Silicon Motion Technology Corporation (SIMO) Earnings Call Transcript & Summary

Okay. Good morning, San Francisco, and good evening, Asia. I am Simon Woo, Bank of America tech analyst, covering Silicon Motion and the semiconductor memory chip industry. This session is with Silicon Motion's CFO, Mr. Riyadh Lai. Riyadh has been working as a CFO for more than 10 years. And also, the IR Director, Chris, also here on the call. Again, a special thanks to Riyadh for our Global Tech Conference. This session around 45 minutes duration. So for the first 5 minutes, 6 minutes, maybe Riyadh will spend for the quick introduction and overview. But before we get started, let me read the disclaimer. Company disclosures relating to the individual companies or securities discussed on this call today can be found on the call invitation. All right. Riyadh, would you like to kick off then, sir? Thank you.

Sure. Sure. Sure. Simon, thank you very much for inviting us to your conference. It's a real pleasure to be here again. As many of you know, Silicon Motion is a fabless semiconductor company focused on controller ICs for solid-state storage devices. So these would be SSDs going into PCs and data centers; eMMC, UFS going into smartphones and a host of other applications. We're the market leader in controllers. We ship more controllers than any other company in the world. We are the merchant market leader in SSD controllers for client devices. We're also the merchant market leader for eMMC, UFS mobile-embedded storage going into smartphones and a host of related applications. We've done very well as many of you know. And this year, we're gearing up for a growth here despite the uncertainties relating to the economic -- the global economic issues. We're expecting growth from all 3 of our primary products: our SSD controllers, our UFS/eMMC controllers as well as our SSD solutions, which are largely our highly customized data center SSDs and industrial SSDs that we've been shipping to Alibaba and some of the data center guys in China in industrial applications in the diversified set of markets. So we continue to do well, and this year is going to be an interesting year, and we're gearing up for growth for the full year as well as growth in this quarter with contribution from all 3 of our key products. So with that, let me hand it over to -- in the interest of time, hand it over to Simon, and we can continue with Q&A.

Yes. Sure. Definitely. Great introduction. Well, let's just start with some general questions for the general people which are not familiar with the NAND flash controller that the Silicon Motion is doing a great job for the -- all the NAND solutions there. So maybe how about if you recap quickly why the NAND controller logic chip is so critical in the data center or even for the SSD storage area, even for the 5G phones. Why you think NAND controller is very, very critical component as a logic chip? And how this can offer the better function for the overall memory NAND or data storage function there?

Yes. So we're a critical companionship to NAND flash memory chips. So one way to think about it is the NAND chips are kind of like a building, like an empty building. And in order to turn the empty building into, say, a hospital or an office building or a school or a residential building, you need the controller. The controller is what turns this empty building into these type of application, these types of places where you can store things, where you can live in and where you can -- where you -- a host of different activities, right? So the controller converts the residential -- the building -- this empty building into an SSD, into an eMMC or UFS going to smartphones, the SSDs going to data center, going to PCs and a host of other applications. So we play this important role. And on top of that, even for these applications, say, back to analogy of a building, the NAND chip being the empty building, the NAND controller chip turns it into a hospital, the hospital can be a field hospital or it can be a general hospital or it can be something in between, right? So we play this critical role that enables these empty buildings for storage and turn them into value-added host for the storage of data and where activities can take place on a value-added basis. So in a nutshell, we play a very important role in enabling these NAND flash chips and turn them into highly specialized storage devices.

Great, yes. So just to let to double check. The -- how you break down your revenue mix NAND controller for SSD versus for smartphone or other technology products? So could you update roughly your revenue mix by application then?

Yes. Yes. So last year, we -- about 50% to 60% of our sales were from SSD controllers, roughly 20% -- 25% of sales related to eMMC, UFS and 10% for SSD solutions.

Yes. So 50% to 60% for the SSD products which are used in the PC and data center?

SSD controllers. Yes.

Yes. And then the 25% of your revenue is for the eMMC/UFS, which means mainly for the smartphone embedded controller.

That's correct. Yes. Yes. And we have 10%...

