慧榮科技 (SIMO) 2016 Q4 法說會逐字稿

Ladies and gentlemen, thank you for standing by and welcome to the Silicon Motion Technology Corp fourth-quarter 2016 earnings conference call.

(Operator Instructions) I must advise you that this conference is being recorded today, Tuesday, January 24, 2017.

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 risk 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 recent 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 any changes 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.

I would now like to hand the conference over to your first speaker today, Mr. Jason Tsai.

Thank you.

Thank you.

Good morning, everyone.

Welcome to Silicon Motion's fourth-quarter 2016 financial results conference call and webcast.

My name is Jason Tsai.

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 some of our recent business developments.

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

We will then conclude with Q&A.

Before we get started, I'd like to remind you of the Safe Harbor policy which was read at the start of this call.

For comprehensive overview of the risks involved in investing in our securities, please refer to our filings with the US 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.

2016 was an outstanding year and we end the year with a strong fourth quarter.

Our fourth-quarter revenue of $144 million was the high end of our expectation and is a 47% increase from the same period a year ago.

For full year 2016, revenue increased by 54% to $556 million, a record high.

Our fourth-quarter operation margin expanded to 30% and earnings per share were $0.95.

For the full year, EPS increased 73% to a record $3.34.

Riyadh will discuss our financial results in greater detail later on the call.

The need for more advanced controller technology has been evident now for many years as NAND's improving affordability [succumb] to the cost of becoming increasingly weak and difficult to manage.

As the industry moved to 3D NAND and layer count increases from 30 and 40 layers to 60 and 70 layers this year, in nearly 100 years (inaudible) for sale NAND over the next one to two years.

Unique new problem will continue to arise that will push the limits of controller technology.

As 3D NAND layer count increases, data corruption will get worse from time, usage, and temperature-induced data retention issues.

The effect of [secretary] noises and data rewrite interference increasing data corruption issues will require better controller technologies.

Our expertise is managing these and other NAND flash component issues in different applications with solutions optimized for performance, load power, cost, and ease of use.

Our customers see our solutions are unique and unparalleled in the industry.

We are collaborators and partners of flash makers and we are able to see and understand the inner working of their new NAND technologies years before they are commercialized.

This unique role and position we play in the industry provide us with a long-term visibility of NAND technology trends and control requirements.

Only by being a trusted partner to the flash makers are we able to help resolve their NAND technical issue so their next-generation NAND can be commercialized sooner.

Let me now move on and talk about our key products.

Our highly-customizable hardware plus firmware turnkey clients the controller sales grew [175%] in 2016 and where one of our key growth drivers last year.

Sale of RFID controllers grew further in Q4, a nice quarter of rapid sequential growth.

And our SSD controllers have scaled to where they are one of our two largest product lines.

We are now the clear emerging market leader with a market share of around 30%.

Our primary customers are the NAND flash makers, specifically Intel, SanDisk, and Micron.

And we also sell to significant majority of leading (inaudible) makers.

Our client SSD controllers are used in a wide range of applications from PC to data center and channel markets.

SSD, using our controller, can be found in the PC of all top five Wintel PC OEMs.

Based on our discussion with all of the NAND flash makers, we believe that it is more than likely that in a few years' time most of their client SSD controller requirements will be provided by a merchant controller supplier like Silicon Motion.

The majority of our growth last year was driven by our SATA SSD controller for managing 2D TLC NAND.

For 2017, we expect our new products, primarily our new PCIe, ME, and ESD controllers to drive our growth.

In the second half of last year, we rolled out our first PCIe controller for Intel's 3D NAND.

This year we are expanding our PCIe and the ME product family with additional solutions offering faster performance as well as lower cost products such as (inaudible) and single-chip PGA SSD.

Like our SATA controllers, most of these new products are targeted for PC OEMs, but some are targeted for non-mission-critical data center applications.

While we have been shipping SSD controllers for data center applications, bear in mind that these are computing-grade SSD controllers.

Based on our business with the hyperscale data center and enterprise customers, in particular cloud operators, our customers feel increasingly comfortable with replacing HDD with SSD.

