chips.

So we will have increased cost associated with having the chips actually ASO-qualified.

Let me add one more point.

So in fact, we already paid out 3 14-nanometer chips.

So the current OpEx level you're seeing only supports 2 to 3 chip a year of a 14-nanometer chip.

From a 14-nanometer to 10-nanometer, the increase is much less than from 28 to 14.

So from the payable fee point of view, [going to] 2 to 3 chips is not as significant as previously.

That's very helpful for me.

One quick clarification housekeeping-wise from George.

I wasn't quick enough to write down when you said the non-GoPro revenue guide sequentially for the quarter.

Can you just run us through that, the GoPro amount again?

Yes.

Nonrevenue increase in Q2 would be about 12% to 17%.

That's on $69 million to $72 million total top line.

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

Perhaps a clarification for me.

Just as you look at the CV1 family, I guess, I was under the impression that, that didn't have the video encoding and that would come from another AMBA solution.

But in a number of the applications for CV1, it does sound like it may have video encoding capabilities.

Can you just clarify whether the video encoding is part of the CV1 solution?

Well, for the video encoding function, we haven't talked about our CV1's spec.

So I'll leave that further.

But in the future, we definitely believe all our CV chips have a complete video processing, video encoding and CV functions.

But when we come to talking about CV1, we'll make it clear how we're going to have inside CV1.

Okay.

So the comments then about single-chip drones may not necessarily be CV1; it just may be a future CV1 -- CV family member?

It's possible.

Yes.

Got it.

Okay.

And the second question.

Just I think in your prepared script, you talked about demonstrating at ISC West neural networks running on the security camera.

Just kind of wondering how you see that application rolling out for image recognition, person or package tracking.

What -- how do you see that rolling out over the next few years?

What percent of the market do you think that applies to?

Yes.

So maybe I should address that competition question first that the Nest camera that is being introduced, a lot of people asking about our competition situation.

And also that's related to your question, I think the key point is that all competitors have not changed.

It's still primarily [core comps] selling a smartphone application processor in a high silica; and also a function of low-cost and the low-end video processing channel [that] we talked about.

But additionally, more importantly, our traditional advantage of video processing, video compression and power consumption has not changed either.

But to your point, as we said in the past, a lot of early adopters of a computer vision design now using software implementation like DSP or CPU to [outperform] time-to-market so that our road map to address this problem is starting with our new S5L, which is a (inaudible) CPU, which can do face detection, face recognition and object detection, as we demonstrated at ISC West, as you mentioned.

And we believe those functions are going to continue to be very important, but I think it's going to be the must-have feature moving forward.

There will be more and more computer vision features required for the security camera for different applications.

That's why we believe that our new CV chip family, that will offer much higher performance, and that we'll address this issue far more efficient than any competition CPU and DSP in the future.

And then just the last question on the security market.

You talked about seeing the HEVC strength within the China market.

Wondering if you're seeing any progress in terms of HEVC adoption outside of the Chinese market.

Well, it's an interesting question because yesterday at Apple's conference, they announced they're going to innovate a OS-controlled support HEVC.

So what I mean, although I cannot -- I don't know f or sure what I mean is I think Apple figured out how to negotiate their patent royalties.

We saw 2 major patent pools and as well as other patent holders, especially in terms of video streaming patents.

So I think with that, I believe that the HEVC format finally gets some traction in the market.

And hopefully, with that, we're going to start seeing more and more HEVC requirements in the market.

Especially if we're going to order a smartphone, we'll be able to decode HEVC and display HEVC strings.

Then I think that will open up the whole platform.

It has been delayed for 2 to 3 years, but I think finally, we'll start seeing some indications that the problem will be resolved.

And our next question comes from Brad Erickson with Pacific Crest Securities.

First, just regarding ASO-qual.

Are you able to just quantify that incremental spending you might see next year on our R&D line relative to, say, the CV spending we're seeing this year?

Not yet.

