慧榮科技 (SIMO) 2018 Q1 法說會逐字稿

Ladies and gentlemen, thank you for standing by, and welcome to the Silicon Motion Technology Corporation First Quarter 2018 Earnings Conference Call. (Operator Instructions)

Before we start the conference, I would like to read out the safe harbor statement for you. This conference call contains forward-looking statements within the meaning of Section 27A of 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 conditions 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, uncertainties, and actual market trends and our results may defer 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 the 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.

Now I would like to hand the conference over to your host for the day, Mr. Jason. Over to you, sir.

Thank you, and good morning, everyone, and welcome to Silicon Motion's First Quarter 2018 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 business developments. Riyadh will then discuss our first 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 our Safe Harbor policy, which 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 market -- 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 the 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 us today. Let me first update you on our business, and then Riyadh will review our financials and provide our outlook later on the call.

We are pleased to report first quarter revenue of $130.3 million, which is up 2% year-over-year and down 4% sequentially, and is the highest first quarter revenue in our corporate history. We saw solid sequential sales growth from both our client SSD controllers and eMMC controllers. This trend was, however, offset by similarly soft SSD solutions sales.

For the first quarter, we delivered earning per ADS of $0.71, which is also up year-over-year and also the first quarter record high.

Our business started this year with much more favorable market condition than last year. Unlike last year, our business no longer is negatively affected by lack of the NAND supply and rising NAND prices. Currently, NAND supply is rapidly a variable and NAND prices continue to soften. In Q1, increasing output of 64- and 72-layer 3D NAND had led to modest NAND price reductions. We believe NAND supply will continue to grow as the year progresses and price will decline further, both favorable trends to us.

Lower NAND prices will further strengthen demand for client SSD and demand for our controllers. Currently, interest by NAND flash vendor in building client SSD is much higher than the same time last year. Our SSD controller projects with our 3 NAND flash partners are over 80% higher today than 12 months ago.

PC OEMs are also communicating to us their plan to increase adoption of SSD this year versus a little or no increase in SSD adoption last year.

NAND industry supply will increase much more strongly this year than last year, especially from 64- and 72-layer 3D NAND capacity coming online. 64-layer 3D manufacturing yield are already matured and the flash vendor are keen to further ramp output.

Intel and Micron have already been sampling SSD using their QLC 64-layer 3D NAND, and we believe QLC-based device sales will gradually increase in the second half of this year. Additionally, most flash makers have already begun to sample with customers their 96-layer NAND and are expecting commercialized production in the second half of this year.

All of these technologies, such as higher-layer count from 64- to 96-layer 3D and higher-density QLC not only increased industry capacity and the output, but also leading to significant and rapid reduction in NAND costs per bid. And seeing the NAND industry is highly competitive, reduction in price per bit quickly follow cost reductions.

We believe our NAND flash partners are currently taking advantage of their increasing supply of low-cost NAND to build more client SSD and price their devices attractively to OEM and other customers.

Based on today's NAND flash prices, we expect our SSD controller sale to grow by at least 20% for the full year, and we expect full year SSD controller revenue to exceed what we achieved to 2016.

As previously mentioned, we currently have growing portfolio of SSD controller projects as our 3 NAND flash partners over 80% more than at the same time last year. The majority of these SSD controller projects are for PCIe, NVMe and interface that is designed specifically for faster NAND flash and SSD. The majority of SSD controllers that we are currently shipping on the other hand are using the legacy SATA 3 interface, originally designed for slower hard disk drive, but still extensively used in client devices, data centers and enterprise applications.

Our flash partner are working with their OEM customer to transition SSD technology from SATA 3 to higher-performance NVMe and expect that majority of client SSD sales will be based on NVMe next year. We believe we will benefit from this transition to NVMe and it could, as a result, further increase our SSD controller market share.

We are also excited to be working with 2 flash makers on their more cost-competitive 4 b/cell QLC-based PCIE, NVMe client SSD, which are expected to start shipping in the second half of this year and gradually ramp. Our controller project will help bringing to market the world's first client SSD to be using QLC NAND flash.

We believe our success has been a function of our best-in-class technology and cost-effective solution, but also our unique value proposition to customers.

Our turnkey solution and customizable platform cover a broad range of market from SATA to PCIe NVMe from single-chip VGA SSD to low-cost DRAM-less SSD to premium high-performance SSD, and offer customers option to select from wide range of flash from MLC to TLC and QLC; from legacy player NAND to current 64- and 72-layer 3D NAND and upcoming 96-layer 3D NAND.