The total cost of ownership continues to decrease and have reached breakeven on lower-cost versus the HDD.

Additionally, we are now gearing up to enter the enterprise-grade SSD controller market because the addressable market for these controllers are expanding rapidly, with growth coming from both expanding demand for high-performance enterprise and hyperscale applications and also from displacement of mission-critical HDDs by SSDs.

We have already retooled some of our client SSD controllers to repurpose them for enterprise-grade applications, initially for our own channel SSD solution.

Retooling include special-purpose firmware that have been specifically architected and optimized for data center application requirements.

We are now also developing our own enterprise-grade controller based upon our deep understanding of both NAND flash and data center requirements and target introducing this controller to customers next year.

I will be talking more about this activity in the near future.

We finished 2016 with a broad portfolio of high-quality projects with our OEM partners.

That will be our foundation for continuing strong growth in 2017.

Currently, we have 20% to 30% more SSD controller projects with OEMs than a year ago.

Our client SSD controller will continue to be our fastest-growing product line.

Let me now turn to our EMC and UFS controllers.

2016 was a phenomenal year for our (inaudible) EMC controllers, growing by about 50% and continued to be as big as our client SSD controller business.

Our growth came from a number of different factors, including our primary NAND flash partner gaining share and then focusing on fast-growing Chinese OEMs such as [Opal] and [Vebos].

The TAM for EMC enlarging as ultra low-cost smartphones and tablets began using EMC instead of their NAND and government subsidize in China for 4G smartphone spurring upgrades.

In 2017, we expect our EMC plus UFS sales to grow at least as fast as overall smartphone market growth.

Currently a significant majority of all Android smartphones from low-cost mainstream to high-end smartphones are using EMC.

Most of the EMC is (inaudible).

Our new EMC controller this year will be focused on bringing more cost-effective high-capacity 3D NAND to mobile devices, as well as (inaudible) with higher performance DDR4 mobile DRAM in eMCP.

We are also seeing EMC usage (inaudible) to smart TV, set-top boxes, and new classes of IoT devices.

The market for embedded memory remains dominated by EMC today and we expect the use of UFS to remain limited to the very high end until the end of 2017 with the rollout of [AP] that support uMCP at the end of this year it is very likely that UFS and uMCP will become a more material part of the mobile embedded memory market in 2018.

Our UFS 2.1 controller was launched with a flash partner late last year and they are assembling today with leading smartphone OEMs.

Separately, we expect our UFS controller for SK Hynix uMCP solution will launch in late second half of 2017 coinciding with the timing of uMCP support from various AP providers.

UFS expected to be small but still contributing meaningful revenue to us this year, growing more significantly in 2018.

Finally, I will speak about our SSD solutions.

Our SSD solutions also had a phenomenal 2016.

SSD solution sales grew over 3 times from the previous year, driven by both our customized Ferri industrial SSD and channel hyperscale SSDs.

Our Ferri BGA single-chip industrial SSD grew as we accumulate a growing portfolio of long-tail sales to OEMs in Japan, but also expanding in Europe and the US with applications ranging from multifunction printer systems to automotive and server applications.

These FerriSSD sales will expand further in 2017.

Our channel very high-performance hyperscale SSD grew as our large Chinese hyperscale customer placed very large orders for their data center in 2016 and have awarded us with new projects that will start scaling in Q2 2017.

However, our channel sales in 2017 will be affected by flash variability as our customers and our company are both unable to secure commitment from sufficient supply of high-grade NAND at reasonable terms for our products.

Based on our discussion with the flash makers, we are increasingly optimistic about the quality and timing of new 3D NAND supply.

We believe the supply of 34- and 72-layer 3D flash will increase significantly in the second half of 2017 as the yield and manufacturability improve and new capacity comes online.

However, flash availability will remain tight throughout this year, especially as the increasing supply coincides with a strong seasonal demand for flash in the second half of the year.

With tight flash supply expected to remain throughout 2017, NAND flash vendor will continue to seek to capture more economies by prioritized solution sales ahead of wafer sales.

Our channel SSD are being affected by this situation.

This situation may improve in second half of 2017 as more supply comes online.

