奇景光電 (HIMX) 2017 Q2 法說會逐字稿

Good day ladies and gentlemen, and welcome to the Himax Technologies' second quarter 2017 earnings results conference call.

(Operator Instructions) As a reminder, this conference call may be recorded.

I would now like to turn the conference over to Greg Falesnik, Managing Director of MZ North America.

Thank you, operator.

Welcome everyone to Himax's Second Quarter 2017 Earnings Call.

Joining us from the company are Mr. Jordan Wu, President and Chief Executive Officer; and Ms. Jackie Chang, Chief Financial Officer.

After the company's prepared comments, we've allocated time for questions in a Q&A session.

If you have not yet received a copy of today's results release, please e-mail greg.falesnik@mzgroup.us or access the press release on financial portals or download a copy from Himax's website at www.himax.com.tw.

Before we begin the formal remarks, I'd like to remind everyone that some of the statements in this conference call, including statements regarding expected future financial results and the industry growth, are forward-looking statements that involve a number of risks and uncertainties that could cause actual events or results to differ materially from those described in this conference call.

Factors that could cause actual events or results to differ materially from those described in this conference call include, but are not limited to, general business and economic conditions, the state of the semiconductor industry, market acceptance and competitiveness of the driver and non-driver products developed by Himax, demand for end-use application products, the uncertainty of continued success in technological innovations as well as other operational and market challenges and other risks described from time to time in the company's SEC filings, including those risks identified in the section entitled Risk Factors in its Form 20-F for the year ended December 31, 2016, filed with the SEC in April 2017.

Except for the company's full year of 2016 financials, which were provided in the company's 20-F and filed with the SEC on April 12, 2017, the financial information included in this conference call is unaudited and consolidated and prepared in accordance with U.S. GAAP accounting.

Such financial information is generated internally and has not been subjected to the same scrutiny and review, including internal auditing procedures and external audits by an independent auditor, to which we subject our annual consolidated financial statements and may vary materially from the audited consolidated financial information for the same period.

The company undertakes no obligation to publicly update or revise any forward-looking statements whether as a result of new information, future events or otherwise.

I will now turn the call over to Ms. Jackie Chang.

Thank you, Greg and thank you everybody for joining us.

Our outline for today's call is first, I will be report Himax's consolidated financial performance for the quarter on a GAAP basis and provide supplementary results on a non-GAAP basis which excludes share-based compensation and acquisition-related charges.

Second, I will provide the third-quarter outlook.

Finally, Jordan would discuss our business and product highlights, after which we will take questions.

Our 2017 second quarter revenue and GAAP earnings per diluted ADS [came in] at midpoint of our guidance, while gross margin exceeded the guidance, for the second quarter we reported net revenue of $151.7 million with a gross margin of 23.8% and GAAP loss per diluted ADS was $0.004.

The second quarter revenue of $151.7 million represented a decrease of 2.2% sequentially and a decrease of 24.5% year-over-year.

I will go through the issue was causing the revenue decline below.

Revenue from large panel display drivers was $52.1 million, down 12% sequentially and down 22.8% year-over-year.

Large panel driver ICs accounted for 34.4% of our total revenue for the second quarter, compared to 38.2% in the first quarter of 2017 and 33.6% a year ago.

The decline was due to temporary slowdown of our large-sized driver IC business caused by earlier misses of certain customer's new design in projects as we have reported in the last quarter's earnings call.

We have overcome the engineering hiccup and business will be back on track starting the third quarter.

In spite of the lukewarm sales in the first half of 2017 our engineering collaboration and design-in activities with large-panel customers across China, Taiwan and Korea all remain robust.

Such activities will lead to future rebound in sales momentum.

Revenue for small and medium sized driver came in at $70 million, up 5.1% sequentially and down 22.7% year-over-year.

Driver ICs for small and medium-sized applications accounted for 46.1% of total sales for the second quarter as compared to 42.9% in the first quarter of 2017 and 45% a year ago.

Sales into smartphones increased 2.5% sequentially and down 51.5% year-over-year.

The less than satisfactory result in the second quarter was caused mainly by weak sentiment in the China market since most brands were preparing for new models based on 18 versus 9 displays, which Jordan will elaborate a bit later and therefore turn cautious in building inventory for legacy 16 versus 9 displays.

In addition our sales were affected by the shrinking addressable market for pure TFT-LCD driver ICs, a significant portion of which is being replaced by TDDI and AMOLED technologies, as we indicated in the previous earnings call.

As Chinese OEM customers began to replenish inventory for the new release in the second half, we have seen strong recovery in the third quarter.

Our small and medium-sized driver IC revenue for automotive application increased 15.6% sequentially and 50.6% year-over-year.

We are happy with the continuous strong momentum in this space.

Our drive IC using tablets increased around 8% sequentially, but declined 13.7% year-over-year for weak overall market demand in the product segment.

Revenues from our non-driver businesses were $29.6 million, up 0.9% sequentially and down 31.1% versus last year.

Non-driver products accounted for 19.5% of total revenues, as compared to 18.9% in the first quarter of 2017 and 21.4% a year ago.

The sequential increase was primarily due to NRE contribution.

The year-over-year declines were due to discontinuation of LCOS and WLO shipments to one of our major AR device customers who decided to end the products production as we've reported before.

To a much lesser extent, lower sales of touch panel controllers and power management ICs also contributed to the year-over-year decline.

Our GAAP gross margin for the second quarter was 23.8%, up 70 basis points from 23.1% in the first quarter and down 230 basis points from 26.1% for the same period last year.

The sequential increase was a result of a more favorable product mix and higher NRE.

The year-over-year decline was primarily caused by unfavorable product mix and margin decline in the driver ICs product lines.

Now let's take a look at operating expenses.

