慧榮科技 (SIMO) 2017 Q2 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the Second Quarter Silicon Motion Technology 2017 Earnings Conference Call.

My name is Annie and I will be your conference moderator for today.

(Operator Instructions) Before we begin today's conference, I have been asked to read the following forward-looking statements.

This conference call contains forward-looking statements within the meaning of Section 27A of the Securities Act of 1933 and Section 21E of the Securities Exchange Act of 1934, as amended.

Such forward-looking statements include, without limitation, statements regarding trends in the semiconductor industry and our future results of operations, financial condition and business prospects.

Although such statements are based on our own information and information from other sources we believe to be reliable, you should not place undue reliance on them.

These statements involve risks and uncertainties and actual market trends, and our results may differ materially from those expressed or implied in these forward-looking statements for a variety of reasons.

Potential risks and uncertainties include, but are not limited to, continued competitive pressure in the semiconductor industry and the effect of such pressure on prices; unpredictable changes in technology and consumer demand for multimedia consumer electronics; the state of and any change in our relationship with our major customers; and changes in political, economic, legal and social conditions in Taiwan.

For additional discussion of these risks and uncertainties and other factors, please see the documents we file from time to time with the Securities and Exchange Commission.

We assume no obligation to update any forward-looking statements, which apply only as of the date of this conference call.

And I now would like to hand the conference over to your first speaker today, Mr. Jason Tsai, Senior Director of Investor Relations and Strategy.

Thank you.

Please go ahead.

Thank you.

Good morning, everyone.

Welcome to Silicon Motion's Second Quarter 2017 Financial Results Conference Call and Webcast.

My name is Jason Tsai and with me here is Wallace Kou, our President and CEO; and Riyadh Lai, our Chief Financial Officer.

The agenda for today is as follows.

Wallace will start with a review of our key recent business developments.

Riyadh will then discuss our second quarter financial results and provide our outlook.

We'll then conclude with Q&A.

Before we get started, I'd like to remind you of the safe harbor policy that was read at the start of this call.

For a comprehensive overview of the risks involved in investing in our securities, please refer to our filings with the U.S. SEC.

For more details on our financial results, please refer to our press release which was filed on Form 6-K after the close of market yesterday.

This webcast will be available for replay on our website, www.siliconmotion.com, for a limited time.

To enhance investors' understanding of our ongoing economic performance, we will discuss non-GAAP information during this call.

We use non-GAAP financial measures internally to evaluate and manage our operations.

We have, therefore, chosen to provide this information to enable you to perform comparisons of our operating results in a manner similar to how we analyze our own operating results.

The reconciliation of GAAP to non-GAAP financial data can be found in our earnings release issued yesterday.

We ask that you review it in conjunction with this call.

With that, I will turn the call over to Wallace.

Thank you, Jason.

Hello, everyone, and thank you for joining our earnings call.

As you know, overall NAND availability continues to be very tight.

While hindered supply has been improving gradually, continuing strong demand in cellphone, enterprise, and hyperscale data center [applications] continue to limit availability.

Hindered supply constraints have been affecting our module maker customers, hyperscale customers, and even NAND flash partners.

As a result, despite our strong pipeline of SD design wins with customers and growing portfolio of unique solutions, our ability to grow sales has been limited.

Our second quarter sales, therefore, came in slightly below our expectation and second half sales will also be impacted.

We believe this is a temporary issue caused by NAND industry supply transitioning to new 3D technology.

By the end of the year, we expect our growth to start to rebound.

And in a few quarters' time, as the industry supply accelerates, this headwind should turn to a tailwind.

Riyadh will discuss our -- about our financial performance in greater detail later on this call.

Let me talk more about NAND industry supply condition as this the biggest factor limiting our near-term growth.

As you know, NAND flash makers are already in commercial production of 64-layer 3D NAND flash.

For example, our controller already is shipping in the world first commercially available client SSD using 64-layer 3D NAND, Intel's 545 SSD.

64-layer 3D NAND components from Intel, Micron, Samsung, Toshiba and WD are now all commercially available.

This flash maker has been reporting smoothly improvement in manufacturing yield and industry supply has been gradually increasing.

However, initial 64-layer 3D supply remains inadequate.

And for the time being, supply continues to be tightly rationed.

Supply is limited not only by current manufacturing yield and current number of the fabrication tool sets but also longer production cycle time relating to producing higher layer count 3D NAND, specifically for wafer deposition and [action].