We're working on it now.

It is a major project.

I think it will add costs to next year.

I'll probably be able to talk more intelligently about it on the next call.

But we can expect OpEx to continue to increase to absorb that next year.

Increase as a percentage of revenue or?

Yes.

It very well could increase as a percentage of revenue, but we haven't got the final -- and really, the final structure on how to go forward with that yet.

Got it.

That's helpful.

And then secondarily, based on what you know from your work thus far in this lab, what are kind of the most suitable applications within automotive you see developing from an opportunity set?

Will it just kind of be basic ADAS functionality at first?

Or do you see more imminent potential for level 2-, level 3-type applications?

And the second, just where do you kind of intend on differentiating versus some of the more established players in that space?

Thank you.

So first of all, of course, our focus with a big lab technology in CV1, we'll be demo-ing ADAS function, for sure.

But I think going through a autonomous car level 4, level 5 is still our ambition and we want to achieve in the future.

And CV1 is a great fundamental to achieve that goal.

But like you said, the early revenue will be coming from the ADAS market.

In fact, we do believe the ADAS market does exist not only on the auto OEM side, but also on some of the aftermarket side.

So we will try to address both with our CV families.

And the way we differentiate with a big player is we saw this lab acquisition, we got a complete software stack of autonomous car.

Basically, what I mean is we have a software that -- we can use that software.

We control one software that can drive a car.

I think very few companies can say that.

And now we have not only the software but also the hardware, which optimized for the algorithm and the software implementation of this lab software.

And we are probably one of the very few who have the combination of differentiated hardware and complete software stack that can do autonomous car today.

And we are trying to -- going to try to demo it, but I think that's definitely how we kind of differentiate.

And of course, because of our differentiated technology, on the hardware side, we expect that we deliver higher performance and much more power consumption.

That's another thing, which is consistent with our advantage of our video processing technology.

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

I have one just quick housekeeping for George, then a follow-up for me.

So George, I noticed on the balance sheet, there was maybe a $9.5 million sequential jump in intangible.

What was that?

Yes.

Under the new accounting regulations, some of these long-term operating leases now have to be capitalized.

So our lease on our software we use to develop the chips, which is very expensive cadence software, became a capital asset.

So we recorded the asset and the liability, and we will now start depreciating the asset versus recording it as a lease payment.

The P&L impact is 0.

Got it.

Got it.

Okay.

Perfect.

And then for me, just going back to auto, going to continue the conversation there.

Some of the established players in the market in ADAS, they sort of built their businesses through some Tier 1 partnerships.

To what degree are you guys involved there?

And is there opportunity for you and as you come out with CV and as you come out with some of the new solutions to establish some Tier 1 partnerships to help you grow in that market?

Absolutely.

In fact, in the last 3 years where I spent a lot of time to try to build relationships with Tier 1. In fact, when we mentioned ship starts with non-ADAS products, by working on the video recorders, working on e-mails and working on the surround views, now we are engaging with a lot of Tier 1s, not only just discussion of the project but also development of new projects.

And we're sure that we will become a viable supplier to those Tier 1s.

I think that's a huge step.

And during the meantime, all of 3 markets, doesn't matter if it's recorders emails or surround views, they all require ADAS road map for those kind of products.

And with that, because we are the supplier with the current solution and we have a chance to talk to the ADAS team to understand their requirement as well as what they want from us.

So I think our Tier 1 engagement has been there for 3 years and getting stronger through our current products, hopefully, we can convince that to use our CV1 or CV family chip for the ADAS in the future.

And that does conclude our Q&A session.

I would like to turn the call back to Mr. Wang for closing remarks.

Thank you for you all for joining the conference call today.

And particularly, I would like to thank all our employees for their continued dedication and support.

I'll talk to you next time.

Thank you.

Ladies and gentlemen, thank you for participating in today's program.

This does conclude today's program.

You may all disconnect.

Everyone, have a great day.