GAAP operating expenses were $37.1 million in the second quarter of 2017, up 8.1% from the preceding quarter and up 21.4% from a year ago.

The sequential and year-over-year increases in the second quarter are in line with the operating expense budget that we reported in the last earnings call.

The sequential increase was primarily the result of rising R&D expenses in the areas of 3D sensing, WLO, TDDI, and high-end TV.

The year-over-year increase, on top of the same reasons above, was also caused by annual merit increases.

In addition, NT dollar appreciation against the U.S. caused our salary expenses to increase around $1 million as we pays the bulk of our employee salaries in NT dollars.

GAAP operating margin for the second quarter of 2017 was minus 0.6%, down from 10.9% for the same period last year and down from 1.0% in the previous quarter.

The sequential decline was a result of lower sales and higher expenses in the quarter and the year-over-year decreases was a result of lower sales, lower gross margin and higher expenses.

Second quarter non-GAAP operating loss was $0.4 million, or minus 0.3% of sales, down from 11.1% for the same period last year and down from 1.3% a quarter ago.

Again, the sequential decline was a result of lower sales and higher expenses in the quarter, while year-over-year decrease was caused by lower sales, lower gross margin and higher expense.

GAAP net loss for the second quarter was $0.6 million or $0.004 per diluted ADS compared to GAAP net income of $1.4 million or $0.008 per diluted ADS in the previous quarter and GAAP net income of $19.8 million or $0.115 per diluted ADS a year ago.

Second quarter non-GAAP net loss was $0.3 million or $0.002 per diluted ADS compared to non-GAAP net income of $1.7 million last quarter and non-GAAP net income of $20.2 million the same period last year.

Turning to our balance sheet.

We had $185.9 million of cash, cash equivalents and marketable securities as of end of June, 2017 compared to $179.3 million at the same time last year and $199.5 million a quarter ago.

On top of the above cash position, restricted cash was $107.2 million at the end of the quarter as compared to $107.4 million in the preceding quarter and down from $138 million a year ago.

The restricted cash is mainly used to guarantee the company's short-term loan for the same amount.

We continue to maintain a very strong balance sheet and remain a debt free company.

As of June 30, 2017, our inventories were $147.

7 million, little changed from $148.3 million a quarter ago and decreased from $186.7 million at the same time last year.

Accounts receivable at end of June, 2017 were $163.2 million as compared to $187.9 million a year ago and $167.7 million last quarter.

Days sales outstanding was 96 days at the end of June 2017, as compared to 90 days a year ago and 97 days at the end of the last quarter.

Net cash outflow from operating activities for the second quarter was $1.2 million as compared to an inflow of $13.1 million for the same period last year and a inflow of $5.5 million last quarter.

The sequential decrease was mainly due to income tax payments of $8 million, a year-over-year decrease was the result of lower profitability.

Capital expenditures was $11.9 million in the second quarter of 2017 versus $1.7 million a year ago and $2 million last quarter.

The second quarter CapEx consisted mainly of capacity expansion for WLO production lines and ongoing payment for the new building construction.

As reported in the last few earnings calls we are increasing CapEx right now to enlarge our WLO capacity within the current headquarters to meet certain anchor customer's strong and urgent demand.

We are also constructing a new building to house further WLO capacity, the net generation LCOS production line and additional office spaces.

This is our Phase 1 expansion which is $80 million as we announced in the last earnings call.

We declared an annual cash dividend of $0.24 per ADS during the second quarter totaling $41.3 million, which will be paid out on August 14.

Our dividend is determined primarily by the prior year's profitability.

Our decision to pay out 81.4% of last year's net profit demonstrates our continued support for our shareholder base and strong confidence in the near-term return outlook for our newly increased CapEx and our overall long-term growth prospects.

The CapEx budget for 2017 and the dividend for the year of 2016 will be funded through our internal resources and banking facilities if so needed.

As of June 30, 2017, Himax has $172 million ADS outstanding unchanged from last quarter.

On a fully diluted basis the total ADS outstanding are $172.5 million.

Now goes to 2017 Q3 guidance.

For the third quarter of 2017 we expect revenue to be up 23% to 30% sequentially.

Gross margin is expected to be up 1% sequentially depending on our final product mix.

GAAP earnings attributable to shareholders are expected to be in the range of $0.013 to $0.025 per diluted ADS based on $172.4 million outstanding ADS.

Non-GAAP earnings attributable to shareholders are expected to be in the range of $0.03 to $0.042 per diluted ADS based on $172.4 million outstanding ADS.

For the third quarter of the 3 product categories, we expect large panel driver IC to increase around 10% quarter-over-quarter those for small and medium-sized panels to be up by around 20% sequentially and non-driver IC business is expected to increase around 90% sequentially.

As we have done in the past, our third quarter GAAP earnings per diluted ADS guidance have taken into account of our expected 2017 grant of restricted share units or RSUs to the term at the end of September.

The 2017 RSUs subject to our Board approval is now assumed to be around $3 million, almost all of which or $0.015 per diluted ADS will be vested expense immediately on September 30th.

The grant date in comparison the 2016 RSUs totaled $12 million out of which $9.2 million or $0.043 per diluted ADS was vested immediately.

The grant of RSUs will lead to higher third quarter GAAP operating expense compared to the other quarters of the year.

I would now turn the call over to Jordan.

Thank you, Jackie.

Despite the decline in the first half in our business, we anticipate a strong recovery in driver IC segments and exciting opportunities in the non-driver segments over the remainder of the year and beyond.

Looking ahead we believe our overall financial performance will be resilient.

Now let me provide you with some details behind our guidance and trends that we see developing in our businesses.

Firstly on our large panel driver IC, we have a positive outlook on the growth momentum in our large display driver IC business.

