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

Okay. Great. Thanks, everyone, for joining us. So I'm Randy Abrams from our Asian semiconductor team based out of Taiwan for Crédit Suisse. And we're happy and pleased to also have Taiwan company, Silicon Motion. And with us, we have Riyadh Lai, CFO. Also joining, Chris Chaney, who's IR. Just for the quick background. I think many familiar with it but a leading merchant supplier of memory controllers, supplying into smartphones and also, PC and enterprise SSDs. So format for today is fireside chat. Feel free any time -- e-mail me questions, we'll try to slide it into the questions. And I think just for Riyadh, I think your camera is tilted 90 degrees.

Yes. All right. Let me...

Got you now. That's better. Okay. So just to kick off -- so we have about close to 30 minutes. The first one to get everyone up to speed. If you could start just now with a high level snapshot of your market position in each of the main segments that you're in, like SSD, mobile and solutions. Just how the mix and market position looks at this stage? And then we can dive a bit further into details.

Right. So we have 3 primary products. We have our SSD controllers, we have our -- serving the client market, PCs, other client devices as well as our initial presence in the enterprise part of the market. We -- our second product line is our eMMC+UFS, serving the smartphone industry, but also IoT and other applications, including Chromebooks. And then our third product line, our SSD solutions. Within our SSD solutions, we have 2 primary products, our Ferri industrial SSDs and our Shannon data center SSDs. So our Shannon products are primarily for Chinese hyperscalers, including Alibaba and Baidu. Our Ferri industrial SSDs are going into a broad mix of applications, from traditional industrial usage to commercial equipment, to data center boot code storage, to automotive, so quite diversified. For these 3 products, our client SSDs, we are the merchant market leader. We ship more and more than any other merchant supplier out there. And our market share is almost as big as the leading NAND flash makers market share. So we have about 30%, 1/3, somewhere in that ballpark of the client SSD controller market by far, the merchant leader. In the eMMC+UFS, mostly for the mobile market, we are also the merchant leader. Our market share is somewhere in the mid-teens, somewhere in that ballpark, low to mid-teens.

Okay. Great. And on the business outlook, to recap that, third quarter, I think, originally, middle of the year, you guided for a bit of a moderate decline, but then it ultimately actually topped guidance by good 5 points. So things look like it picked up some through third quarter, and then fourth quarter when you guided another 3% to 10% sequential rebound. So if you could go through maybe what you saw midyear kind of what disrupted the business a bit? And then now how it's closing out the year, given the guidance -- with the improving trends you're seeing driving some returns?

Yes. Yes. At the start of the year, so this is prepandemic. We had pretty good visibility for the rest of the year, and everything just went haywire. So this is turning out to be a very complicated year. But nevertheless, we're still driving pretty solid growth through all of our product lines, not to mention our overall top line growth. But the growth of our products have become a lot more lumpier than we had expected. We got impacted on the smartphone side from COVID, affecting China, and then the rest of the world in Q2, resulting in very difficult market dynamics for our large customers to predict. So leading to inventory building and the need to digest the inventory and then impacting our forecast for our eMMC+UFS. On the other hand, our larger client SSD product line got impacted by flash allocation, as the world went into hibernation, working from home, learning from home. This resulted in very strong surge procurement of data center SSDs and driving NAND allocation away from client devices temporarily towards the hyperscalers, so it impacted our business specifically in Q2. But that has all gone away. And so we saw the hyperscaler surge procurement ending -- ended pretty much midyear. And so that freed up a lot of flash for our client SSD business in the third quarter, and so gave a strong boost and availability of flash going into client SSDs, with our flash vendors reallocating towards this type of application. Not to mention PC demand continued to be very strong. So back to your question about our Q -- your comment about our Q3. Pretty solid, better expected growth, primarily coming from our client SSD product line, driving -- being driven by the better availability in NAND as well as continued strong pull from PCs because of the pandemic-related purchases of some PCs by consumers and corporates. Our eMMC+UFS, on the other hand, pretty much performed as expected. We had very sharp growth of that product line in Q2, despite the end markets for hands is dropping off but -- and so that led to a situation which we had foreseen in Q3 of needing -- our customers needing to digest that. So we're now seeing a pretty good progress in that end, so progress towards a rebound of that into Q4.

Okay. Great. Usually, it seems, if we look at the overall NAND market, I feel like the Goldilocks situation is probably the best one. And in a way, I think from a -- if we get a little bit of oversupply that keeps pricing coming down, triggers elasticity, that's typically the best for the unit side. If you could give a snapshot, just as we close out the year, how the broader NAND market is looking in terms of the balance for fourth quarter and then setting up for 2021? Because sometimes the things out of your control, but the broader NAND market, kind of feel what's happening in your business?