Yes. SSD solution means you get the NAND chip from the maybe Korean or Japan, U.S. makers and then you put your own controller and then you provide the entire SSD solutions for the maybe China e-commerce company?

That's correct. That's correct. So we build these highly specialized SSDs like our open-channel SSDs for Alibaba. We have our industrial Ferri SSDs for a host of different end markets.

Yes. Okay. Let's narrow down the, maybe, your business environment. How about -- how do you assess the memory chip industry cycle? When we look at the Taiwan foundry company, they have delivered a great -- the double-digit top line growth and the great gross margin, op margin. That's the maybe foundry or logic chip area for the past 10 years. But the memory industry, for example, 2018 was very great. But the 2019, quite severe correction in terms of the ASP trend. And then the 2020, early 2020, some recovery, but with the coronavirus, some risk factors we can see. So how do you assess overall the memory chip industry cycle? Maybe with the coronavirus issue or after the COVID-19 pandemic, yes?

Simon, that's a very difficult question you have there.

Sorry.

The NAND industry is very volatile, it's very dynamic and part of it is the result of this very capital intensive and very competitive market. You have 7 large vendors of NAND flash, you have 2 in Korea, 3 in the U.S., one in Japan and a new player coming out of China. So very competitive, capital intensive. And so every time you have big step-ups in capacity, the difficult entity of matching the capacity to demand and then you have volatility, especially in -- since the start of the year from COVID-19 and how that's been impacting demand, right? So there's some puts and takes, obviously, from what we've seen so far. Q1 smartphone -- those are -- let me step back a second. If you look at NAND flash, 3 primary applications that consume NAND flash. You have data center enterprise market; you have the mobile market, smartphone market; and then you have the third market being PCs and other kind devices, right? So this year, at the start of the year, with COVID-19 starting with China, the whole Chinese market was impacted and so did much of the global tech food chain. And so smartphones got hit in Q1 severely. And going into Q2, the rest of the world being affected by COVID-19 and the virus turning into a pandemic, and we started seeing the impact spreading into -- smartphone sales being impacted to the rest of the world. But on the flip side, because of the sheltering, the lockdowns, working from home and online learning, this is leading to much stronger demand from the data center guys for collaborative tools, for videoconferencing and host of other stuff that you need in order to work from home and online learning as well as entertaining yourself from home, right? So a lot stronger demand from the data center guys. And then in order to facilitate the online learning and working from home, we started seeing very strong demand for PCs and SSDs going to PCs towards the end of Q1. And this is going to carry through Q2, and we're seeing our orders for SSDs extending through Q3.

Yes. That's great. So let me check some details for the -- maybe NAND industry or controller area. So how do you feel the overall channel inventories of the NAND flash chips or NAND controllers? I mean the -- among the module makers who are PC OEMs, the smartphone OEMs or even the hyperscalers. How do you assess overall the memory chip inventory or controller inventories that's there?

On the controller side, our customers, whether module makers, flash maker or OEMs, they don't normally keep inventory or controllers there. Their inventory is sitting as our inventory. So whenever they need controllers, whenever they're building controllers, they -- or building SSDs or other types of storage devices, they'll pull the parts from our stock, right? So we're maintaining that inventory. And so on their side, the inventory is usually very lean and continues to be that case. On the NAND side, inventory levels is harder to gauge because we're not involved in -- most of our business -- a vast majority of our business, 90% of our business has nothing to do with the building of storage devices. So it's very hard to have clarity in terms of what the inventory levels are within the food chain, whether they're the NAND guys, the distributors or the module makers. And so unfortunately, there's -- it's part and parcel of our business as a -- primarily as a fabless.

Yes. Sure. So maybe 2 data points for the NAND flash industry. Number one, the spot market price has been weaker in the past few weeks. And the separate is the Samsung, for example. Samsung Electronics has recently announced their new CapEx plan for their newly constructed fab in Pyeongtaek, Korea. So Samsung is going to be quite active for the NAND chip production, capacity expansion. So maybe, number one, how do you assess the NAND spot market price weakness after the -- some recovery through the late 2019 and then the Q1, but sort of the mid-to-late second quarter, NAND spot market price corrected. The matter of the on-demand issue or supply issue? And then if Samsung increases the NAND chip production, maybe it could be a good thing for Silicon Motion? Or half and half? Or best thing? What's your view, sir?