While global TV shipments may experience 0.4% year-over-year decline, China's share of global TV panel shipments is projected to reach 33% in 2017, compared to 29.5% last year.

Being a market leader in the large display driver IC business in China, we will capitalize on China's rising market share.

Not only have we refreshed our product roadmap and delivered better product costs to our existing and newly added customers, we have also secured new design wins, particularly in 4K TV to solidify our growth for the remainer of the year.

Looking forward, 4K TV penetration is still on the rise and Chinese panel customers are still ramping new advanced generation fabs over the next few years, including a brand new Gen 8.5 fab and another Gen 8.6 during the second quarter of 2017.

This will help further grow our revenue and market share in the large panel segment in 2018.

In the small and medium-sized panel segments, our driver IC sales for automotive applications have enjoyed over 30% annual growth over the last few years, well surpassing the market average.

Still more panels are going into vehicles, with the number of units expected to increase from 135 million in 2016 to 200 million in 2022.

We have successfully engaged literally all of the major automotive panel manufactures worldwide for long-term partnerships and secure many of their key projects pipelined for the next few years.

The most significant segment in our small and medium-sized panel drive IC business is ICs used for smartphones.

Our customers have started to replenish inventories after the lackluster first half and we have added more design-wins for 18:9 displays.

Sales into smartphones already rebounded in June and we expect a strong recovery into the third quarter and fourth quarter.

In our last earnings call, I discussed how 18:9 displays are becoming a trend and how we expect higher TDDI penetration in the smartphones going forward.

Both of these trends held true and continued to accelerate in the second quarter.

In order to increase the effective viewing area of the display without enlarging the overall size of the phone, new aspect ratios and bezel-less designs are essential.

Most of the brands are now preparing for the change into new displays featuring the so called full HD+ and HD+ resolutions, which have an aspect ratio of around 18:9 to 21:9.

We predicted this market shift and have been working hard to get [ourself] ready for this new trend.

We have been awarded several important projects for major brands.

New design-in activities are ongoing and we're already starting mass shipments for some of the projects, helping boost our third quarter revenue for the segment.

In terms of our progress in TDDI, I reported in the last earnings call that we made further investments into R&D and customer engineering to catch up with our customers' requests for fast product ramps.

I'm pleased to report that our Full HD+ TDDI solutions have drawn tremendous interests from Tier-1 brands and most panel makers in China, Japan and Korea, primarily because our TDDI solutions enable super-slim bezels for customers' panel design.

We expect our Full HD+ TDDI solutions to be a significant contributor to our revenues in the fourth quarter and beyond.

Now very briefly on the touch panel controller IC, while the discrete touch panel control IC is being quickly replaced by TDDI, we expect this product's revenue will in fact grow over 40% in the third quarter as some of the earlier design-in projects featuring our on-cell solutions started volume shipments for Chinese smartphone brands.

Now I will talk about AMOLED displays.

We have joint development projects with many of our Chinese OLED panel customers and have delivered product samples to some of them in the second quarter.

With Chinese smartphone brand's AMOLED adoption forecast to reach 18% in 2017, Chinese panel makers have committed tremendous capital to build 7-10 brand new OLED fabs and are driving at full speed to pull forward the mass production schedule.

AMOLED is set to become mainstream in the global smartphone market in the near future with penetration potentially reaching as high as 50% by 2020.

Once the Chinese panel makers start to mass produce OLED displays, we believes OLED driver IC will be one of our growth engines for our small panel driver IC business.

Now on our non-driver IC business segments.

The segment has been our most exciting growth area and a clear differentiator for Himax in the past few years.

Now let me share some of the exciting progress we made in the last quarter, as well as future growth opportunities.

First I will touch on our 3D sensing total solution.

We believe 3D sensing is among the most significant new features for the next generation smartphones.

Our SLiM product line, which stands for structured light imaging module, are based on structured light technology is a state of the art total solution for 3D sensing.

Our goal is to provide total solution with performance, size, power consumption and costs all suitable for smartphones and tablets.

We offer fully integrated structure light modules, with the vast majority of the key technologies inside the module also developed and supplied by ourselves.

These critical in-house technologies include advanced optics using our world leading WLO technology, laser driver IC, high precision active alignment for the projector assembly, high performance near-infrared CMOS image sensor and last but not least, an ASIC chip for 3D depth map generation.

The fact that all of these critical building blocks are developed in-house puts us in a very unique position.

We are able to react quickly and tailor our solutions to customers' specific needs.

It also represents a very high barrier of entry for any potential competition and a much higher ASP for us.

While we prefer to offer a total solution, we can also provide the aforementioned individual technologies separately to select customers so as to best accommodate their specific needs.

Thanks to our absolute technology leadership, our progress made with the fully integrated structure light 3D sensing total solution module is very exciting.

We are seeing strong demand for 3D sensing solutions from numerous Tier 1 customers.

We are in close collaboration with select leading smartphone makers and partners right now, aiming to bring our total solution to mass production as early as 2018 to meet our -- as early as early 2018 to meet our customer's aggressive launch timetables.

Moreover, given that we are offering highly integrated solutions with ASPs much higher than those of individual components, by the time we start shipping our total solutions, they will be a major contributor to both our revenues and profit, and consequently create a more favorable product mix for us.

Furthermore, our 3D sensing will be a game changing technology for a wide range of applications.

The smartphone space is our current focus, however, we believe over time it will be a necessary feature for applications such as AR/VR, industrial, IoT, AI, automotive, robotics, military, surveillance and drones.

In the last earnings call, we reported that this year's CapEx will be significantly higher than usual.

We also reported the urgent addition of a new WLO capacity to meet the rush demand of a certain customer.

This new capacity is located in our existing headquarters in which we retrofitted space to make room for the new equipment.

We are pleased to report that the project is going smoothly as planned.