Overall, 2016 was a record-breaking year for Silicon Motion and we finished the year strong, both financially and with a solid pipeline of projects that will grow in 2017.

Our growth in 2017, however, will be constrained by limited flash variability for channel SSD.

Additionally, our NAND flash partners could potentially grow faster if they had more NAND today to meet the current strong demand.

The NAND industry remains very dynamic.

We expect significant accounts for incremental 3D flash supply to come online in the second half of 2017.

When we and our customers have higher certainty about the rollout of a new supply, we may look to revise up our guidance.

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

Thank you, Wallace, and hello, everyone.

I will summarize our financial results and provide our guidance.

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

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

Our Q4 revenue decreased 9% sequentially as client SSD sales growth was offset by seasonally weak eMMC sales and SSD solutions project transitions.

Our Q4 revenue is 47% higher than the same period last year.

Our revenue for the full-year increased 54% because of very strong client SSD and eMMC controller sales and SSD solution sales during the year.

Our Q4 client SSD controller sales grew over 5% sequentially with growth coming from our NAND flash partners.

Sales to our module makers, on the other hand, were flat sequentially as they continued to be affected by flash tightness.

For the first quarter, we expect our SSD controller sales to decline seasonally, comparable with what you would expect of the PC component food chain.

Based on our pipeline of SSD controller projects, especially our new PCIe NVME SSD controllers, we expect sales to start growing again in Q2, which should lead to full-year 2017 growth of 20% to 25% from over $170 million in 2016.

Our Q4 eMMC controller sales declined over 5% sequentially, in line with past seasonal patterns.

In Q1, eMMC controller sales will likely be flat, again similar to past seasonal patterns, and for full year 2017 we expect sales to grow about 5% from over $170 million in 2016.

We are expecting 2017 eMMC controller sales to grow in line with smartphone growth and, as Wallace had talked about, our deep relationship with Hynix continues to grow stronger.

Our Q4 SSD solution sales declined more than 40% sequentially as our big 2016 hyperscale SSD project ended and we do not expect our new 2017 products to ship until Q2.

Our SSD solution sales will likely continue declining further in Q1 due to limited availability of high-grade flash components at favorable prices.

In 2017, while we expect SSD solutions to rebound beginning in Q2 with our new hyperscale projects, for the full year sales should decline 15% to 20% from over $80 million in 2016 due to lack of flash availability.

Currently, we and our large hyperscale customers are both unable to obtain sufficient supply commitments of NAND components at reasonable terms to grow our sales in 2017 and we do not believe this situation will change in the first half of 2017.

As a recap of our revenue trends, we expect our upcoming Q1 revenue to decline 11% to 16% sequentially, primarily because of seasonally soft client SSD controller sales, flat eMMC controller sales, and declining SSD solution sales.

This is in line with certain NAND flash industry leaders, such as Samsung, who we believe are planning on reducing their NAND bid shipments in Q1.

For full year 2017, we expect revenue to be flat to up 10%, primarily because of strong client SSD controller sales growth plus modest eMMC controller sales growth, offset by the short-term decline in SSD solution sales.

Let me reiterate that our full-year revenue guidance is based on current tight NAND flash market conditions as we see it today.

We are expecting significant improvements in NAND flash supply in the second half of 2017, and if this were to happen sooner or we were to receive stronger confirmation of this improvement, we may look to raise our guidance.

In Q4, we had two 10%-plus customers, Hynix and another NAND flash vendor, down from three 10%-plus customers in Q3.

In Q4, our two 10%-plus customers accounted for 40% of total sales.

Our Q4 gross margin increased to 50.2% from 48.9% in the previous quarter, due to sales of a more favorable product mix of less SSD solutions and more SSD controllers.

In Q1, we expect our gross margin to be in the 48% to 50% range, which reflects less contribution from our client SSD controller sales because of seasonality factors.

For full year 2017, we expect our gross margin to be in the 49% to 51% range, in line with our long-term 50% gross margin target.

Our Q4 operating margin increased to 30.4% from 28.7% in the previous quarter, due to lower operating expenses.

We throttled back bonus and other compensation accruals so that they converged with actual payments.