Additionally, NAND disks are also limited because flash makers are currently focused on producing lower capacity 256 gigabit 64-layer 3D components to maximize production yield and will start transitioning to manufacturing higher capacity, more cost effective, [5-trail] gigabit components when yields improved further.

Based on current constrained NAND supply, flash makers have been actually rationing their limited supply of flash and allocating more for their own enterprise [SSD].

At the start of this year, our flash partners reduced their client's SSD forecast and as they were redirecting their flash toward their own enterprise SSD . This rationing in favor of enterprise applications increased in the last few months.

But as you know, our NAND industry is very dynamic and recently, as the flash makers' supply has gradually increased, they have been increasing availability of flash towards third-parties building enterprise SSD in addition to their own enterprise SSD but at very high prices.

While our enterprise industrial SSD solutions are benefiting from the better availability, higher pricing is impacting our profitability in the short term.

Currently, flash supply remains inadequate for client SSD.

We are confident NAND supply will continue to improve over the next few quarters, and supply will likely accelerate as the manufacturing yield continues to improve, more tools are put into production and flash makers start transitioning 256 gigabit 64-layer 3D to 512 gigabit and introduce 4 bit per cell QLC 64-layer 3D.

We have been working with our flash partners on this and are targeting the commercial shipping of our [curiosity] client SSD controller to them in the first half of 2018.

Additionally, many of our flash partners have been making very good progress in developing 96-layer 3D NAND, both TLC and QLC.

And we expect these components and our client SSD controllers that support the next-generation 3D NAND to begin coming to the market by the end of 2018.

Finally, I also highlight that flash makers have been investing aggressively in CapEx and some recently disclosed that they plan to further increase CapEx this year.

It certainly seems demand for NAND is very strong, while our industry had been facing NAND tightness since May 2016.

This will reverse over time and supply will become plentiful and stay that way for a few years to come.

Structurally, the NAND industry has always been a cyclical industry, and we do not believe industry dynamics have changed.

Let me now provide an update of our key products beginning with our client SSD controllers.

As we have previously discussed, NAND flash tightness has been affecting our client SSD controller sales since mid-2016 when our module maker customer started complaining about NAND availability issues.

Nevertheless, our overall client SSD controller sales continued to grow in the second half of last year as NAND flash makers redirected their flash from multi maker to their own client SSD products.

At the start of this year, the NAND flash maker redirected more of their limited supply of flash to their enterprise SSD.

In the first few months this has been accelerating, directing even more flash away from client SSD and this is affecting our client SSD controller sales in Q2 and pushed our new projects and expected [RIM] to the fourth quarter.

Our pipeline of new products remains unchanged.

In fact, it has grown but timing of these new projects is not expected to ramp until later this year, driving a material rebound in our fourth quarter SSD controller sales.

Despite current NAND tightness, I'm pleased by the progress and execution of our team.

We made a solid progress in further diversifying our client SSD controller customer base to reach deeper into PC and other device OEM markets, and in upgrading our unique hardware plus firmware turnkey controller solution for both [SATA] and PCIe NVMe client SSD to further strengthen our industry performance leadership.

An example of this is the world's first commercially available client SSD using 64-layer 3D flash uses our controllers.

We enable our customers to rapidly come to the market using our latest generation, most cost-effective NAND.

We believe we are well placed to benefit when flash supply availability improves.

Last quarter, we introduced to you our enterprise SSD controller business.

This quarter, I am pleased to announce that we have added another OEM customer and also for enterprise-grade SSD for hyperscale data centers.

Additionally, our first project has been pulled in and we now expect to deliver commercial samples of our first enterprise SSD controller to our customers by before end of this year.

For both hyperscale customers, we are engaging directly with them.

Many of our enterprise SSD controller projects for our hyperscale customers are for open-channel NVMe SSD.

Open-channel SSD controllers are designed specifically for delivering not only high-data rate performance but, more critically, low and a predictable latency a new (inaudible) with current standard enterprise grade SSD.

Open-channel SSD controllers also enable whole devices to directly manage the NAND components, which enables significantly scalability to storage and creates flexibility to the [in-demand] software-defined storage architecture.

We are also adding this capability relating to software-defined storage with our recent acquisition of (inaudible).

Now let me turn to our eMMC and UFS controllers.

Sales of our eMMC controller grew in the second quarter as SK Hynix pre-built inventory, which was part of their internal planning decision, which we believe is related to their current upgrades.

Our new UFS controller program remains on track with our flash partners and we continue to expect initial shipments to our top 5 Android OEM to begin later this year.