A major ramp of the new WLO capacity has already started at the beginning of the third quarter and will accelerate throughout the remainder of 2017 and beyond.

Now let me move on to our other WLO business updates.

Advanced wafer-level optics or WLO is one of the key technologies enabling 3D sensing, AR devices and many other applications.

At the present time, 3D sensing is the top priority of our WLO business.

Leveraging on our exceptional design know-how and mass production experience in WLO technology, we are able to produce the world's most compact optics required of 3D sensing while achieving superior performance.

In addition to 3D sensing, we also have ongoing collaborations with customers in developing light-guide for AR glasses and micro displays using our WLO technology.

Now moving on to the other major CapEx project of this year, i.e., the construction of a new building.

The progress has been good to date and everything is proceeding according to schedule.

The new building, located near our current headquarters, will house additional 8 inch glass WLO capacity and provide the extra office space we desperately needed.

The new building will be completed and ready for personnel and equipment move-in by the end of 2017 or early 2018.

Its timely completion is particularly critical for 3D sensing total solution business as it will house the new WLO capacity needed for multiple smartphone customers.

Judging from our customer's enthusiasm, we are planning to kick start the Phase II capital expenditure beyond the Phase I $80 million that we announced earlier our much sooner than expected in order to fulfill the strong 3D sensing demand for the next 2 or 3 years.

The Phase II capacity will still be located in the new building.

In fact, the new building has sufficient room to house capacity much in excess of the Phase II expansion.

We expect the Phase II investment to provide a handsome return and will entrench Himex, as the top tier customers' 3D sensing go through supplier for its leading technology, a reliable capacity support in this up and coming industry with tremendous growth potential.

As reported before among of all the components in our 3D sensing total solution, the only two items requiring capital expenditure for us are advanced optics, utilizing our in-house WLO production line, and active alignment for which we have developed a state-of-the-art solution.

The two items are not outsourced because they require highly differentiated manufacturing know-how and are critical factors of our competitiveness.

I would report the Phase II CapEx plan in due course.

Now on to our CMOS image sensor business update.

We continue to make great progress with our 2 machine vision sensor product lines, namely near infrared or NIR sensor and Always-on-Sensor.

Our NIR sensor is a critical part of the structured light 3D sensing total solution, which I I spoke on above.

Our NIR sensors overall performance is far ahead of those of our peers in 3D sensing applications.

We currently offer a low noise HD or 1 megapixel and 5.5 megapixel NRI sensors and are planning to add more to further enrich our product portfolio.

Our NIR sensors deliver superior quantum efficiency in the NIR range especially over 940 nanometer band, which is critical for outdoor applications.

Our Always-On sensor or AoS solutions provide super low power vision -- super low power computer vision which enables new applications across wide variety of industries.

The ultra-low power Always-On vision sensor is a powerful solution capable of detecting, tracking and recognizing its environment in an extremely efficient manner using just the few many watts of power.

We are pleased to report that we already have one major global brand leveraging our AoS in their new high end TV models which have already hit the market.

For the traditional human vision segment, we see strong demands in notebooks and increased shipments for multimedia applications such as car recorders, surveillance, drones, home appliances and consumer electronics among others.

I mentioned earlier that one of the critical elements of our 3D sensing total solution is an ASIC for 3D depth map generation.

We are able to develop the ASIC, thanks to our unique in-house capability in developing video ASICs for customers.

Equipped with the ASIC, our 3D sensing total solution can substantially reduce the power consumed while processing 3D sensing, enhance personal data security, accelerate the 3D depth map generation and free up a smartphones processor for other applications.

We view this unique capability as a significant competitive advantage.

It has been and will continue to be one of our key drivers in the success of our 3D sensing total solution.

I will now turn on to our LCOS update.

Even though the market is still in development stage, we continue to see heavyweight companies allocating major R&D resources and budgets in their push for AR goggle devices.

Our list of customers continues to expand and now it covers many of the world's biggest tech names.

In addition to AR applications, we are pleased to report that we are making great progress in developing high-end heads-up-displays for automotive applications.

This represents a significant long term growth opportunity for us.

Our technology leadership in this space has little competition, which is evidenced by our partnerships with Tier-1 companies who have launched their AR products with Himax inside, as well as our partnerships with the world's leading AR glasses producers.

That concludes my report for this quarter.

Thank you for your interest in Himax.

We appreciate you joining today's call and we are now ready to take question.

Thank you.

(Operator Instructions) Our first question comes from the line of Thomas Sepenzis of Northland Capital, your line is now open.

I am wondering Jordan, if you can talk a little bit more about your structured light solution and how that might compared to a VCSEL in terms of overall cost or capability, what the advantages might be, and when you think you might actually see some revenue from that product?

Thank you.

Thank you, Tom.

There is actually a lot of misunderstanding around the marketplace and I've been seen research reports lately because apparently people started to speculate about the up and coming smartphones having 3D sensing capabilities.

Now, I said this is misunderstanding because one should not compare structured light against VCSEL.

VCSEL is one the 2 major types of lasers, and structured light, can use both VCSEL and the other type of laser which is called Edge Emitter.

Okay now VCSEL and Edge Emitter there are pros and cons among the -- between the 2 and in fact, we have our current in-house solution in terms of total solution we are using primarily Edge Emitter type but we are also working with our customers on VCSEL type meaning our technology can actually (inaudible) cope with both.

so again, I mean one should not compare VCSEL with structured light, a better comparison if I may elaborate on the question, I think it's very important for everybody to understand about 3D sensing.

Structured light technology, a right comparison with this technology is something called Time-of-Flight or ToF or another type, which is little mentioned is stereoscopic or dual camera type.

I will talk about dual camera because dual camera is already in the race.