In Q1 operating margins should decrease to 23% to 25% due to much higher R&D tape-out expenses, including costs of several controllers at 28 nanometers.

For full year 2017 we expect operating margin to be in the 27.5% to 29.5% range, similar to our 2016 operating margin of 28.2%.

2017 operating expenses should increase at a rate similar to our revenue growth.

Note that our operating expenses this year will be front-end loaded, as we will be spending much more on expensive R&D tape-outs in the first half of the year than in the second half with most related to SSD and eMMC controllers.

We ended the fourth quarter with 1,122 employees, 28 more than at the end of the previous quarter.

For full-year 2016 we added 149 employees; we expect to hire fewer employees in 2017.

Most of our new hires continue to be R&D engineers.

Our effective tax rate in Q4 was 23.8%, higher than the 16.8% in Q3 because of a one-time accrual adjustment at our Taiwan entity for conversion from NT dollars to US dollar tax accounting.

For full-year 2017, our model tax rate is 20%, higher than the 18% effective tax rate in 2016 and previous 18% model tax rate due to expiration of certain tax benefits in Taiwan.

Our Q4 EPS was $0.95, lower than the Q3 $1.07.

For full-year 2016, our EPS was $3.64, 73% higher than 2015's $2.11.

Total stock-based compensation in Q4 was $8.4 million, higher than the $6.5 million in the previous quarter due to the seasonal timing of RSU awards, consistent with past years.

For Q1, stock-based compensation should decline to within $3.2 million and $3.7 million, again due to the seasonal planning of RSU awards.

For full-year 2017, stock-based compensation should be within $14.5 million and $16.5 million, consistent with 2016's full-year stock-based compensation expense of $17.4 million.

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

We had $277.8 million of cash, cash equivalent, and short-term investments, $8.6 million more than in the previous quarter and $92.6 million more than a year ago.

In November 2016 we paid $7.1 million of dividends to shareholders, the first quarterly installment of our annual dividend.

In Q3 last year, we borrowed $35 million from two banks as a replacement for intercompany lending, because the cost of bank loans is comparable to the cost of using internal funds and with less hassle.

For example, without needing transfer pricing studies.

We expect that with our strong operating cash flow we can repay that loan within 12 months.

In Q4, we had $25 million of loans outstanding after repayments of $10 million.

This concludes our prepared remarks.

We will now open the call for your questions.

(Operator Instructions) Mehdi Hosseini, SIG.

Thanks for taking my question.

Two follow-ups.

First, can you please help me understand the kind of NAND that you use for high-grade SSD?

Is that still planar MLC or does it require more dense 3D NAND?

Then a follow-up; what would be your revenue guide if there was no NAND shortage?

So regarding the first question, the NAND we use for our enterprise grade SSD solutions is a planar 2D MLC NAND with the enterprise grade.

And to your second question, Mehdi, it is a bit of a hypothetical, but what would be very clear would be that if we had much better availability of NAND for our Shannon business our revenue guidance for the full-year would have been much higher.

Much higher than 10%?

This is a hypothetical question, but since there are certain -- a lot of uncertainties, including how much more supply would be available from the NAND flash vendors versus what we currently see.

And so it would have depended on the extent of availability of NAND to feed into our products.

Okay.

Just going back to my first question, I just want make sure I understand.

If high-grade enterprise SSD requires planar MLC and will migrate into [KD] NAND, does that imply that some of your NAND suppliers are actually going to add planar NAND capacity to alleviate the shortage?

No, no NAND maker will increase planar NAND capacity.

Actually, planar NAND capacity will be reduced because some equipment are going to be used to produce 3D NAND.

Okay.

So maybe I misunderstood; you said for high-grade SSD you use 3D NAND MLC, correct?

With a planar NAND, MLC.

2D NAND.

2D NAND, okay.

So it seems to me that (multiple speakers) 2D NAND.

So that implies that this shortage could sustain beyond 2017, because everyone is converting planar to 3D and the 3D NAND that is coming online is more TLC.

And this is where I'm a little bit confused.

What gives you confidence that the shortage (technical difficulty)?