Well, for our business, the longer-term driver for SSDs, adoption of SSD versus HDD, the continued proliferation of storage devices that use NAND as the key building block for storage of information. This is dependent on NAND availability continue to scale. And so the build-out of more fabs is certainly very beneficial for us and the overall industry. The other bit is NAND prices continue to get cheaper and cheaper. I mean this is the long-term trend, and this comes from new generation NAND coming out every -- pretty much every year of 64 layers to 9x layers, the 128 layers or so, 144 layers and so forth. So this ongoing march of new generations of NAND coming out, this will continue. And with each generation, the NAND, dollars per gigabyte will continue to get cheaper and cheaper. Near term, it's a little difficult to plan in terms of short-term NAND prices because there's a lot of supply-demand shifts. For example, if data center spending continues to be very strong for the rest of the year, while everything else is lackluster, it could be NAND prices continue to be fairly stable. But if the data center spending falls off towards the end of the year, then NAND prices could correct. But these are near term and it's really quite -- the price of NAND is quite volatile, but the longer-term trend is very clear. Longer-term trend is dollar per gigabyte will continue to come down because of cost structure of NAND component. Every generation getting cheaper and cheaper, and because this industry is very competitive, the price curve will match up with the cost curve. And we've seen this happen over the course since start of this industry and will continue to be the case. And so this will lead to greater proliferation of SSDs and related storage devices as well as the displacement of hard disk drives in what's -- in client devices in other parts of the market.

Yes. That's great. And then the -- sorry? Okay. So let's look at the evolution, the growth trend of the NAND industry. 10 years ago, for example, NAND chips mostly used in the flash card and then the partially smartphone. After the smartphone era, we sold the client SSD. And then these days, enterprise solution SSD. So in the -- after the data center theme or enterprise solution, do you see another area for the growth like autonomous driving or some industrial, while, eventually, maybe NAND flash can replace entire hard disk drive everywhere? So could you share your view for the evolution for the past 5 or 10 years? And then the -- for the next 3 years, 5 years, 10 years, which application will lead the NAND industry growth? And then how the Silicon Motion prepares to enjoy the new growth momentum?

Yes. So Simon, you're absolutely right. This -- our industry has gone through a lot of change over the last 10 years. I joined Silicon Motion back in 2007. And then -- this was before smartphones. And so NAND chips were primarily going into memory cards, micro SD cards, USB flash drives and mp3 players. Remember those things, mp3 players? And then very quickly, with NAND chips beginning -- continuing to be cheaper and cheaper, and with robust controllers, NAND was widely became -- widely used in smartphones. So if it weren't for NAND chips and solid controllers, there probably would not be the smartphones that you see today, right? And then subsequent to that, the rollout of SSDs going into PCs and into the data center market. And with the PC side, it's only because NAND chips got so cheap that even from a dollar value perspective becoming as cheap as HDDs and certainly the characteristics of SSDs. I mean it's light, low power, thin form factor. These very -- these sort of characteristics are beneficial for ultralight PCs, made inroads into -- quick inroads into the PC market. And with NAND prices continue to come down, already more than half of all notebook PCs are using SSDs. And in a few years' time, the rest of the notebook market will also convert into SSDs. We're seeing this happening on other types of client devices. The game console guys towards the end of this year, they'll start rolling out. For once -- for the first time, game consoles using SSDs. You also see SSDs going into other client markets, into the external storage, the surveillance market. And with the rapid growth of the hyperscalers, very strong use of SSDs in order to widely distribute content, for transactional purposes, for the -- for example, the last mile streaming of video or in order to handle the very large sets of transactions seamlessly at the e-commerce guys. So very strong uses of SSDs, and we'll continue to see this. There's still a lot of growth in the PC market, in prime market. The opportunity on the data center market is blue sky. And furthermore, as you mentioned, there likely going to be a host of new applications that we just -- frankly, just don't have the vision to see today. But from our perspective, what we can see is the trend of NAND becoming more and more difficult to manage with these successive generation and us, as a controller maker, planning the technologies necessary to continue to manage these and to turn them into upcoming types of applications are necessary for whatever devices that were to arise. I mean we -- surfing back into 2007, we certainly did not see smartphones, but once smartphones started coming around, we knew how to manage the SSDs -- sorry, the NAND chips into the eMMC that were necessary for the Android smartphones and quickly became the market leader and then into subsequent applications, and most recently into the data center market.