Yes. Yes, there are a lot of moving parts. The basic dynamics that we've been seeing, beginning with Q3, and will likely play through Q4 and also the first half of the year. The hyperscaler procurement of NAND and the search procurement ended midyear. So going to the third quarter, freed up, and so led to softer NAND pricing. So we saw NAND pricing beginning to slide in Q3 modestly, and continue to side modestly in Q4. And likely will probably also slide modestly in the first half of the year, which, as we all know, it's principally the soft season of the NAND industry because of a lot of it being -- going into consumer electronics. So this will continue to, we believe, play out near term. The other dynamic that will play -- come into play is the continued expansion of NAND capacity, especially the continued rollout of the next-generation NAND components with even lower cost structure. Now these are the NAND components with even higher layer counts, 3D NAND with higher layer counts. So with the higher layer count, the cost structure comes down. Dollar per gigabyte will continue to decline. And we're seeing a lot of the next-generation being planned for significant capacity expansion. So one of the key factors that we need to continue to monitor is how smoothly they're able to roll this out. So paradoxically, the smoother they're going to be the NAND flash guys are able to roll it out, that means more 3D, more bits, more stronger bit growth, and that could have an impact on NAND prices negatively. But at the same time, those who are able to do it earlier have better yield with the next-generation versus their competitors. We'll have cost advantage. So it's to each one of the vendors incentive to be the first to achieve better yields.

Okay. And switching gears to SSD. It's -- I believe, we can run over half of your revenue. In that category, it looks like a strong year. I think on the last conference, you mentioned pacing 20% year-to-date versus probably broader markets, maybe 10% growth. Could you address after these couple of years strong growth, where we're at now for penetration? And do you feel we're starting to get to that maturity stage? Or are there some things on the horizon that could continue to drive pretty decent growth for this business?

Right. So absolutely, this year, we're growing roughly 2x the market in terms of SSD controller sales, twice as fast as the overall market. So in this market, the end applications are client devices, of which a little over half are PC-related. So on the PC side, the adoption of SSD is much higher than other categories within the client devices. Other two categories are consumer electronics, including game consoles and surveillance and other external storage classes. And so these are less penetrated by SSDs. So initially, the rollout of SSDs began with the lower density application. So PCs, in general, using -- able to run with lower capacity. And so to transition from HDD to SSDs happen a lot faster. And when you go to game consoles, generally higher capacity and then external storage, even more capacity. So the higher the capacity, the more expensive NAND becomes. And so it takes a little bit more time for NAND prices, dollar per gigabyte, to come down, so we get the inflection versus hard disk drive. So already in client devices, PCs, we've already seen pretty significant transition. Notebook PCs, all very high usage of SSDs already, especially on commercial notebooks, but also pretty high on consumer notebooks, too. Desktops, a bit lower, but that's already also happening. The game console guys finally making the move into adopting SSDs. So this will play out over the next few years. The other consumer electronic surveillance, we're just taking baby steps into that. So there's a lot more availability of opportunities for client SSDs. So in terms of -- if you're -- put a number behind that, we're a little over half through the transition cycle, right? So 500 million client devices, a little over half through the transition cycle. So plenty of opportunity for us from an end market perspective. But for us, it's not just the end market, but also market share gains from -- that will be driving our growth.

Okay. And I think we can expand on both of those. Just one more end market, where we're coming off a PC or where notebooks probably grow 20%. And there's a lot of different views on how it looks postpandemic. But if you could give a perspective of how your customers are viewing it now for next year. If they're looking at it in terms of a growth year or we're factoring in some deceleration as we come out of -- or hopefully come out of COVID a bit by second half with the vaccine?

Yes. Yes. So relating to PC demand, it's helpful to have wind behind our back with continued strong PC demand. But in a bigger scheme of things, from a sensitivity perspective, that's not the primary driver of our growth. The primary driver of growth is the displacement of HDD. So even if PCs were to, say, flat growth, down 5% or up 5%, that doesn't move the needle very much. What moves the needle a lot more is continued adoption of SSDs versus HDDs. And on top of that, our continued market share gains.

Okay. And so that -- if you could talk a bit about it. I think you have talked about an additional NAND customer. You've already had four, and potential for a fifth one. I think if you could talk of dynamics where you're at? And what could drive that additional share? I think in the opener, you said about 1/3 market share. But maybe timing for picking up that additional NAND customer, timing to start driving toward 40%, if that's still the case?