We provide our flash partner with hardware plus firmware turnkey solution for their SSD targeting retail and other fast-turnaround markets. These turnkey solution are also widely used by our module maker customers. We also provide our flash partner with highly customizable hardware plus firmware platform for developing SSD that they can tailor their own firmware for their OEM customer and application-specific need. Additionally, we then provide our customer with a match on the ground local engineering support in the U.S., China, Taiwan and elsewhere.

Now let me turn to our open-channel NVMe SSD controller for hyper-scale data centers. Our first hyper-scale customer began sampling and [testing] our controller late last year, and the feedback continue to be positive. We anticipate this controller entering production midyear. Separately, early in April, our second hyper-scale customer began sampling our open-channel SSD controllers.

Open-channel SSD technology continue to evolve and progress rapidly. Microsoft has been leading the standardization of open-channel SSD technology with Project Denali, as of which we recently began a joint development partner separately. Alibaba and Chinese hyperscalers are developing a different open-channel SSD standardization and we are also involved in this as well.

We believe, over the next few years, open-channel SSD technology will be broadly adopted by most of leading hyperscalers for their data centers, and we are well positioned to grow as a leading controller supplier to this market.

Let me now turn to our eMMC controllers. In the first quarter, revenue rebounded sequentially as SK Hynix began replenishing their depleting eMMC controller inventory. Longer term, the market for eMMC-embedded memory, which is primarily used in Android smartphone is an ex-growth market.

Our strategy is to actually manage this product line to deliver stable revenue stream and reduce the significance to our overall business as we rapidly grow our SSD controller and SSD solution sales.

While our relationship with SK Hynix remain strong today and we continue to very effectively support SK Hynix with our very competitive low-cost, high-performance eMMC controller, we are also working very closely with Micron as their UFS controller partner. We believe our UFS controller sale should scale next year as UFS adoption increases, and this should offset the expected gradual decline of older eMMC technology. This combination of continuous sale of old eMMC controller at SK Hynix, plus our newer UFS controller sale growth at Micron should enable us to deliver a stable revenue and cash flow stream.

Now turning to our Shannon data center SSD and Ferri industrial SD solutions. Revenue declined seasonally as expected this quarter, and will recover sequentially in the second quarter as the new project scales. Unlike first half 2017 when NAND supply was short, sales of our SSD solution in this year will not be facing any NAND supply constraints. We expect our Shannon SSD sale to grow in the second quarter as new Alibaba projects scale with their growth in data center space. Separately, we are also expecting our Ferri SSD sale to rebound in the same quarter as our product is diversified further into automotive infotainment systems, data networking (inaudible) storage and other applications.

Let me conclude by saying that we are off to a good start in 2018. The first quarter was largely as expected, with more NAND supply coming to market and NAND flash prices declining modestly. We expect these trends to continue throughout 2018, which will benefit our business, especially our client SSD controller sales.

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 then provide our 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 yesterday.

Our Q1 revenue decreased 4% sequentially, in line with guidance. For Q2 2018, we expect revenue to increase 3% to 8% sequentially with growth led by our SSD solution sales. For full year 2018, we expect revenue to increase 5% to 10% with bias towards the upper half of this range, led by accelerating demand for our SSD controller, strong growth from our SSD solutions and stable eMMC controller sales, partially offset by declining memory card and flash drive controller sales.

I will now walk you through the performance of our key products before taking you through key elements of our P&L and the rest of our financials.

In Q1, sales of our SSD controllers grew 10% sequentially, which is follow-on growth from our SSD controller rebound that began in Q4 2017. In Q2, we expect our SSD controller sales to grow further as new programs begin to ramp. For full year 2018, we believe our SSD controller sales will likely grow at least 20%, upward bias from our previous expectation.

While we saw some modest NAND price reductions in Q1 and believe price reductions could accelerate in Q2, we remain unsure to what extent we could benefit from this for the rest of the year. We are noting that bias is on the upside, and we will likely enjoy sequential SSD controller sales growth for the rest of the year. We also reiterate that if NAND SSD prices were to fall meaningfully, our SSD controller sales growth will likely be stronger than currently anticipated.

In Q1, sales of our eMMC controllers grew 15% sequentially, as SK Hynix continued to rebuild their depleted eMMC inventory. In Q2, we expect our eMMC controller sales to be stable. For full year 2018, we believe our eMMC controller sales should be flattish, in line with 2018 smartphone industry forecasts.