Yes. That's great. And then the -- looking at the -- overall the industry competitive landscape, yes, Silicon Motion obviously is quite outstanding company as a fabless company. You design the chips and then you ask this -- the fabrication, maybe TSMC mainly. But your local competitors also doing things and they design the chip and then they ask UMC or TSMC or even the SMI systemic, China. So the similar business model, but how the Silicon Motion can differentiate your NAND controller design and foundry sourcing and then the marketing distribution? And then how you compete with even the U.S. companies like Marvell or others. So what's the Silicon Motion's differentiation or what your management can do better job versus competitors there?

Simon, as you know, our industry is already quite consolidated. There aren't that many number of meaningful competitors I can count in one hand. And part of our success is success that we built up over the last -- the course of the last 10-plus years. We were -- when I joined back in 2007, we were already the market leader in terms of controllers. But for -- back then, the only types of devices were cards and flash drives and mp3 players. So back then, we were already the market leader and developed a host of cutting-edge technologies back then for managing NAND and delivering the types of applications that host devices needed, and then became the market leader in each one of the successive applications that we've entered from the external storage devices to the eMMC/UFS-embedded storage in smartphones and then going into the SSD for PCs and other client devices. Each one of these markets that we've entered, we've been able to use the expertise that we built up over the course of the decade and entered successfully because of the strength of our technology relating to the hardware, which is the -- our ASIC plus firmware and be able to do this very successfully in expanding our end applications and customer base. So we've been -- as the leading supplier, we have the benefit of being able to scale out our technology more broadly than our competitors and because of our skill to stay extremely focused on doing this. So with these successive generation, we're the first point-of-call for customers who want solutions. We -- the value add that we provide to them is we help our customers be -- to stay and become even more competitive in the storage devices that they're bringing out, whether for OEMs or for channel markets. It's all about the value added. And we need to continue to do this to stay relevant. So today, it's mostly related to the PC and smartphone markets. And already -- we've already started moving into the data center hyperscale market, where we've already started showing our initial value added [indiscernible], to grow this over the course of the next few years.

Yes. Yes. Yes. Just a follow-up question should be -- actually, this is -- this question raised by long-term investors, which cover Samsung Electronics, Korean chip makers. Some investors are saying, when they talk with Korean chip makers, they've been saying they want to have in-house NAND flash controller rather than using the Taiwanese or U.S. NAND controller. So the question should be, what risk factor is if Korean chip makers continuously use their in-house controllers and then if they expand the NAND chip production, so they try to combine their own controller, their own NAND chip and then they sell the entire NAND solution product to the OEMs or hyperscalers. In that case, what could be the Silicon Motion's reaction to this? Or opposite way, the memory chip makers -- these memory chip makers, they -- their track record for the logic chip area, not great. So they should use maybe Silicon Motion's controllers. But so far, Korean chip makers using the more in-house controllers. So how do you assess the risk factor with the Korean chip makers or the growth opportunities for the future?