In the near term, we expect UFS adoption by smartphone OEMs to remain low as NAND flash makers continue to price UFS at a significant premium relative to eMMC embedded limiting the appeal to smartphone OEMs.

These issues have also delayed application processor support.

I will now talk about our enterprise and industrial SSD solutions.

The ongoing tightness of NAND availability continues to impact our SSD solution business, both Shannon on the enterprise and Ferri industrial SSD limiting growth this year.

Without a doubt, had it not been for NAND supply availability issues that we are facing today, both Shannon and Ferri would be growing meaningfully this year.

That said, high-grade, high-density 3D NAND flash components are increasingly available for our enterprise SSD solution but at a very high cost.

Our channel design activity with Alibaba and other Chinese Internet companies, whether for their public and private cloud projects, continues to expand as our customers are seeing value in both our unique controller capability but, just as importantly, valuing enterprise SSD solution with specific customer design features.

In the second quarter, we started scaling our second-year program of customer design SSD for Alibaba's data centers.

Our highly customized single-chip VGA from form factor Ferri industrial SSD business has been performing very well.

Our portfolio of high mix, low volume long-term projects range from commercial and industrial to automotive and networking applications.

With the customers ranging from Japanese OEMs to U.S., European and Chinese OEMs, we are currently facing a situation where we have significant amount of sales that are not going to fulfill because we are not able to procure enough NAND flash.

Overall, while we continue to face headwinds from ongoing but temporary NAND flash tightness, we are excited by our strong pipeline of new design wins and backlog across all of our major product segments.

We are, however, very well positioned to rebound when more new NAND supply continues to come online.

I will now turn the call over to Riyadh to discuss our financial performance and outlook.

Thank you, Wallace, and hello, everyone.

I will summarize our financial results and outlook.

Before I begin, I would like to reiterate that our comments today will focus primarily on our non-GAAP results unless otherwise specifically noted.

A reconciliation of our GAAP to non-GAAP data is included with the earnings release issued today.

Our second quarter revenue increased 4%, sequentially, driven by the expected ramp of new SSD solutions projects and growth of our eMMC controllers.

Our embedded storage product sales continued to be roughly 80% of our total revenue in the second quarter.

Let me provide additional color on our key products.

Our client SSD controller sales declined by almost 10% sequentially as temporary high-NAND flash prices and high-client SSD prices slowed adoption rates and flash vendors redirected their limited flash supply from client SSDs to enterprise SSDs.

Based on our customers' rolling forecasts, we expect our client SSD controller sales to decline further in the third quarter before rebounding materially in the fourth quarter.

Beyond this year, we remain confident about rapid growth for our client SSD controllers.

The cost of these new and next-generation 3D NAND components from 64-layer to 96 layer, from TLC to QLC, are significantly cheaper than previous generations of NAND.

Over the next 3 to 5 years, the majority of client HDDs will be displaced by increasingly less expensive and higher performing client SSDs.

And as the industry's leading merchant client SSD controller supplier, we should continue to benefit from this multi-year trend.

eMMC controller sales grew nearly 5% sequentially as SK Hynix prebuilt inventory as part of their internal flash capacity planning.

Our eMMC controller sales will likely decline in the second half as our flag partner is planning on allocating flash from mobile towards SSDs.

Longer term, we believe our eMMC plus UFS controller sales should be stable.

We are working to reduce our customer concentration risk by increasing business with our other NAND flash partners.

Additionally, slowing smartphone sales are being offset by new categories of growth applications from smart TVs running Android OS and Chromecast TV dongles to IoT devices like Amazon Alexa speakers and other connected devices.

SSD solutions increased over 50%, driven by strong growth from both our Shannon hyperscale SSDs and Ferri industrial SSDs.

Our Shannon SSDs are benefiting from our second-year Alibaba project ramp.

Our Ferri SSDs, on the other hand, are benefiting from the snowball effect of accumulating a growing critical mass of diversified, industrial, commercial, automotive and networking OEMs.

Overall, SSD solution sales should grow further in the third quarter and peak for the year.

Longer term, we are seeing very strong demand for our unique Ferri SSDs from a very broad set of customers and applications.

And for our Shannon SSDs, hyperscalers will continue to meet highly customized enterprise grade SSDs, which we are very good at designing.

For the full year, our client SSD controller sales should be flat, eMMC down slightly and SSDs solutions down by less -- SSD solutions down by less than previously guided.

Our second quarter gross margin decreased to 48.7% from 51% in the previous quarter, largely due to higher-NAND costs relating to our SSD solutions.