In Q1, sales of our SSD solutions declined over 25% sequentially because of seasonally weak Shannon and Ferri SSD solution sales. In Q2, we expect our SSD solution sales to rebound sharply sequentially as new programs ramp.

For full year 2018, based on our current portfolio of Shannon and Ferri customers and projects, we believe our SSD solutions should be able to grow at least 20% with strong growth from both our Shannon and Ferri products.

Our Q1 gross margin increased sequentially from 46.6% to 48%, largely due to growing mix of sales from higher gross margin controllers and declining lower gross margin SSD solutions. We expect our Q2 gross margin to be in the 46.5% to 48.5% range, as we expect lower gross margin SSD solutions to grow faster than higher gross margin controllers.

For full year 2018, we continue to expect gross margin to be in the 47% to 49% range.

Our Q1 operating margin decreased to 22.1% from 24.1% in the previous quarter because of lower revenue and higher operating expenses, slightly offset by better gross margins. We expect our Q2 operating margin to increase to 22.5% to 24.5%. Operating expenses in Q2 should be similar to Q1.

For the full year, we expect operating margin to be in the 23% to 25% range. Total headcount at end of Q1 increased to 1,262, which was 12 more than at end of Q4.

Our effective tax rate in Q1 was 14.6%, lower than 16.6% in the previous quarter. We expect our effective tax rate for full year 2018 should be roughly 15%.

Our Q1 EPS was $0.71, lower than the previous quarter's $0.79, but higher than $0.70 a year ago.

In Q1, stock-based compensation in our operating expense, which we exclude from our non-GAAP results was $3.1 million, lower than $8.1 million in the previous quarter, due to seasonal timing of RSU awards and consistent with timing of past years.

For Q2, we expect stock-based compensation of $0.6 million to $0.7 million and $14.8 million to $16.8 million for the full year 2018. We had $346 million of cash, cash equivalent and short-term investments at the end of the first quarter, $20 million less than in the previous quarter, but $42 million more than a year ago.

In Q2, we paid $10.8 million of dividends to shareholders, the second $0.30 per ADS quarterly installment of our annual $1.20 per ADS dividend that was announced in April of last year.

This concludes our prepared remarks. We will now open the call to your questions.

(Operator Instructions) We have the first question from the line of Gokul Hariharan from JPMorgan.

First of all, I had a question on enterprise controller side. I think Wallace had mentioned in his prepared remarks that you are starting to see some more traction, especially among Microsoft as well as Alibaba-led consortium. And you also mentioned it will -- your first product will get into production. Could you talk a little bit about a couple of things? Any thoughts on when you expect this to contribute to revenues? What is the kind of competition you are facing in the open-channel space, given that many of the established controller companies don't seem to be very supportive of open-channel? And lastly, is this going to be restricted to 1 or 2 hyperscalers or is it a consortium, which could be adopted across multiple hyperscalers just like what happened in the server space with open compute? That's my first question.

So let me answer your first question. We expect our first open-channel controller production in middle of this year, but volume would be very limited. We believe our open-channel controller enter meaningful production is really 2019 next year. We just become member of the Project Denali open compute project host by Microsoft and we are joining as a discussion member. We believe from the industry, Microsoft would like to define a standard interface for their open-channel SSD controller end product because enterprise SSD is not equivalent to data center SSD. We believe majority carmaker would like to see what tailor-made SSD solution to meet their specific need. So we believe this trend will become broader and broader and more attractive to favor data center makers. Although initially, NAND maker hesitate to participate because they will reduce value for NAND maker SSD solution. But we believe eventually, NAND maker -- currently, at least 2 to 3 decided to participate to support this event. Because China will have their own standardization of open-channel, it will be very favorable to Silicon Motion to participate in both organization and to adopt our intellectual property from controller to firmware knowledge for NAND to our Shannon, the driver layer FTL that gave a tremendous opportunity to serve the growing market. We believe the market will become meaningful maybe from late 2019 and 2020, and we should see really significant sale revenue grow from 2020.

Okay. That's very helpful. Could you also share a little bit on the client SSD controller side? What are the recent updates that you're seeing from your customers, given you mentioned that you have seen 80% increase in terms of the number of projects that are -- that you're working with your NAND flash partners?