Well, for Samsung, they've been able to have the benefit of scale, right? The largest supplier, they supply at least 1/3 of the NAND industry. And so because of their first mover, they've been able to successfully vertically integrate it. But like everybody else, I mean all the other NAND flash makers, they all see controller technology as critical for turning their empty spaces into -- that can store stuff into value-added storage applications. And then you can only convert your empty space into -- the storage media into a value-added storage device through the controller. So the value-added is straightforward. But the question is, how can you develop the technology that is constantly shifting? New generation flash coming out every year. So this is an ongoing treadmill. With new applications, new categories of applications. You've got controlling NAND, you got to deliver the types of storage devices, the many flavors of SSD that continue to change from SATA 3 to PCIe, NVMe, Gen 2, Gen 3, Gen 4, Gen 5. So this is an ongoing investment that's critical for staying relevant on the device side. You can do it yourself or you can outsource. The value of outsourcing and -- which we've demonstrated to our customers and so it's been -- has led to an increase in outsourcing. Opportunities for us is we can do it cheaper and we can do it better. We can amortize our expense over a much larger base. And we can develop technology with -- we're learning from a host -- with a broader set of experience than if you were just to do this internally, right? So better economics, more experience and aggregation of this leading to us being value-added partner to our customers. But at the same time, we also recognize that for some of our flash partners, there may be certain products that they want to do internally that's strategic, that they feel that they can add more value themselves than to outsource to us. So end of day, it will always be some combination. Some products, they will want to devote their -- use their own expensive resources because they feel that they can -- by doing this, they have more control over their own technology road map and differentiation. While others that are more mainstream, more commoditized, more competitive, they outsource because we help them stay relevant. So we believe it will -- longer term, it will stay -- still remain a combination of certain products. They will seek to use their own resources, and others, they outsource to us. And this has been the case since we started being a controller maker.

Yes. Definitely. That's great. A little bit technology-related questions then. NAND chip makers, they are working on the -- not only the 128-layer 3D NAND, but 1yy node coming. So such higher layers require more advanced controller or not much relevant? Or you can charge extra dollars for the multi multilayer 3D NAND?

Yes. Absolutely. With each successive generation NAND, it's all about getting more storage out of the same real estate, right? And so this is -- continues to lead to NAND that are less and less robust from a storage integrity perspective, from a programming race perspective and a host of other issues. And so our controller technology continues to evolve and become more and more sophisticated. And this has been the case way back when NAND was 2D. As you shrink the geometries and trying to pack more cells in the same real estate, the errors and endurance issues magnified. And going into 3D, as you stack them higher and higher, they're like skyscrapers, taller, the less robust and furthermore within each one of the floors of the skyscraper and then the rooms on each floor, you're building out these bunk beds from one bed to 2 bunks to 3 bunks and now to 4 bunks. The QLC, the quad bit products, right? So with each successive generation, the cost continues to come down rapidly because of the stacking, because of the multilevel cell technology. And these are -- the trade-off is the NAND cell as a storage becomes less and less robust and you need the controllers to overcome these issues. And then on other side, the device side, the devices -- the storage devices using NAND, they want better and better performance. They want faster read/write data rates, right, better endurance and a host of other value added. For example, going from PCIe -- SATA as a PCIe Gen 3, significant step-up in data rate and going from PCIe Gen 3 to Gen 4, again, another big step-up in data rate and other value-added functions from storage device. And so you got this -- you have this widening delta between storage device performance where OEMs and consumers want stuff that are better and better, and on the flip side, the actual NAND storage cells used in these storage devices getting weaker and weaker with these successive generation.

Yes. Yes. Another value-added controllers should be UFS versus the eMMC. So UFS portion now, what, 20%? Or what's the mix ratio between eMMC versus UFS these days?

For us from our first quarter shipments, we are -- already 2/3 of our segment sales already UFS. More broadly, the overall market, the market is still primarily eMMC. Last year, the UFS was just found primarily in premium phones. So premium phones which are roughly -- premium Android phones are roughly, say, 15% of all phones are there on that ballpark. Those were almost all UFS last year and will continue to be. Last year, we started seeing certain -- small segment of the mainstream phone starting to adapt to UFS, and that will increase significantly this year. So we're going to be a big beneficiary of this growth. So in bigger, broader terms, even if smartphone sales were to decline, but because of this rapid transition from eMMC to UFS, the UFS use -- unit growth will be quite strong, and we will benefit from that.

Wow. So already 2/3 of your -- the smartphone controller is already UFS. 2/3?

That's correct, yes, 2/3.

Wow. That's interesting. Yes, we have only couple of minutes. So second last one, sir. How do you see the overall impact from all the U.S.-China trade issue tension? So do you feel any order decline from China big Tencent maker? Or how do you see the impact from the U.S.-China issue these days?