Because NAND flash component prices have increased sharply, the lag between our earlier SSD pricing agreements with customers and subsequent NAND procurement arrangements have led to margin compression.

This margin compression will worsen in the third quarter as our average cost of NAND components in inventory increases before reversing at the end of the year when pricing arrangements are reset.

Margins from all our controllers, on the other hand, remain stable sequentially and will remain stable for the rest of the year.

Our second quarter operating margin decreased to 23.9% from 26.4% in the previous quarter, primarily due to lower sales and lower gross margins.

In the third quarter, operating margins will fall further as sales fall and gross margins contract further from SSD solutions accounting for an increasing proportion of our sales and SSD solution margin being impacted by the use of expensive NAND flash components.

This issue will reverse in the fourth quarter when our client SSD controller sales rebound and our Shannon program for Alibaba, this year, ramps down.

Our total headcount this quarter increased to 1,153, which was 4 more than the first quarter.

Our effective tax rate in the second quarter was 22%, lower than the 27% in the first quarter but still higher than our long-term model tax rate of 20%.

We anticipate our effective tax rate to remain elevated for 1 or 2 more quarters due to the mix of operating companies that are profitable and incurring tax expenses with entities generating pretax losses.

Our second quarter EPS was $0.71, slightly higher than last quarter's $0.70.

Total stock-based compensation in this quarter was $1 million, lower than the $3 million in the previous quarter due to the seasonal timing of RSU awards and consistent with timing of past years.

For the full year, we expect stock-based compensation in the range of $14.4 million to $15.4 million.

Let me now talk about a few key items from our balance sheet.

We had $306 million of cash, cash equivalents and short-term investments, $1.6 million more than in the previous quarter and $86.9 million more than a year ago.

Our business generated less cash flow this quarter because of the timing of working capital movements, specifically, higher accounts receivable related to the timing of sales during the quarter and higher inventory levels from temporary accumulation of NAND components.

In May 2017, we paid $7 million of dividends to shareholders, the third $0.20 per ADS quarterly installment of our annual $0.80 per ADS dividend.

For the third quarter, we are expecting revenue to decline 3% to 8% due to temporary headwinds to our business from NAND tightness and flash maker NAND allocation decisions.

Gross margin in the 45% to 47% range, a temporary decline from our normalized gross margin range due to higher NAND component costs in the short term relating to our SSD solution sales.

Operating margin in the 19% to 21% range as a result of declining sales, lower gross margin and stable operating expense.

Our current order forecast and backlog for the fourth quarter gives us strong confidence that revenue will rebound and gross margin will return to more normalized levels.

New projects that were expected to ramp in the third quarter have been pushed out to Q4 to coincide with better NAND availability.

Before we turn the call over to Q&A, I would like to comment about our share repurchase announcement and management's intention to buy our shares.

As we said throughout our call today, current softness in our market is temporary.

The NAND flash industry has always been very dynamic and cyclical.

We are going through a period of NAND supply tightness.

NAND tightness will disappear when more NAND supply comes online.

The very strong headwinds that our business is experiencing will likely become a strong tailwind in a short period of time.

During this period of time when our business is experiencing headwind and our share price is underperforming, this presents a unique opportunity for our company to repurchase our shares and for management to accumulate more of our shares at valuations that are attractive in our opinion.

Our board has authorized a new $200 million 1-year share repurchase program.

Additionally, our CEO, many of our fellow executives and myself have notified our company that we intend to buy our shares subject to compliance with relevant laws regulations and policies.

This concludes our prepared remarks.

We will now open the call to your questions

(Operator Instructions) Your first question comes from the line of Mehdi Hosseini of Susquehanna International.

Two follow-ups.

Wanted to better understand the impact of your largest customer, especially with their ramp of 3D NAND.

Based on their quarterly earnings, it was very obvious that they're running late.

Their 72 layer is also late.

And I am -- I want to better understand how that particular customer and unavailability of their 3D NAND is impacting.

And to what extent are you trying to improve your exposure to the other NAND manufacturers?

And I have a follow-up.

Mehdi, our largest customer is a customer of -- buys our eMMC controllers.

Here, we had a fairly decent first half but our eMMC sales will decline in the second half of the year as our customer allocates flash from mobile applications to SSD.

So the full year, our eMMC sales will decline a little bit compared to last year.

Sure.

But if they're running late with their 3D NAND project, doesn't that constrain their ability to supply inexpensive NAND into the client SSD market?

100% of our current eMMC project with (inaudible) was 2D NAND.