We have a pretty broad [evolving] things in 2017 last year. Majority are design projects are focused on 3D NAND and from 64-layer TLC to 96-layer TLC as well as QLC. We will be the first controller maker to support 64-layer QLC SSD, and we do see very broad range of the product line. And through the pipeline, we have quite strong confidence in our real sale revenue and

(technical difficulty)

and really exceed 80% compared to last year. And we're going to grow at least 20% our sale revenue compared with last year.

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

Wallace, you mentioned that NAND pricing is declining moderately and you expect it to kind of decline further as you flash makers take advantage of the mature yields. Wondering if you could -- now that you're about a quarter into the year, give us a sense of what the attach rates for client SSDs, kind of the rough ballpark are going to be this year in the desktop notebook market versus kind of where they were last year?

We see really the first half of the notebook really is -- has softened and we really do not see the strong growth. However, adoption rate of SSD momentum is very strong compared with last year because the price declined sharply especially for OEM. The NAND price decline for the channel is slowly gradually, but NAND price decline to OEM, I've seen is sharper. And we see it's about 1/3 to the PC market, about 1/3 to [PaaS] compared to this year.

Okay. Got it. Okay. So the NAND pricing to the OEMs is much sharper versus the NAND pricing to the module makers? I just want to understand that.

That's correct.

Okay. All right. That's interesting. Okay. And then your kind of shift to PCIe, NVMe, the real trends in the industry, which demonstrates your innovation and technology. Can you talk about that transition from SATA 3 to PCIe and NVMe? Is this broad based across the industry? What's driving the move to faster interfaces? And how are you then positioned competitively in PCIe and NVMe as these are more advanced interfaces, controllers?

I think our second-generation PCIe and NVMe controller that launched last year are doing very well. In fact, most of our new client SSD win with these new controller is our unique turnkey solution and customizable platform, which help our customer quickly bring to market high performance. From the market per se because Intel and Samsung driving PCIe into the market very aggressively, so we do see PC OEM in commercial product line, 70% are already adopted the PCIe, but the consumer line is still below 50%. But in the channel, I think channel is still slow. Majority of the channel -- probably 70% are still in the start-up, but we do see the transition due to 3D NAND, higher-density 3D NAND, faster 3D NAND coming and costs going down. That's why it would trigger the demand for PCIe SSD to replace the SATA as well as replace majority of the HDD. And we do believe, in 2020, the PCIe BGA SSD will become even more attractive to replace HDD as well as replace HDD in desktop. So that's why we see the PCIe momentum will become stronger and to replace SATA next year in 2020.

And last question on that, Riyadh. The ASPs, I think, in the past you said, are twice as high versus the SATA controllers SSDs. Does that still remain the case?

Well, our -- Raji, our ASPs in general for our client SSD products on average is roughly about $5. Obviously, for newer products, we are going to be rolling out our newer products at higher price points. But as they become mainstream, our ASPs will converge towards the $5 average.

But next year or -- later this year or next year, you are going to see an ASP bump as they transition to the this PCIe SSDs.

If there are any ASP bumps, they are going to be temporary. Our business model is not about pricing our products increasingly higher, but rather to provide a same sort of affordability to our customers, so that they can be competitive in their markets. And the more competitive they are, the better we benefit as their controller partner.

The next question comes from the line of Mehdi Hosseini from SIG.

I have one follow-up question for Wallace and one for Riyadh, and thank you for all the details. It seems to me that the long-term -- longer-term growth opportunities are very attractive. So for Wallace, I want to better understand as how the cloud guys and their procurement becomes very different and distinguishable compared to the enterprise. When I look at NVMe market is dominated mostly by Samsung and Intel, those 2 effectively account for 75% of the NVMe market, either -- it doesn't matter client or enterprise. But I think that market share is up for grabs when we look into the cloud and -- especially with the open-channel, it creates opportunity. And I want to get a better sense from you or maybe some more detail. I think you kind of alluded to this early on, but help us understand the market competitiveness or competitive landscape with open-channel? And how that becomes very different than enterprise? And I have a follow-up.