Yes. So when you look at our business, we have 2 classes of customer. We have the module makers and we have the NAND flash makers. And through these 2 customers, through these 2 customer bases, we're going into the channel market and also into the OEM market, PC and other device OEM markets. And so these -- the flow of our controllers through the SSDs to -- for example, the PC OEM market, is flowing through -- primarily through the NAND, our NAND partners, but also through our module-maker customers, like Kingston and others. So when you look at our market share between these 2 segments, we have very high market share in the module into the channel market. In some of the channel markets, for example, the regional [indiscernible] market, our market share is somewhere at least 75%. But when you're looking at the PC OEM market, our market share is much lower. So when you balance the two, you get to that 1/3 sort of market share. So there's a lot more opportunity for us on the PC OEM side of the business. And when you then look at our products, where our products are now headed, we have extensive PCIe Gen 3 and also now also PCIe Gen4 design wins. Our Gen3 will continue to carry the weight of the company through next year. Most of our growth next year will still be coming from Gen 3. And these Gen 3, the primary growth drivers, will be from our NAND flash partners, with their products primarily going into the PC OEM, where we're able to continue to gain market share. And then we have the Gen 4 products coming in. Our Gen 4 products don't start shipping until mid of next year -- middle of next year. And with that, we've picked up a Korean customer. I'm delighted to talk more about this. I mean, this is a -- this project is an expansion of our long, and deep, and broad engagement with this NAND flash partner of ours. So we've been working with them for a long, long time. We provide our customer with a very compelling solution that enhances their the SSD competitiveness in the marketplace. And this engagement of ours, it's supported by our Korean team, led by our General Manager there, who has a decade of senior executive experience at this NAND flash partner. So we now have very deepened our tentacles into our -- this customer. And this design win also expands the breadth of our business engagement with this customer, and will deepen further when they complete their recently announced acquisitions. So we're very excited about where this is all headed.

Okay. Who do you think triggered the change? Like in the mobile business, that customer started with merchant, like your solution on the eMMC and then went in-house. Now it looks like the opposite and always the fear was that the captive would take over the business to their own controller. What do you think the trigger was? And how I guess, how sticky in terms of -- is it still going back and forth where they evaluate in-house versus now an engagement where they start pushing a lot more toward merchant?

Well, our customer has been using their own controller, and now they're expanding to use merchant for an SSD controller. So this is, in fact, a movement the other way. With our previous engagement on the eMMC side, it was we supplying all of the eMMC, and they wanted to diversify the supplier base. And so rolled out their own controller to not be completely dependent on sole supplier basis. So this -- it's good to have a healthy mix of vendors. And when there are not too many merchant suppliers that can provide the same services, some parts of their program will be dependent on internal programs. But at the same time, we are very well-placed for continued expansion, especially with their current acquisition, which will be completed shortly. Intel, their client SSDs, are all running on our controllers. We've gone through many generations of it -- of their products. One of the reasons why our customer is keen to do this acquisition is because of Intel's QOC technology leadership. All of Intel's client SSDs today are using QLC technology. Intel is TLC technology leadership. They supply more TLC parts than any other vendor out there today, many generations of mature TLC product in the marketplace. And today, all of these are using our controllers. So this will continue, and we expect that as long as we continue to deliver on this, we'll be very well-positioned for the years to come because of our ability to very effectively manage QLC technology. Something that we believe very few people -- very few, whether internal, external parties, have the capability of doing well.

Okay. And it sounds like you'll pick up share once PCIe4 gets going, that's where you pick up the new design win. I guess, two questions on that transition. First, I mean, over here in Taiwan, [indiscernible] made a kind of a statement about being early with the AMD platform. I'm curious how your engagement, if you would engage more on the EMD side in the future? And as we go to PCIe 5, we'll play out any differently, we'll still be really ramping up with the Intel platforms? Or do you see getting good market share on both CPU platforms?

Yes, absolutely. We're agnostic. I mean, we're really timing more for when the higher volume starts rolling out, right? The AMD part came out earlier, but the volumes are small because it's really targeted more for the desktop applications versus the higher volume notebook rollout. So it's -- where we're at is now the beginning of the rollout of the notebooks. They're able to use the Gen 4, right? So this is now tied with the next-generation of AMDs as well as Intel's 11th generation chipsets, right? So they are designed specific for this. When you look at even on the controller side, until the new wave of controllers starts coming out, where we're at the forefront, the previous generation of Gen 4 controls were not optimized for notebooks. They were very power hungry. I mean, they had the Gen 4 interface. They had the throughput speed, but we're not designed to be able to be used effectively for notebook application because of the high-power usage. But now with ours and some of our -- some of the other vendors' products, the Gen 4s are now being optimized more for notebook. So we feel we are well-positioned and Intel and AMD's parts, the next-gen that are being optimized for Gen 4 for notebooks, these are all -- there will be more generations, gradual expansion of parts that are going to be even better tailored for the industry.

Good. One just last question on the SSD side. I'm curious if you're seeing any effective capacity from China for YMTC, it does look like 64-layer working, even getting to sample and product on 128. So to the extent they're hard to ramp in capacity, how you see that coming in? And also your position if it starts to be some capacity into the market?