So although they're transitioning to the 3D NAND, the majority will be reserved for SSD, our high-end UFS product line.

So our current eMMC product will not be affected even if they delay or slow down transition from 2D to 3D NAND.

Okay.

And then you mentioned the 4 bit per cell and the QLC, and I'm under the impression that you will start sampling that in the second half.

How will that impact your blended pricing?

Is that going to have an impact?

Or is this just another transition in 3D NAND that would require your customers to upgrade their controller?

We mentioned about 4 bit per cell TLC availability, we start to see samples at this moment.

We believe it will be commercialized in the first half of 2018.

We believe that will help the big growth of the whole NAND industry and it has nothing to do with our gross margin because we're selling controllers, manage (inaudible) flash, shipping SSD products from our NAND OEM customers.

And we believe that we're helping to improve the overall industry from the NAND supply and into the big growth.

Mehdi, let me also add that, generally, for all of our new controllers, when we launch new controllers, our ASPs are always initially always a little higher but they gravitate towards our segment average over time.

Your next question comes from the line of Mike Burton of Longbow Research.

I wanted to follow up on mobile.

I missed your comment on UFS.

Are your top customers still on track to ramp in the second half?

And have you seen any movement at Hynix to doing an eMMC controller in addition to their UFS one?

I think our UFS program remains on track and we are waiting for broader AP support and smartphone design using UFS.

Still only high-end AFC APs supporting UFS.

Until broader, more mainstream APs supporting UFS and UMCP, the market for UFS will remain limited to Samsung, Qualcomm high-end platform.

We understand some OEMs and some NAND makers will develop a new eMMC for high-end 3D NAND and -- but that's their decision.

We believe our current eMMC also already support certain 3D NAND for the customer.

Okay.

And then also regarding the inventory correction at that customer.

Is your expectation for your full year guide that we should see sequential growth in mobile coming back in Q4?

No.

You should expect our mobile -- our eMMC sales in the second half of the year to be down compared to the first half of the year.

Q3, our large customer, they built inventory, prebuilt inventory and we believe this is relating to their fab equipment changeover.

But for the second half of the year, specifically, we believe that they're going to be reallocating their flash more towards SSD.

So this will impact their procurement of eMMC controllers for the second half of the year.

So for the full year, you should expect our eMMCs to be down slightly

Okay.

And then just you mentioned the mix effect, Riyadh.

I was wondering if you could help quantify what kind of impact that's had on margins to you in the near term here as we look to that potentially becoming a tailwind as 64-layer starts to ramp?

And then any other signs of increasing competition that's affecting pricing from either Marvell or internal solutions?

In the near term, our lower gross margins are a function of our SSD solutions increasing as a percentage of our business.

And these products have lower margin due to higher NAND flash costs.

By the fourth quarter, we should start seeing our SSD controllers rebound strongly and our SSD solutions decreasing as our Alibaba products -- projects ramp down, which should lead to higher gross margins in the fourth quarter.

As flash supply improves and our mix of business improves with more controller sales in the fourth quarter, we are confident that our gross margin will improve.

When flash supply is readily available and prices come down with new generation 3D in mass production, we expect our SSD solution's gross margin to normalize again.

In addition, we do not see any meaningful changes to the competitive landscape for SSD.

We have historically dominated the client software controller market and we continue to lead the market.

Our same generation PCIe product lines are coming online later and we have a very strong design win pipeline and we're going to see the production from Q4 moving to 2018.

The next question comes from the line of Rajvindra Gill of Needham & Company.

Riyadh, a question on the forecast for client SSD for the year.

I believe you said it was going to be flat, if that's correct.

Just wanted to get an understanding of the math there because, if I remember correctly, client SSDs in Q1 was down 20% sequentially and, in the second quarter, it was down 10% sequentially.

In Q3, you're guiding it to be down slightly.

So in Q4, it's going to require at least 50% sequential growth, if not higher, to get to the flat year-over-year.

So just wondering if you could maybe provide a little more insight on that ramp in Q4 and if the math is right or not?

Raji, I'm not quite sure about your math but for the full year, we're expecting our client SSDs to be flattish compared to last year.

It will be down.

It was down in the first few quarters but we're expecting a very strong rebound in the fourth quarter leading to flattish client SSD sales for the full year.

Okay.

And what gives us kind of confidence in terms of the visibility?

Because the tightness of the NAND flash, we saw signs of that at the beginning of the year.

It hasn't loosened and there's expectations that it will start to expand but that hasn't come to fruition in the past.