Okay. Let me try to explain a little bit about the business model of open-channel. The way Microsoft want to define industrial standard for open-channel is they would like to change this model. They are able to procure the NAND from different vendor. They can also procure the controller from different vendor and to consign to manufacture partner to make SSD. So the SSD sale will be more NAND independent, because the [effort to] FTL move to the driver layer and the host maker able to control. So this is the fundamental business model change. Now looking for Silicon Motion, how we play into open-channel. We are a controller maker. We understand NAND very well. We understand the FTL flash management, the [GABI] collection, the ECC, everything very well. And we also know the driver layer very well through our Shannon system. Their solution have already developed to driver layer FTL serve the China and major customer. That's why we are able to integrate all different technologies together seamlessly, maintain a good performance and no latency and to tailor for customer need. That's why we are able to launch 2 different products very quickly for China customer this year. And looking at the landscape, NAND maker, they hesitate to jump into the total solution due to each of carmaker might be needing your customization, need a very, very tedious support for many different special functions because different cars have a different purpose, have different application. They are more [object] oriented. There are different than enterprise SSD. So we believe they give us not only opportunity to serve the carmaker directly, also gives more opportunity to serve NAND maker, which they did not have such a solution. And so I cannot mention who they are, but it really is very exciting for Silicon Motion to open the door to [end] open-channel for car industry for call makers.

Great. Crystal clear. Now let me turn to Riyadh. You have been making investment to help you with these longer-term opportunities. When I look at the revenue and operating profit to employee, it peaked in 2016 and it has been declining as you reinvest in the company. When should we see a turnaround in this? Is this more of 2020 when cloud opportunities happen and the scale comes in? Is that when we should see a rebound in like a dollar operating profit per employee? And is that when the leverage is going to come back into the model?

That's a very good question, Mehdi. We actively track our revenue per headcount. And our revenue and headcount had been improving quite nicely until last year. Last year, unfortunately, our revenue did not grow, but at the same time we were -- we are committed to continuing to invest ahead of curve in order to make sure that we are properly positioned for our longer-term growth opportunity. For example, the open-channel data center SSDs, right? So we continue to have to add engineering resources in order to make sure that we are -- we remain ahead. But at the same time, we have to -- we suffered through last year where revenue declined a little bit, so resulting in our overall dollar per headcount coming down less than optimally. Now as we get ourselves back on the revenue growth curve and a lot of the headcount investments have already been put in place, we're not expecting the heavy sort of headcount increases that we've had in the past. And at the same time, when our revenue starts to scale again, we should get ourselves back on the trend where our revenue per headcount becomes more optimal.

But is the leverage in the model -- and I know we're talking about the longer term and there are some uncertainties, but -- would help for investors to understand -- or to have a grasp of when would the leverage come in a given all the investments that you have already made. Is that more like second half of '19 when open-channel scales? Or should we just wait and see how your new products are going to evolve and turn into high volume manufacturing?

Well, the first trigger for us would be in the near term when the NAND prices fall more dramatically. Because the faster NAND flash prices fall, the more our client SSD controllers are going to be able to scale, right? And so when that scales, we're going to have must faster revenue growth and the metric -- dollar per headcount will start looking much better. And longer term, obviously, when our data center SSD controller starts contributing, we'll also have the benefit of that. So the near term -- the uncertainty that we're facing is when will NAND prices come down and how fast will prices come down, so that we can start having stronger client SSD growth. And that will be the largest determinant of our dollar per headcount. But longer term, one of the targets that we have in managing our business is trying to get ourselves to roughly 30% operating margin profile.

The next question comes from the line of Sujeeva Desilva from Roth Capital.

So my question is on the SSD solutions business here. Can you talk about whether you are able to improve the gross margins on this business on a secular basis? Or whether they are kind of fixed around NAND prices, where they have to pass the NAND pricing moves through to the customer?

We believe last year, we did suffer the NAND shortage and our procurement for NAND is not really that great. So I think the gross margin is going to suffer continually through the second quarter. However, I believe, around the third quarter, we should see better margin improvement for our SSD solution and we should be able to benefit with the new NAND price decline and that will also benefit our SSD solution for our third quarter.

Let me also add more to what Wallace said, just so that you have the right expectation. You should not expect our SSD solution's gross margin to be comparable to our controller gross margin. By definition, most of the components in our SSD solution are NAND that needs to be procured. And in a typical SSD, NAND could be 80%, 90% of the bill of material. So by definition, these products are going to be below corporate average in terms of gross profitability.

Okay. And then the other question on the SSD solutions is, where the revenue size for '18 will come in will be kind of a record year for you guys and what the long-term growth rate for that business can be? An estimate there?

I think we -- I think the major growth for our SSD solution for Shannon is really the open-channel SSD for next year. If we execute very well, we should be able to see very strong growth in 2019. It's all the execution and customers, their, really, commitment. So far, we see the momentum is very strong, and we're preparing to launching the solution aggressively next year. This year will be a small volume production.