So it all depends the combination of NAND chosen from different projects, because when the 3D NAND output improved significantly that will release certain 2D NAND obligations to certain customers.

That's why we believe we're going to see better visibility, better outlook from 2D NAND available to our customers in our channel business.

In addition, the new 3D -- the new enterprise projects will adopt 3D NAND.

The new project.

Right.

But the prices are more like 2018 volume ramp?

No, it's 2017.

Okay, great.

Thank you.

Anthony Stoss, Craig-Hallum.

A couple questions for you.

Stepping back, looking at the growth of the markets, maybe you can share with us a range of a multiyear growth rates, what you expect from all the industry's it serves.

Again if there's no NAND shortages.

Then, secondly -- also a two-part -- by the end of 2017 what percentage of your client SSD business do you think will be PCIe versus SATA?

And also if you can help us with the range on how much higher the ASPs are for PCIe?

Thank you.

Let me start off with your questions about market growth.

On the SSD side, with client storage devices we are looking about 500 million units of client storage devices, of which roughly about a quarter are already using client SSDs and the remainder are still using client HCD.

So given how fast NAND flash prices are falling and given how much NAND flash capacities are being built and will continue to be built, we're fairly confident that in, say, three to five years time the significant majority of the 500 million units are going to be going into -- using client SSDs.

So that should give you a framework in terms of how much, how quickly the market is expected to continue to grow.

On the eMMC side, eMMC plus UFS combined, most of these products are going into smartphones and so the expected growth rate for these type of products should be well in line with smartphone growth.

So if smartphones are expected to continue to grow 5%, then that's what you should expect for this market.

When PCIe versus (multiple speakers).

I'm saying we ship our PCIe controller solution to Intel and two model makers into 2016 and for 2017 we have numerous design wins and project out for PCIe with multiple flash vendors and many model makers.

So we are also introducing new PCIe controller in 2017 to, firstly, expand our product portfolio.

This is including high-end performance and including lower cost and including for [DRAM] (inaudible) and for single-chip PGA solutions.

So the (inaudible) price range is very, very broad.

It could be from $10 and down to $4 or $5 range for (inaudible) business.

So I think it's a very wide range of PCIe portfolio.

We have a three- to four-generation new product coming in the second quarter of 2017.

Great.

Thanks, Wallace.

Suji De Silva, ROTH Capital.

Thanks.

So Shannon revenue, can you talk about what that was for the industrial SSD; what they will be from the $80 million?

How much you expect that to decline in 2017?

And will it still be a single customer driving most of the revenues in 2017?

First of all, it's not a single customer.

We have multiple customers for our Shannon business and on the Ferri side we have a very diversified set of customers.

All together, our Shannon plus Ferri products, we did about $80 million of sales last year, 2016, and we are going to be growing that -- it will be declining 15% to 20% this year, 2017.

Great, thanks for that.

Then in the smartphone upgrade cycle and the use of UFS, what do you think the cutover rate is to UFS for your mix versus eMMC exiting 2017 or toward the back half?

As we said, we expect the UFS will remain very high-end smartphone and until uMCP deposits become popular, supported by the new AP end of 2017 and UFS or uMCP won't be the material set mainstream product line.

We believe 2018 will become more meaningful for UFS and uMCP.

Thanks for the color, thank you.

Mike Burton, Brean Capital.

First, on the mobile side, I was wondering if you could give us your view of downstream inventories.

Have you seen anything abnormal this year versus normal trends prior to the lunar new year?

Then looking out further, any thoughts about the seasonality of the mobile piece this year?

Regarding the channel inventory, we did not see -- actually inventory in the channel is for our customer side; we did not see anything abnormal.

And we see the smartphone growth will remain 5% annually and we will see a steady growth quarter to quarter.

Okay, thanks.

And then as a follow-up to Tony's question on ASPs for PCIe, I understand there's a range there.

But can you help us understand, as that becomes a larger portion of revenues, does that have an impact on gross margin or OpEx relative to your SATA controllers?

Naturally, it won't because our SATA controller is also in transition to be more cost competitive new-generation controller to support 3-D NAND and we have both offer DRAM and (inaudible) the solution to our customers.