Dual camera by definition, in theory can generate 3D picture or 3D image but it's not really a suitable solution for smartphones, because dual camera require two cameras to be separated apart quite far away and your smartphone simply doesn't a space to take that distance.

And that is why although dual camera has been around in smartphone for quite some time.

You don't -- you never see dual camera in those smartphones enabling 3D sensing because of this very reason, because they are not separated far away enough to really generate the 3D data.

So you are -- what we are seeing today is for the dual camera to only enhance your 2D photos, enhance the image quality of the 2D photos.

So that's what the dual camera does today and I think it will continue to do so in any foreseeable future.

Okay.

So, but because dual camera by definition it sounds as it can generate 3D, that's why I thought I should mention it, but let's just knock it down, because this is not really suitable for smartphone.

Now the race where they've come down into two type of technology, one is our structure light and the other one is ToF or Time-of-Flight.

Before I go on to compare the two, I must mention our total solution of today, is structure light type but we have worked with partners, third-party partners on ToF type as well.

And in fact we are -- in our optics and laser, our projector side we can totally do ToF.

And we need to sensor because our sensor today CMOS sensor which is by definition a structure light type of sensor or receiver.

For ToF we require, we will need another type of sensor which is ToF sensor.

We do have a plan to kick-start a ToF project, problem is we are so tied up today, so we can't really, we don't really have the manpower to do so.

But I think probably next year, we will probably start our ToF solution development as well.

So, technology-wise, it's not really that difficult for us.

Now get back to the main issues of the pros and cons(inaudible) of the two technologies.

Structure light by definition it's like you mention you have a projector, you have a receiver, the same goes for both there you have algorithm to compare the projection side and receiving side and that's how you generate your 3D data.

The difference is structure light, you kind of project a structure, an invisible structured service or image, right?

And,it hits the objects, bounce back, and you have a receiver to catch it.

And then, you compare it.

So you are really talking about an image.

But with ToF, which is Time-of-Flight which by definition is basically you try to measure the outgoing signal, the timing of outgoing signal, and signal coming back, you catch it.

You measure the time difference and the speed of light.

That's how you measure the distance and therefore create the 3D map.

So in terms of ToF rather having projected out a page image, it's really a spot of dots (inaudible), a spot of dots (inaudible).

So their theories are very different.

Now on pros and cons, I would mention a few areas of difference.

The first one is resolution, which I think is the most critical among all because if you look at our smartphone camera experience, we started with -- VGA are even sub VGA which is about 300,000 pixels.

And before you know over a short period of time, a few years, now, we are talking about well over 10 megapixels already.

So the lesson is that once consumers has it, they want better resolution, no matter what.

So there consumers are hungry for the resolution.

I think this will hold true for 3D sensing.

ToF in resolution is its major, major drawback because light travels very fast.

So you need, your pixel needs to take a lot of photos in order to get the accuracy, i.e., your pixel size needs to be rather large and therefore you cannot shrink your pixel size, it's a physics limitation, right?

But the structure light type, our sensor is very similar to the RGB,(inaudible), the image CMOS sensor that are already used to in our smartphone camera today.

So the sensor resolution can be enlarged.

It's a matter of Moore's law.

It can enlarge very quickly, very fast.

It's a easy effort.

So we are betting on structure light initially, primarily because we have this firm belief that with structure light we can easily create a roadmap with enhanced resolution every year.

With TOF -- it's very, very difficult if not impossible.

So the resolution comp between the two is by the differences by order of the magnitude.

So is very, very very big.

The second comparison is about precision that's how you can capture the detail of your 3D so to speak, right?

In here structure light because you're talking about a page of image, so we've seen somewhere around, I would say 5 or 6 meters, structure light outperforms ToF in precision.

But beyond that, because it's about -- the speed of light, so the longer resistance is actually easy for them.

So ToF I would say 5 or 6 meters or above ToF is probably its strengths area, however back to the first point of resolution, when it's that far way people do also require a high resolution as well.

So again that gets back to ToF's weakness.

By anyway in precision, both has its strengths and weakness.

And the third area is the capability to be integrated with your RGB sensor.

Bear in mind structure light, the sensor is really identical to your image sensor.

So you're having, one page of RGB sensor, image against another page of black and white 3D sensor image, so it's very, very easy to have each enhancing the other and even to synchronize them and generate 3D photography.

While with ToF is very, very different.

As I mentioned earlier, the ToF sensor by nature is very, very different than your RGB sensor so by definition is very, very difficult to integrate and particularly with the lack of resolution, the integration even if you can achieve that, the effect will not be satisfactory.

So that's integration capability, in here structure light also has benefit.

And the fourth area, ToF has a advantage [cushion] which is X and Y side.

I talk about the dual camera, the projector and receiver needs to be separated far away.

To much lesser extent, for structure light but structure light because your projector and receiver with an image.

So the projector and receiver also needs to have some distance, that distance must be accommodated by your smartphone.

But with ToF there is almost no distance required because they are measuring the time of -- the speed of light, so they don't really need a distance between receiver and projector.

Therefore their X and Y dimension can be a lot smaller compared to structure light, that is ToF advantage.

And finally assembly, the (inaudible) easiness for module assembly.

Structure light again because of its very short distance between projector and receiver the assembly is a lot easier to handle compared to structure light.

So, I mean as I said there are pros and cons between the two and (inaudible) and which one will win over but I think my bet is that(inaudible) they will kind of coexist in the market place for a very long time.

They are just, they will just be used for the right applications, right range, right scenario, right situation.

So I, my bet is(inaudible) both will coexist.

And that is why I think in the long term actually, not very long-term, in fact probably starting next year we'll start to develop a ToF solution.

I'm sorry for the very long answer but I think it's a very important question.

We are very early stage in this new technology.

So I think it's a good idea that we kind of set the record straight for this very important question.