Yes, that's a tough question again. But what we're seeing is likely this decoupling of technology of food chain because of the U.S.-China issues. So China, in order to be self-sufficient, you have YMTC growing. And so this, in our opinion, is more customers and more demand for controllers in China as well as in the rest of the world. So more competition with the rest of the world being more commercially driven versus China being more strategically driven and leading to strong growth in the western market versus China, both strong growth on both markets.

So, so far, no any direct impact from the China handset makers? I mean any order cancellation -- I'll put this way, any order increase for their inventory buildup? No big change?

Well, we saw -- yes, we saw weakness in Q1, but that's more COVID-19-related with the lockdowns in China impacting demand as well as manufacturing and food chain, but everybody is now back online. And so in recent months, we've been seeing pickup in China domestic demand. And certainly, the phones are now -- the factories are all back online and so phones are going everywhere. But the bigger issue is less about the U.S.-China issue versus -- more related to the impact from the stay-at-homes, the sheltering in the U.S. and Europe in Q2 and what that means to smartphone sales and the follow-on economic consequences with rising unemployment and potential issues from that towards the second half of the year.

Yes. So basically, you are still -- you're seeing the second half uncertainties? So management, they don't want to provide any numerical guidance for the second half year?

Yes, it's -- second half for our smartphone, there's -- the visibility is still limited, though from a procurement forecast from our customers, the forecasts are fairly limited. But from a more conceptual perspective, we feel fairly good that our UFS will experience fairly strong growth in the second half of the year, in fact, for the full year because of this transition from legacy eMMC to UFS more than offsetting whatever downside in global handset shipments.

Yes. And then separately, the SSD growth was still on track broadly because of the remote work economy, online transaction increase. At least this area quite safe, still showing the good momentum, right?

Yes. That's right. That's right. So we -- for our SSDs, we -- SSD controllers, we see order patterns through Q3. Q4 is still pretty murky, but the broader theme of HDD transitioning to SSD and that transition more than offsetting downside on PCs, that should lead to strong -- fairly solid SSD growth for the industry as well as for ourselves this year.

Wow. Yes, that sounds good. So we spent roughly 40 minutes. This is great. So how about if you close today's session with maybe your very consistent shareholder return policy, I mean, the quarterly cash dividend. And also you used to have big budget for the share buyback program. So how about if you conclude today's call with some capital return policy. Thank you, Riyadh. Really appreciate.

Yes. Thank you, Simon. It's good to be here, but -- and I'll wrap up with just some thoughts about what we're doing with returning capital to our shareholders. We've consistently been strong believers that our -- whatever we don't need for operational purposes, we'll return to our shareholders. And we've done that primarily through our dividend program and more opportunistically through our share repurchase programs. And so what we've done on the dividend side is, as our business has become more successful, our cash flow continues to grow. We've grown our dividend in line with that. And so as our cash flow continues to grow then we have visibility towards the growing cash flow, our Board may continue to like to increase our dividend in upcoming years. So next revisit of a dividend will be towards -- in Q4 this year. And so this will then depend on what our visibility for our business and cash flow looks like into next year. For our share repurchase, this again is related to business visibility and how our share price is at current point in time. Unfortunately, as much as we love to continue to buy back our share, we just are -- just would not be appropriate to be buying back our shares when we don't have the ability to communicate solid visibility through this year. I mean we've -- we no longer have a full year guidance, and this is a result of more limited business visibility and so this has also affected our decision to temporarily refrain from buying back our shares. So with that, thank you very much, Simon. And it's a real pleasure to be here at your conference -- your virtual conference today.

Yes. Great. Yes. Thank you very much. Maybe see you next time at the another conference. Really appreciate it. We -- it's already over 10:00 p.m., almost 11:00 p.m. Taiwan time. Really appreciate, Riyadh. Thanks, Chris as well.

You're welcome.

Great. Great. Thank you, Simon. Stay safe. Bye.

Yes. Thanks, audience. We are closing this session. Thank you. Thank you, everyone.