For PCIe, I think be a higher ASP and we have a variety of PCIe solutions (inaudible) we can maintain better gross margin compared with 2016.

Let me add further to what Wallace said.

Our ASPs for every category of products are generally very, very stable and what we have is a natural cycle of products, products cycling in and products cycling out.

We constantly have a newer generation of products coming into our product mix, like the PCIe NVME controllers, and our mainstream products, over time, will become mature and will cycle out, like our SATA SSDs.

And so the combination of products that are coming in at higher gross margin -- higher ASPs and higher gross margin, as well as older parts where ASPs are falling and gross margins are coming down.

Through a blending of these products coming in and exiting our product portfolio, we are able to maintain our pricing and then gross margins at fairly stable levels.

Great.

Thanks, guys.

Betsy Van Hees, Loop Capital Markets.

Thanks so much for taking my question.

I just wanted to go back for a little bit of clarification.

So when you guys gave your Q4 guidance, you went over some 2017 metrics and it sounds like the only thing that is changing is the Shannon business, because you had guided it at the time that you expected it to be up 15% to 20% year over year and now you expect it to be down 15% to 20% year over year.

But the client SSD controller and eMMC controller business, it sounds like you are continuing to keep those guidances for 2017.

Am I understand that correctly?

That is correct, Betsy.

Okay, great.

Then I was wondering if you could help us understand in terms of the Shannon business, you had expected NAND supply to be tight in the first half of 2017 when you gave your guidance in Q4 -- for the Q4 quarter.

So could you help us understand what has really changed so much in the Shannon business that it's going to be down so dramatically versus up 15% to 20% year over year?

What happened in the NAND supply dynamics in the last couple of months?

While NAND manufacturers are actually bringing new 60 layer 3D capacity online, it takes time to improve the yield, because the yield-related uncertainty of flash partner able to provide firm procurement forecasts involving our controllers and their new capacity.

So we have see previous design-in product from last year using planar mode 2D NAND MLC.

We have a new design project with 3D NAND, so because it cannot reach confirmed pricing favorable turn, our customers and channel were not able to procure sufficient NAND for the project.

Let me also add, markets are very dynamic.

For example, Samsung today in their earnings call they explained to everybody that expectations of big growth for the full year are going to be a lot less than what investors were expecting or what they had previously communicated.

I believe they were previously communicating 40% bit growth and now they have taken it down to a 30% bit growth for the year.

Similarly, for the first quarter, most people were expecting Samsung's bit shipments to grow in the first quarter and now they are communicating that their bit shipments are going to be declining in the first quarter.

So it should give you a sense of the dynamic nature and how fast it changes in our industry.

Thanks so much for the clarification.

That was very helpful, the understanding between the 2D NAND and the 3D NAND and where the shortfall is coming in the Shannon business.

I greatly appreciate that.

And then, Riyadh, just wanted to ask a question.

I think you mentioned that the OpEx is going to be higher in the first half of the year and then it was going to come down in the second half of the year.

So how should we be looking at OpEx for Q1 this quarter, your guidance?

Our Q1 OpEx and our second-quarter OpEx are going to be higher than in prior quarters because of R&D tape-outs.

Most of the tape-out this year will take place in these two quarters.

Most of our tape-outs, as I believe we previously explained, are going to be 28 nanometer and these are very expensive.

These are necessary investments that we need to make as most of them are for SSD and eMMC controllers.

After these tape-outs are done, however, our second-quarter operating expenses will be much lower.

This is going to result in our operating margins being lower in the first half than in the second half.

Also, we will have better operating leverage in the second half because of stronger revenue growth.

For the full year, we believe we are going to be able to deliver operating margin at levels similar to last year.

And just to add a bit more, in Q1 our operating margin, as we previously said, should decrease about 23% to 25%.

Again, this is all largely due to higher R&D tape-out expenses.

Okay, thanks for that.

Appreciate it.

Then my last question for you, Riyadh, is on the tax rate.

You said it was going up to 20% from 18% and you mentioned several factors, one of them being tax credits in Taiwan that are -- you are losing.

Do you expect those tax credits to come back anytime in 2017 or are those completely gone?