Our next question comes from the line of Tristan Gerra of Baird.

Also a quick technology question regarding the headup display you've mentioned some progress if first, you could talk about the timing and of potential ramping and what revenue opportunity and also compare it with the DLP or micromirror MOEMS technology that some competitors seems to be implementing for headset displays if you could just give us a quick overview of the differences in technology and why you think you are well positioned there?

Firstly, quite frankly the market is slowing down.

There are couple of very high profile launches, but they hit the sweet spot of the consumer's demands, if you will, and we were fortunately enough to have participated in both in fact in many others as well.

So I think for us in our customers we have kind of learned from the lessons and basically they are some technology issues but more importantly are cost issues.

Technology issues include the very famous field of view and certainly the weight and the size and the mechanical design, the industrial design of the wole gear(inaudible) and what not but I think the most important thing is [property]probably cost and the -- and because of this issues the developers are not particularly enthusiastic (inaudible) about developing as particularly for AS glasses and that just adds to the problem.

So I think what us and our customers are working very hard to do behind the scene, if you like, is try to solve all these problems [hope] hopefully once and for all and there are -- and so you're probably not hearing like higher-- high profile or our announcing break through (inaudible) LCOS technologies again but there are actually a lot of activities going on as we speak behind the scene.

Having said though -- we do anticipate if not end of this year probably sometime next year there will still be customers trying to launch products, new products, tried to further test the market but I think again what we and the industry leaders are doing right now I believe is fair to say is that we want to sit back and really learn from lesson and really tackle the problem that we learned from the past.

As far as comparison with DLP and Scanning Mirror.

But Scanning Mirror involves laser and (inaudible) actually DLP as well, they are there -- as far as I (inaudible) can tell, sparkling remains the major, major issue, technical issue and it is really very, very difficult to solve.

And I think and actually, I don't want to elaborate too much here DLP is never really a serious player in goggles for a lot of technological barriers, DLP is quite suitable for like strong illumination projector.

But when it comes to goggles type of displays not really suitable.

Some customers are starting on scanning mirror more carefully right now but again we're also hearing a lot of very difficult to solve problems and the fact that the scanning mirror has been around for very, very long time for as long as (inaudible) LCOS, and it's nothing new to goggle customers.

People have been studying scanning mirror ever since we got started on this goggles business.

But I haven't seen anything really concrete so far.

So I think it's fair to say that the LCOS remains the leading technology, when it comes to AR devices -- AR goggle devices but the again I mean don't expect anything major this year or even next.

Having said that though I mentioned briefly in my prepared remarks, HUD or head-up display, I think it's a very exciting area.

I'm talking about very high-end head-up display.

People are using LCDs, (inaudible) lotonpoly LCD displays for head-up display.

Automotive is a good market, it's a big market, it's up and coming it's going to be here for long-term it's fantastic but I think they only occupy the low end and (inaudible)LCOS is the best is the most suitable for very high end and we actually have been working with certain leading edge (inaudible) customers for quite a few years but you'll appreciate when it comes to automotive, the pre-qualification, preparation for mass production takes very long time.

So I think it's a more a probably 2019 or on wards kind of story for Himax but when it happens, it will be very long term it's going to be very steady, very stable and I think it's very excited the kind of image we can generate on your windshield.

I've seen numerous (inaudible) demos and I think it really is super-exciting.

Our next question from the line of Sujeeva Desilva of Roth Capital, your line is now open.

On the 3D structured light solutions, the total solution versus the component parts the total solution for Android when can we expect that to ramp?

What's the ASP difference there and what king of ROI do you get on that it's the largest endeavor for you guys?

Right.

I mentioned in my prepared remarks, our total solution, which is targeting precisely the Android camp.

We are talking about early 2018.

So that means we are working, we have real customers we are working on designs, we are solving engineering issues, we are and what not so we are talking about stuff that are very real and our customers give us tremendous pressure for timetable they want it faster but, but I think we are way ahead of our peers because of the fact that we can offer total solution, I think that's most critical.

And as far as ASP is concerned now you have -- you have front scanning -- front sensing and rear sensing.

Front sensing by definition will be lower cost and rear sensing vision is more powerful i.e., higher cost.

The industry is likely to start with front sensing because it is lower cost and also engineering wise it's easier to handle.

So in terms of front sensing I would say for our total solutions, it's early stage so I may have to change my answer over time but I'm not talking about $15-20ish.

The thing is for a smartphone I think and then for rear sensing, I think it would be (inaudible) $20 plus.

The thing is we don't believe anything beyond $30 is sellable for a smartphone.

I mean you can force your customer into it, but then there is no demand.

So what's the point?

So our goal, our job is to squeeze the cost, squeeze the power consumption, squeeze the size while still making good profits and provide affordable price for customers.

So I think for your modeling purposes I would say start with $15 is probably a good starting point for front camera.

But I think we are thinking about [$15-ish to $20-ish].

Return on capital it is very high.

It depends on how you define it.

Now -- bear in mind -- I mean Himax comes from fabless business model and you'll be wondering we are announcing $80 million of CapEx and this is only Phase I and who knows how much it's going to be Phase II and so and so forth.

Right.

In my mind, we are still fabless why because we offer total solution.

So, it is only tiny part, the optics i.e., WLO and the active alignment because there's no readily available solution around the marketplace we've to develop it on our own, it is only this 2 parts that we have to put in our own CapEx.

I mentioned in my prepared remarks, we believe this is critical because it's all competitive advantage it involves a lot of confidential know-how and we are very much in the leading edge worldwide.

However, if you think about the real CapEx for the entire module.

You talk about the wafer foundry (inaudible) you have our logic chip which involves advanced semiconductor processor; CMOS image sensor also requires tremendous capital.