So I'm just wondering, it's a pretty high bar -- sequential bar in Q4 for the client SSDs to get to flat.

So I know you talked about based on customer rolling forecasts, any other insights or elaboration would be helpful.

I think we have a very strong SSD design win pipeline through our second-generation PCIe with a major NAND maker as well as PC OEM.

And so even we believe NAND will gradually improve for the supply, but our client SSD is going to rebound strongly from Q4 this year, no matter if NAND supply improve or not improve because our NAND partner and OEM customer, they're going to secure the NAND, support the major launching new PCIe program and new SATA program with the new 3D NAND

Okay.

Got it.

So we should see some new PCIe programs, 3D programs that, I would presume, would come with a higher ASP as well.

That's correct.

Starting to ramp in Q4.

Yes, that's correct ...

Is that fair to assume that we are going to get kind of a richer ASP mix with these new PCIe designs?

That's correct.

As you know, for all of our new products, ASPs and margins are always better.

But over time, they will converge towards our mean.

Okay.

And last question for me and I'll step back.

On the diversification in the eMMC business away from the top customer, can you maybe talk about the progress on that front?

It seems like you've demonstrated excellent quality and dominance in the eMMC merchant market and it would make sense for other flash vendors to start to use your controller.

Wondering if maybe you could give us thoughts on that?

And how would you balance that with some of your top customer -- with the top customer as well?

If you're going to, perhaps, diversify, is there a conflict there?

Or how you think about that?

First of all, I want to say we stay very, very close (inaudible) with SK Hynix and nothing changed with our business.

For eMMC, we're seeing we're going to stay flat.

How it grows is as the market grows in smartphone.

We also diversify and with the new design with several key OEM customers and we can now name the other potential NAND maker because we believe when smartphones become mature, when the solution was stable and some NAND partners going to also to third-party with our eMMC controller because we are more stable, more cost-effective and fitting their needs.

NAND makers tend to go to high-performance and higher -- newer generation UFS development and competition, I think that's the strength with the 3D NAND.

But a majority of eMMC is going to stay with the 2D NAND because the density has so many varieties and that will be more stable with our solution.

And we do see our eMMC growing, expanding from smartphones, go to set-top box and moving to automotive and networking application and that becomes a more desirable and for a more balanced model for us to grow in long term.

Next question comes from the line of Suji Desilva of Roth Capital.

On the 3D NAND, just to understand the supply demand dynamic here.

You talked about availability improving.

Do we get to a point where there's a more normal supply/demand situation?

If so, how far away production yields are kind of -- or is it a situation where it remains in tight supply for several quarters to come?

I think flash makers are making good progress in ramping up new 3D NAND.

And some are even accelerating CapEx investments to ramp new capacity sooner.

So we are optimistic about the long-term NAND supply growth.

And flash makers, over the past week, have talked about how their yields and output for 3D NAND are ahead of expectations and some are even talking about next generation 96 layer and QLC NAND to try even faster bit growth next year.

I think this year, because a majority of the industry bit growth will be back half-loaded, but that positions 2018 to be a meaningful year in terms of supply growth.

So we are confident in maybe early 2018 or mid-2018, the supply and demand will become more balanced and made as a normal growth.

Okay.

And then you talked about eMMC trying to diversify from phones.

Can you tell us what percent of the eMMC business is not smartphones?

I just don't understand where you're starting out here.

Yes.

It's roughly about 10% based on our product sales.

About 10% of our eMMC sales are going into non-smartphone applications.

Okay.

And the last question.

Can you give us a similar sort of split notion of the SSD solutions business, kind of how much is the Shannon business going into the hyperscale and how much is (inaudible) just a baseline up to this point?

No, we don't break out -- break that down.

The next question comes from the line of Mike Crawford of Riley.

The SSD solution side, you're talking about new pricing arrangements that you're expecting to reset at year-end.

Can you talk about the timing and duration of these type of agreements?

It varies quite a lot.

We have a diversified -- very diversified set of customers on our Ferri SSDs and on our Shannon side, while more concentrated, we have a lot of various different type of projects with various -- lots of different types of programs.

And so there's going to be a lot of variability.

But what we've had is pricing arrangements that are set ahead of NAND procurement.

And with NAND flash prices increasing rapidly, it's been affecting our margins, causing margin compression.

But these should start changing as the year progresses, as we get closer to the end of year when a lot of the arrangements reset.

For example, some of our Ferri products have been designed 3 years ago with the major automotive makers.