Betsy, some of our tax benefits expired.

These are tax benefits that we enjoy in Taiwan.

They expired at the end of 2016, so this is leading our effective tax rate to inch up to about 20%.

We are actively managing this right now and we believe that next year our tax rates could return to our long-term 18% model tax rate.

Okay, thanks so much.

I appreciate it.

Jaeson Schmidt, Lake Street Capital Markets.

Thanks for taking my questions.

Just curious if we should still think about the core, that card and USB business, declining in that 10% this year or if you have had any changes to that outlook.

The markets for memory cards in the [factory] are mature.

We focus on high end, more differentiated [UHS-1, UHS-2], and USB 3.0 products to better differentiate from our competitor.

The marketing for expandable storage long-term expected to decline around 10% annually.

But with the tightness this year our expandable controller business will be even more impacted and will likely decline more than 10%.

Okay, that's helpful.

Then just wondering if you've seen anything out of the ordinary from a pricing standpoint within the client SSD space.

With the tight NAND supply, frankly speaking, we see the NAND price and (inaudible) price is moving up because of supply shortage.

Unless the NAND supply situation improves, I think the current client SSD pricing will be stable.

Our total pricing also will be stable.

All right, thanks a lot.

Mike Crawford, B. Riley & Co.

Thank you.

Wallace, you talked about gearing up to enter the enterprise-grade controller market.

Is that beyond just this first 12.8 terabyte drive for Alibaba?

That is correct.

I think our enterprise strategy today is in three parts.

First our channel enterprise at the target very high end, mainly in critical application at a large internet and a cloud player in China.

Second, our customer are using our computing grade SSD controller in the data center application already today.

And, third, we are already developing enterprise-grade SSD controller and then will be launching these products in 2018.

In markets beyond just China?

We will focus on the channel initially in China and then we expand to outside China to US and other countries.

Okay.

And what do you mean by retooled?

Retooling is in -- most of our controllers today the firmware is tailored for client SSD or PC-OEM.

For enterprise-grade data center the priority and focus is different.

For example, they now focus on high-performance sustain rate, so your firmware and algorithm need to have a different copy collection, different feature to maintain sustained performance in data center applications.

I see.

And when you are addressing the enterprise market, does that mean that there's much longer kind of design-in cycles or is that not materially different from what you've been experiencing so far in the client market?

Regarding the controller, we retooled firmware for the market.

Naturally, the cycle is very quick and I think (inaudible) had already started shipping client-grade SSD product into low-end enterprise clouds data center applications.

For real enterprise-grade controller, I think design cycle is longer to be qualifying large-scale enterprise customers.

Okay, thank you.

Then the final question in that regard is it's my understanding enterprise controller ASPs can get up even into the $100 range, well above the $5 to $10 we're talking about for PCIe client controllers.

In 2018 when this business becomes more meaningful for you, what range of ASPs might you see for your enterprise controller efforts?

We cannot comment on ASP right now.

I'm not sure whether it will be $100, but I believe ASP will be much higher than client SSD controller.

All right, thanks.

Ravindra Gill, Needham & Company.

Thank you for taking my question.

I just wanted to see if you could elaborate on, Wallace, your comment on the allocation of capacity for the shim SSDs.

You had said the difference between wafer sales versus solution sales or system sales.

Wondering if you could elaborate further on that point.

It's very clear; when NAND supplies are tight, every NAND maker has his own priority to allocate the NAND for distribution.

So some portion will delivered to committed their major OEM customer and some will, based on this model for example, maybe focus on smartphone or enterprise SSD, not the client side.

Some NAND makers maybe prefer to sell NAND components which have a better margin than sales solution in the client SSD, our example product.

So every NAND maker has different priority allocating NAND, and especially for a module maker, normally they have a lower priority to receive the NAND when NAND is in shortage condition.

Based on that logic then (multiple speakers) go ahead, Riyadh.

Based on the logic then and based on your guidance for client SSD still growing 20% to 25%, clearly they are -- the flash vendors are prioritizing SSD or NAND supply to client SSDs, because that's where they are seeing strong demand from that segment.

For some NAND makers that is correct.