In our case we use BSI even more expensive, even on the receiver side, we have a effective view of the camera right so we provide sensor and our partner will provide lens which requires tremendous CapEx as well, even IR filter requires CapEx, not to mention laser, even finally-- the final assembly of the whole module also requires a lot of CapEx and actually it's quite laber intensive (inaudible) so we decide to outsource all these and we have actually put together and I'm very proud to say we have put together an A Team, involving Himax ourselves, I talked about in-house technologies in this 3D sensory total solution.

We have put together an A team covering laser, lens, IR filter, foundry and module-assembly houses.

I would say, if we cannot do it no one else can do it.

This is really an A Team of the world, we have to pick and choose the best in its own individual area to be in partnership with us and in due course we're probably announced one by one, our partnerships.

Last but not least, also the smartphone AP platform we have (inaudible) teamed up the best as well.

So that is why quite honestly we are seeing very little competition, and customers are just pushing us for better timetable and for the delivery.

So Sujee, I think I believe I've answered your question.

And our next question comes from line of Charlie Chan of Morgan Stanley.

Hi, Jordan, hi Jackie.

Congrats for a very strong quarter and upbeat 3D sensing outlook.

So my first question is regarding your non-driver IC business it is essentially doubling in third quarter.

So can you give us some of breakdown of the components in non-driver ICs the revenue mix and how does each of them grow in the third quarter.

I can't really give very detailed breakdown.

But the major, major, major growth area is WLO, naturally.

We highlighted starting from 2 quarters ago, I believe 2 or 3 quarters ago that we are having -- customer urgent demand, big capacity requirement (inaudible) and that's why we need to retrofit our -- existing headquarters to make room for the capacity.

And we also announced that if everything goes as planned start the third quarter will be mass production and everything is going as planned.

So, that additional CapEx, again I will be highly reluctant to talk about our ROI but it's quite good.

So that is a major contributor.

So that is the optics part and also another strong growth area is our RGB sensor, timing controller these are major growth areas.

The one area that has fallen a lot in the revenue is LCOS.

Last year at the same time, we did have this major customer launching and mass producing its AR devices and then they ended the production.

So if you look at year-over-year,that is a major loss.

So since the WLO is key growth driver in 3Q.

I remember kind of mentioned that those are non-driver ICs in particular WLO enjoys a better margin but why company margin only improved by 1 percentage point in third quarter, can you elaborate a little bit?

It's called -- it's something called weighted average, so it's weighed out by other things and also it's very early stage mass production.

We're talking about the first quarter only.

So you can image there will be little bit of learning curve and also the capacity is not yet utilized in full, although based on this anchor customer's behavior and projection, it will become fully utilized (inaudible) very soon.

So all these has affected the gross margin.

But having said that - it's indeed a good margin.

And my second question is regarding your raised CapEx for very strong 3D sensing demand.

So, I'm not sure if you can quantify the revenue contribution from 3D sensing in 2018.

And then if you can break down into Android smartphones and none Android smartphones revenue contribution, that would be great.

I mean non-Android you are talking about really only one customer in the world.

I cannot comment one or the other, whether or how we're our dealing with (inaudible) that customer, right?

Or even wether, right?

so I really can't comment on this particular question at all.

Overall, I think it's too early to make projections for 2018 because of the fact that we haven't even started mass production yet.

But I mean it's really easy math, if you care to do your modeling, I already disclosed about $15 up to $20 dollars of ASP.

And then you multiply that by the number of units per month and then you got the idea -- the answer is if everything goes as planned the revenue contribution will be tremendous, it will be a major game changer for Himax.

Himax will be looked at very, very differently next year, and the year after from now I believe.

But it really depends on how many units the customers want to ramp, what kind of model they are designing for their smartphone, the consumer reception, our capability to supply in time and so on and so forth.

But if you extend the timetable slightly longer then I think all these issue will be, will not, will not be issues anymore because I firmly believe, consumers will love this and our customer, we will have the first customer then we have the new customers and then both new and existing customers will add models and then very soon, you'll be talking about the penetration of smartphone using, having 3D sensing I believe and when that happens I think (inaudible) again you do your modeling, it will be tremendous growth.

Lastly, so you said that you had much ahead of your competitors in that, 3D sensing total solutions so may I know how many competitors in the market and if you can name 1 or 2 of them.

I know they are very behind but in the coming 3 to 5 years, who are the potential competitors or what type of company could be the potential competitors to your 3D sensing business?

Thank you.

I think there are numerous reports already it doesn't come from me, it comes from lot of speculation, rumors and reports that Apple is going to launch its new iPhone with 3D sensing.

So if anybody claims to have 3D sensing revenue this year.

I think it must be from Apple right because no other smartphones has this features just and I'm also speculative because I don't know when Apple is going to do that or not.

But there are lots of lots of reports that you guys know better, but starting next year it will be more interesting, because as I said, hopefully the Android camp will start to see the first one and the second one, hopefully next year.

And, so, in terms of competition, I will be reluctant to name names and provide comments to my peers.

It's fair to say that the number is very, very few.

So if you ask me to name 2 or 3 or even 4. I'm starting to have trouble because I don't know if I can have that many serious competitors by now.

But if I may try to address your issue from a different angle I think the combination so-called competition as far as I can see it comes from 2 kinds of backgrounds.

The first one is from algorithm because 3D sensing you need a 3D depth map generation it requires algorithm, right?

and this type of company tend to have good algorithm team probably a small team, very smart but they tend to lack the other software components, which are very, very challenging to achieve, not to mention the capability to integrate them.

I'm talking about NRI sensors, the WLO (inaudible) the laser and what not and so on so forth right.

So that is the drawback of the algorithm type of players.

The other type of player comes from components technology provider background and in one particular case they they've done that through a lot of acquisitions and with that they can -- on paper put together a so-called solution, but what they lack is algorithm, they don't have algorithm, and they don't have semiconductor background and that makes it very, very difficult for them to develop algorithm.