And through the designing, through the auditing, everything, we signed a supply agreement and secured pricing.

And now, we're facing the NAND price up.

So it's very hard for us to change the selling price.

So there are many different examples like this.

So we have to waiting for when the NAND supply becomes smooth, more balanced, I think that our SSD solutions gross margin will go to a normal situation.

Okay.

And then the other question relates to the buyback.

I guess part A, is it your intent to use the full $200 million authorization over the next 12 months.

And then part B would be that even if you do that, plus the $30 million a year dividend you're paying, you're still -- about this time next year would have about $150 million of cash left on the balance sheet.

And I'm wondering if, with the tightness in the market that might be affecting other competitors, whether you see any distressed opportunities to buy capacity or capability from other players who might be stressed?

Mike, could you rephrase your question, especially first about the share buyback?

Sure.

So the first part, Riyadh, is if it is the company's intent to use, to execute on the full $200 million 12-month authorization.

So Mike, our board makes the decision is really trying to show that we have confidence our company will rebound and back to the growth mode.

Of course, you also make very -- we also consider every (inaudible) availability and we might consider more things.

But the $200 million (inaudible) new cash coming which should have sufficient cash for us to make other choice or directions.

Let me also add.

From a finance perspective, Michael, is to buy as much of our shares as cheaply as possible so that we can maximize the accretion to our company purely from a perspective of finance relating to a share repurchase.

Okay.

And the second part of that question just related to, even if you did execute on the entire buyback, you still would have, by my estimate, some $150 million left of net cash on the balance sheet a year from now after paying the dividend, after the $200 million dollar buyback.

And given, today, that others must be feeling some pain given the supply tightness in the market, whether there were M&A opportunities that might be opportunistic at this point that you're looking at.

Well, Mike, we actively look and evaluate M&A opportunities.

But as you know, we don't have an extensive track record of actually pulling a trigger and buying companies.

As Wallace mentioned earlier in our call, recently, we acquired a small company -- a software company that will help us on software-defined storage solutions.

So these are an example of how we deploy capital for M&A perspective -- purposes.

Next question comes from the line of Charlie Chan of Morgan Stanley.

So my first question is to reconcile your SSD shipment growth versus the industry growth.

Because we all know that even the supply is tight but client SSD penetration is still going up, right?

From net -- last year was 40% this year, maybe to 45%.

So this is still like 10% industry growth, right?

So how do we reconcile this disparity between (inaudible) industry growth?

Is that your customer is losing share in client SSD or you are losing share to other competitors?

Charlie, we don't believe we're losing market share.

What we've seen is price of NAND have gone up a lot.

The price of SSDs, client SSDs have also gone a lot.

And so we're seeing adoption rate -- adoption of SSDs by PC OEMs and other device OEMs decelerating.

So the use of SSDs this year, the growth of SSDs this year is significantly slower than what we saw last year.

So this, I think, is a very important factor for our industry.

With upcoming NAND, especially with the 64 and the upcoming 96, the price of NAND, the cost of NAND will continue to come down.

And so longer term, we're very confident that we'll get back on track and the adoption beginning to accelerate again when these cheaper components come around.

But for the time being, with the high price of NAND, high price of SSDs, adoption has been impacted and it's affecting everybody across the board.

Okay.

And my next question is regarding your OpEx, right?

So the new revenue guidance is down single-digit.

How about OpEx in terms of absolute dollars, it's going to decreases as well?

OpEx for upcoming quarter should be quite stable compared to the previous quarter.

So for the full year, should we assume it's still up slightly?

For the full year, we'll be up slightly, yes.

It's important to note that while we would like to dial back our OpEx, we have a lot of very important growth projects like enterprise SSD controllers and other projects that we continue to need to invest aggressively if we want to hit our longer term growth targets.

Okay.

Understood.

So lastly, also a follow-up from a previous question from other investors.

So that the NAND supply will increase comments, right?

So I think at the beginning of the year, you provide guidance based on your assumption from your source, right?

So this time around, what can give you better prediction about the NAND supply [ramp up] timing?

For our rebound of our client SSD controller sales, the guidance -- the color that we are providing is based primarily on the rolling forecasts of our customers as well as backlog of orders.

And these have increasingly been inching up for a rebound in the second half of the year, especially in Q4.

So Charlie, let me add.

Regarding the NAND sourcing for our enterprise SSD solution and we're aggressively moving in a different direction.

And some NAND makers maybe they are -- now within some main NAND makers, they are willing to sell the NAND to us because they want to enter enterprise industry.