Furthermore, for our Shannon business, as we mentioned for many of the flash makers if they can capture better economics by using the flash within their products, they are going to be allocating less as sales of wafers to customers like us for us to build our own SSDs.

For us to build our own Shannon SSDs, we've got to procure NAND wafers from the flash maker and then take the wafers to turn them into SSDs that we then sell to our hyperscale customers.

So there hasn't been any market share shift in the Shannon SSD business, is that correct?

Just a reallocation of NAND, a reprioritization of the NAND flash, given the tightness in the market?

That is kind of the way to look at it because our products are highly-customized products, so we are not in the market of competing against the NAND flash makers who are making -- who are providing standardized SSDs.

The SSDs that we are providing to our hyperscale customer, these are custom designs; these are based on the specifications that our customers have provided to us.

And so we are not competing against the flash makers for their type of products.

Ours are customized; theirs are standard.

But that said, if a hyperscale customer cannot secure sufficient NAND wafers for us to make SSDs, then our customers are going to be turning to standard products for their requirements.

So that is what they are doing then, right?

All these hyperscale customers in China are turning to standard product versus customized products, given the tightness in the supply?

That's correct.

That is the situation we are facing in the short term.

This situation will likely change when there is greater availability of flash, but currently that's how we are -- that's the situation our customers are facing.

They would like to build more customized SSDs where these SSDs are designed specifically for -- tailored to their data center requirements.

But if they can't secure sufficient flash to build them, then they will have to turn to other means to build their infrastructure.

Just last question on that and I will hop back in the queue.

Then what would be the actual business impact for those customers, those hyperscale customers in China if they move to a standard product?

I'm just curious; if the business impact is minor, will they just stick with the standard products versus going back to customized products?

I guess what is the differentiation between more of a customized versus standard?

I think it depends on customer type.

For Tier 1, very large-scale customers, they temporarily might turned to standard product because either they cannot afford to wait and delay product launch.

But for medium and small customer, they can wait and they are willing to pay more regarding to procure the NAND.

So I think, for Shannon, we will grow medium and the small customers in 2017.

To add further, when there's greater flash availability -- there's still a lot of demand for our type of product, these highly customized, from our large hyperscale customer.

So there is still that intention of they want to use the solutions that we are designing for them.

These are products that fit very well within their infrastructure, that perform very well with their requirements and they like to do -- use more of these.

And so availability comes around, we are pretty comfortable that they are going to want to use more of what we are selling.

But as things stand right now, that's just not possible.

Thank you, Wallace and Riyadh.

Appreciate the color.

Tom Sepenzis, Northland.

Thank you.

Just a follow-up on the Shannon business, my question is is there any danger that you have missed any programs?

How quickly can you turn around and give product to customers if NAND flash availability loosens?

We did not lose any customers or lose any programs.

I think just customers -- our customer in China, major customer, because of tight NAND supply in certain NAND types they tend to procure finished-good solutions from NAND makers temporally.

When NAND supply becomes better, I think our (inaudible) were going continue ramp up and our new project will continue ramping up with 3D NAND.

So we mentioned our current product in production is 2D MLC; the new product is a 3D TLC.

To add further, we -- as Wallace had mentioned, we did not lose any customers and we certainly did not lose any programs at these customers.

It just turned out that the volume expected for each one of these programs are now a lot smaller than what we had originally because our flash partner -- our hyperscale customer is just not able to secure sufficient NAND components to build at the original plan levels.

Great, thank you.

Then you mentioned that the removable storage would be down 10% quarter over quarter; that you expect that year over year in 2017, down 10%.

That's correct, yes.

At least 10% down.

Thank you very much.

There are no further questions at this time.

I would like to hand the conference back to today's presenters.

Please continue.

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

We will be at the following conferences this quarter: [SIG annual] technology summit in New York; ROTH investor conference in Laguna Beach; Merrill Lynch investor conference in Taipei; Morgan Stanley investor conference in Hong Kong.

Details of these events will be available on our website.

Thank you and goodbye for now.

Ladies and gentlemen, that has concluded our conference for today.

Thank you for participating.

You may all disconnect.