Because with algorithm you need to have the software and you need to turn that (inaudible) into silicon and that requires an IC design capability and certainly the sensors require IC design capability and so and so forth.

So the uniqueness of Himax is that we cover everything.

Our algorithm is now very much in the leading edge, our optics is in the leading edge judging by our mass production experience already.

Our NRI sensor in terms of quantum efficiency, it's actually more than double any of our peers so far in the marketplace right now.

So you are talking about the 3 most important building blocks of projector, receiver and algorithm.

Now with projector we have teamed up with the best laser provider already.

And on the receiver end, we have the sensor we have teamed up with the best lens provider already, and also IR filter provider, also the best in the world.

So we cannot do this on our own.

We will have to have great partnerships, and I've mentioned the concept of having to put together an A team.

I think, so now Android camp, I talked about we only, we only provide the individual components to select customers actually those are the customers who own their own algorithm, they probably even have their IT division in-house.

And they want to be the integrator themselves, and they being the customer, we cannot compete with our customers.

So for such very select small number of customers, we will deal with them on the basis of individual components.

We have got a lot, a lot of success, evidence for my remarks about our leading edge in the individual components as well.

Now if you put those very small numbers of customers aside, all of the other customers require total solution, including algorithm, including the whole thing and module assembly, otherwise they don't know how to get it started, and I think because of the complexity of integration I think in any foreseeable future this will continue to hold true.

And that is why I think we are in a very, very, unique position, have been able to offer the whole thing by luck and by design we have sensor, we have algorithm, we have optics and that is why -- that is almost 10 years of investment, now we are in this unique position right now.

Our next question comes from Jerry Su of Credit Suisse.

Your line is now open.

Jordan I just want to follow up on the 3D sensing, I think, because you have mentioned that you have team up with some of the very best of suppliers in the projector, the laser, or IR filter, lens, I just wondering for those components that you need to outsource, what is the percentage of their value add versus your value add in that $15 of module.

I cannot disclose the detail but in-house is bigger.

In-house includes (inaudible) which is expensive, the optics and the sensor which is also expensive and also I have to do my math, but we account for a very significant portion that's absolutely true.

Laser is also expensive less so for lens and certainly less so for IR filter and certainly module assembly which is labor intensive.

I still think we -- that also account for, for a meaningful amount of dollar in our build of material.

I cannot disclose the details right now, but our part it's actually significant in particular, well, actually we are, let me, let me just try to do it all over again.

Algorithm chip is expensive, it's a logic it's a rather complex logic chip, the sensor we have 1 mega and 5 mega sensor and it's actually much bigger pixel because of this particular implication compared to the RGB sensor you're used to.

There you talked about 1.1 micron or even 90 nanometer kind of pixel size over here is 2 micron pixel size so you cannot really compare apple-to-apple (inaudible) with that kind of sensor.

So sensor is also expensive, and optics per say is not that expensive but integration of optics together with laser involving active alignment and the sub module, the sub - the projector module process that also cost money.

We do the integration ourselves and active alignment ourselves, and lastly we also provide a laser driver for the laser in-house.

So we do have quite a big number of components provided in-house.

So generally for the algorithm and the sensor, the optics, active alignment and the laser driver all these are in terms of the (inaudible) BOM cost should be greater than those outsourcing parts, is this correct?

Yes.

Having said that though our ASIC chips and sensor, are also (inaudible) outsourced for manufacturing.

So it depends on how you look at it.

But I think that know how right?

Yes we, I mean for those If we get back to our business model which is fabless, right?

You will understand that.

The second question I want to shift to TDDI, you had mentioned that you're already developed the full HD+ TDDI but just wondering because the market trend now seems to be, I think HD+ is also seeing a very big pickup in later half of this year, and HD generally is a market probably has bigger volume because full HD has being cannibalized by OLED or other more high-end product.

So just wondering what's your progress on HD+ TTDI, and when can we see the revenue contribution?

HD+ TDDI's revenue contribution will be starting from Q1 next year.

Some shipping in this quarter right now, Q1 next year we have lined up with top name customers, certainly also second tier customers, but big name customers.

We are very excited very hopeful about that particular project.

And HD+ TDDI small volume production this year-- this quarter some more next quarter and next year.

I'm not sure I agree with that.

I think full HD+ is equally important even more important than HD+, than HD+ for us.

I mean, and in particular, full HD+ for branding customers they want ICs with ram so it provides better, much better ASP rather than (inaudible) ramless because power consumption is issues for top name customers, for second year customers probably for sake of costs they would -- they will use our [remedy] ramless solutions, but we are very, very much ahead in progress compared to anybody in the Marketplace in full HD+ TDDI in particular because the industry is switching to 18:9 aspect ratio and also we pioneered this what we called interlays (inaudible) design approach, which enables the bezel, or the boarder of the edge of the display to be very, very narrow and we're leading edge that is why today, (inaudible) with literally all in particular full HD+ literally all major kind of customers and quite a few big name end customers have adopted it, and again mass production is scheduled to be Q1.

And then a quick follow-up on the HD for the HD+, does your solution supports that less TFT mask, which is like for silicon like 6:7 base design?

Yes.

So for Full HD+ the advantage is bezel-less design, for HD+ the advantage is cost reduction as you said, it reduce the number of mask for our panel customers.

I'm showing no further questions at this time.

I'd like to turn the call back over to Jordan Wu for any closing remarks.

Thank you very much again.

As a final note, Jackie, our CFO will maintain Investor, marketing activities (inaudible) and attend future investor conferences.

We will announce the details as they come about.

Thank you and have a nice day.

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

This does conclude the program, and you may all disconnect.

Everyone, have a great day.