So we are working with them.

And through the price transition, we believe when NAND supply becomes better and pricing going down.

We are working on the Q4 pricing with several NAND makers at the moment.

Yes.

So for that big single customer, Alibaba, do they still consign the NAND or will it become the norm that you will buy (inaudible) NANDs for Alibaba going forward?

So we have 2 different business models.

They will purchase NAND consigned to us.

We can provide a total solution.

But even in the tight NAND supply condition, even Alibaba will have difficulty to procure their NAND in a full scale.

Charlie, for this year, we are procuring a lot more flash.

The proportion of our business with Alibaba where we're procuring the flash, this year, that proportion is much higher than what it was last year.

A big part of our business with Ali involves our company procuring flash for us to build SSDs for them.

Okay.

Got it.

So hopefully, your business can recover significantly in fourth quarter?

We believe so.

Yes.

The next question comes from the line of Donnie Teng of Nomura Securities.

My first question is to follow-up on Charlie's question.

As we know, this year, (inaudible) is also slowing down due to the tightening supply.

But the industry growth is still like 10%, 15% year-over-year.

So you already gave a guidance that client (inaudible) sales to be flat, but you also mentioned how you did not lose market share.

So can we say that could be mainly due to you have some new products not yet launched like PCIe or 3D NAND which carries higher ASP while your legacy products could face some ASP erosion in the maybe first half or in the first 3 quarters?

Donnie, think what you'll see is a lot of third-party analysts dial back -- dial down their forecasts, their client SSD unit forecasts for this year.

Some of the numbers from earlier this year are way too optimistic.

And so what you'll see is some of the analysts beginning to take down the numbers as the year progresses.

Okay.

Got it.

So basically, we should not need to worry about any competition according to our current visibility, right?

That's correct.

Okay.

My second question is regarding to your fourth quarter visibility.

So in terms of the ranking of the visibility, which product line you will have higher confidence?

And how should we think about any potential risk in the fourth quarter?

We're fairly confident about our -- all of -- pretty much across-the-board for all of our key products, whether it's client SSDs, eMMC or SSD solutions.

They all have fairly good -- for all our products, we have a good mix of high-quality businesses as well as more risky prospects.

But overall, feel fairly good about all our key products for Q3 and Q4.

Yes.

We have very strong, very broad, Ferri SSD product PO.

So unfortunately, we can only secure around 20%, 25% of the NAND.

If we can secure more NAND, I think we can do better.

Okay.

Got it.

So I think, generally, the risk should be mainly from like whether NAND supply to be still tight, which -- longer than your expectation?

Or whether the end demand could see any weakness in fourth quarter or coming quarters, right?

So it looks like...

We already considered all the factors.

We believe we already use a [conservative] NAND supply situation to give our guidance for the whole year.

Our last question comes from the line of Tom Sepenzis of Northland Capital Markets.

So clearly, revenue is going to snap back Q4.

Can you talk about gross margins and whether that's going to snap back simultaneously or if that will lag by a few quarters as the pricing of the NAND takes a while to come back down?

In Q4, our gross margins should snap back.

It's going to be coming from 2 major factors.

One is our higher gross margin client SSD products will rebound materially in Q4, while, at the same time, our lower gross margin SSD solutions ramp down as many of our projects, including the major ones with Alibaba, concludes for the year.

Bear in mind, our SSD solutions this year is further impacted by the high cost of NAND that we didn't have in past years.

And in addition, in certain contract price for enterprise SSD solutions will be renegotiated in the beginning of 2018.

So we believe our gross margin will be rebound strongly reasonable from Q4 to next year.

Great.

And then in terms of the revenue decline here in Q3, can you just qualify exactly what's driving that?

I mean is this the shift to enterprise, mainly, from client on the SSD side?

Or is it the mobile business?

I just missed the explanation for why we're seeing an actual decline sequentially in units in Q3.

In Q3, NAND will continue to be very tight and this is leading to temporary rebalancing of NAND inventory by the major suppliers and leading to them dialing further down their client SSD procurements in order to have more NAND for enterprise data center applications.

That does conclude our Q&A session for today.

I would now like to hand the conference back to Mr. Wallace Kou, President and CEO.

I would like to thank all of you for joining us today and your continuing interest in Silicon Motion.

We will be attending several conferences in Asia, U.S. and London this quarter.

Details of these events will be available on our website.

Thank you, and goodbye for now.

Thank you.

Ladies and gentlemen, that does conclude our conference for today.

Thank you for participating.

You may now all disconnect.