In addition, we also have our Ferri SSDs, which continue to do very well. And in this part of our business, we have a very diversified set of customers, who's going into also very diversified set of end applications. And a lot of these projects are long-tailed, and so it's a matter of accumulating more and more customers, going into more and more applications. And we -- so we expect our SSD -- our Ferri SSD solutions to also continue to grow longer term and potentially even at a comparable growth rate to our Shannon products.

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

Just to be clear on semantics. When you say that you're working on 80% more client SSD controller projects than at this time last year, is that a function of unit volume or potential revenue impact or how are your measuring that 80%?

Project numbers. Number of projects.

Given that some projects have the potential to be perhaps mainstream products and others may be niche products. Is -- can you take a stab at potential unit volume and revenue with that?

We cannot talk about detail. But from the number of projects are really involved is 80% more than compared with last year. And due to the new NAND, the trend, we benefit leading in our technology to support 96-layer TLC as well as QLC. That's why we get tremendous work from our NAND partner in helping them accelerate launch into the market for late 2018 and 2019.

Let me also add further. These projects relate to the different types of NAND that are coming out, the current 64-layer flash, the upcoming 96-layer flash. We have QLC, TLC. We also have various other different types of NVMe projects that Wallace had previously mentioned. So it's accumulation of all these projects that are -- at our NAND partners leading to 80% more projects this year than at the same time last year.

Okay. And then Riyadh, in recent quarters, past 4 or 5, there has been some pretty big swings in working capital from quarter-to-quarter, I think more so than what we've seen in the prior years leading up into this. Is that your assessment as well? And how should we think about working capital expansion and contraction going forward?

Mike, there has always been certain amount of lumpiness relating to our working capital, especially when you look at it on a turnover perspective. If you -- and then it's -- and a lot of it deals with when we are building inventory for upcoming growth or when we are drawing down inventory or where timing of accounts payable, we could have a lot of sales, for example, that are happening at the end of the quarter. And that would make the presentation of number look less than optimal. But generally, our payable cycles and receivable cycles are fairly short. And certainly, for this quarter, there is nothing out of the ordinary other than timing matters.

Next question comes from the line of Mike Burton from Longbow Research.

With the open-channel controllers on track to launch midyear, I know you mentioned that -- obviously, a lower gross margin profile. But I'm just wondering how we should think about it versus SSD solutions on the operating margin side? And then also, on the revenue side, is there any cannibalization of your SSD solutions by this open market controller? Or any color on the percent of bill of materials that this controller would be?

Mike, let me start by just correcting perhaps the wrong understanding of our open-channel. Our open-channel controller, open -- these are controllers for our open-channel SSDs using data center. These are not below corporate average gross margin products. Our controller SSD projects are going to be above corporate average gross margin. Our controllers as a whole are above corporate average. And certainly, our SSDs, whether for client SSDs or controllers for client SSDs or controllers for our data center open-channel, they're all above corporate average.

I'm sorry. I meant that the SSD solutions are below on gross margin. I was wondering how that compares versus the open-channel controller on the operating margin side? And then also, the open-channel controllers as a percent of the bill of materials, and if there is any cannibalization of the SSD solutions?

For SSD solutions, Shannon products or Ferri products, these products as a whole are below corporate average in terms of gross margin. And all of these products -- the bill of materials primarily is NAND flash that we need to be procure.

I think -- to answer your question, Mike, open-channel-based SSD solution, the margin will be slightly better than standard NVMe SSD solution. So we are going to charge service fee and software fee and tailor certain software for the customer. So open-channel-based SSD solution, I think, gross margin will be slightly better than standard PCIe NVMe SSD solution.

Okay. Also you mentioned your partnership with Micron on mobile UFS controllers as an offset to some declines in the future at Hynix. I just want to clarify if that was supposed to begin this year or is that more of a comment about next year?

It will begin this year.

Okay. And then last, just other 10% customers in addition to Hynix?

At the end of 2017, we -- for full year 2017, we only have 2 10% customers. The first is Hynix and the second is Intel, just these 2 customers.

There are no further questions at this time. I would like to hand the conference back to the speakers. Please go ahead.

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 and U.S. in the second quarter. Detail of this event will be available on our website. Thank you, and goodbye for now.

Ladies and gentlemen, that does conclude our conference for today. Thank you all for your participation. You may all disconnect your lines now. Thank you.