Yes. We've been seeing supplies of YMC parts -- both YMC components to module makers as well as directly from them. We're obviously working directly with them as well as the module makers who are using YMTC flash for many different applications. I mean, we've been shipping controllers for YMTC-based flash SSDs -- client SSDs going into some of the leading -- some of the top 5 PC OEMs, China-specific PC -- notebook PCs for quite some time. And so it's -- the parts are good. They're able to work in SSDs with our controllers pretty effectively for client -- for PCs and also for smartphones. And they're seeking to expand that into the data center part of the market, too. So it's been a lot of work on YMTC. There -- the next part for them is to get their 128-layer into capacity ramp. So we're all waiting for that to happen. And if that happens, more big growth. So another large customer that we can support. So very excited about where this is all headed.

Okay. Great. And I guess to give a few already getting close to half hour but Into the enterprise business, I just want to go through a couple of channels. But first, your top customer, how the ramp profile? I think you've talked about one program, but you could get up to three programs. But how you see the expansion with Alibaba? And then maybe address after that, the additional other potential customers?

Yes. So let me step back. I mean, for us, when we talk about the enterprise SSD market. There are 2 parts of our business. The first part is the controller side of the business, and the second is the solutions -- enterprise solutions, which being led by our Shannon SSD platform, right? So on the controller side, we've been rolling out controllers that we're then battle testing using our Shannon platform into Alibaba, and then Baidu, and then other Chinese -- and other Chinese Internet companies. And we're -- and we then have rolled out these parts available for other customers, including taking a page out of our successful client SSD playbook, which is about providing integrated hardware, plus firmware capabilities, that allow our customers to very quickly bring products in the marketplace. So we have -- we've rolled out the first integrated turnkey hardware firmware solution that are available for customers to use. So this is an exciting trend for us. But back to your other question about our Shannon business with Alibaba. This year, a year of growth for our Shannon platform. We have one project, similar to last year. We're now working on our third-generation of our open channel SSDs. These are all using our controllers. Our firmware, our open channel driver software, working on the third-generation products for delivery next year. And currently, we're working with -- on three projects versus one project this year. So we should see expansion of our presence -- continued expansion of our presence within Alibaba.

Okay. And just in that, I think when you talked about the other part of the Shannon business and Ferri, which is your kind of goes into industrial auto, maybe there's more competition in those areas. Are there things you're doing to -- that you could do to reposition or to get better profitability just in the face of more competition?

Well, on the Ferri SSD, in fact, we have pretty good profitability that are in line with what you would expect from industrial SSD products. We -- on the Ferri side, we're targeting a niche market that sits between the capabilities of our module-maker customer. We don't want to compete against them. We supply controllers to them for their industrial SSDs. We are also expanding our NAND flash partners' reach into markets that they rather not service. So where we said is in the tailored niche applications, where the module makers don't have the capability to do customized firmware, these are high-touch programs where the NAND flash makers don't have much interest in doing. So by us buying their high-end NAND, and then expanding it into the industrial market, thereby, expanding the overall address of the market, it's beneficial for everybody. Where we've been struggling more on profitability has been with our Shannon non-Alibaba programs where we're delivering standard NVMe SSDs. When we're providing standard NVMe, these are enterprise-grade standard NVMe SSDs, where we essentially competing against our flash against like Samsung and Intel, which is not optimal. But unfortunately, until our open channel SSDs continue to scale and are able to cover our Shannon operating expense, we need to do some in order to do that. I mean, it's a tough situation that we have. We can either take a loss on our Shannon platform by not doing the standard NVMe or we do the standard NVMe to cover operating expense, but suffer lower margins.

Okay. And just before we wrap up, one final question. We didn't touch on the mobile business, but I think that is an area you feel pretty optimistic about for next year. So if you could maybe recap the drivers. Then what could really power some of the growth in mobile? Just aside from smartphone, but it seems like from your customer base, there is some momentum there.

Yes. Oh, yes. Yes. So the smartphone industry is going through a big technology transition from legacy eMMC mobile-embedded memory to UFS Mobile-embedded memory. So last year, approximately 1/3 of all phones had already made the transition. Initially with premium phones. Premium phones because of a high data throughput required to go from devices, from the device and where information is coming, primarily camera. The higher-end phones have the high-spec cameras, with very high frame, or say, 8-K video at 60 frames per second. That type of throughput rate, data throughput rate, you can't use -- it will be lagging if you standard NVMe -- standard eMMC. So hence, the move to UFS -- adoption of UFS. So last year, 1/3 of all phones, primarily high-end phones, started adopting UFS. This year, that's going to -- towards about half of all phones as these high-end cameras may make their way into mainstream phones. And then we'll see more -- a lot more growth next year. So that's the key driver for our eMMC+UFS growth. This year, despite the lumpiness, despite the temporary issue that we had with inventory in Q3, even when you factor that in, our eMMC+UFS year-to-date is growing by 70%. So pretty strong growth, and we expect further growth because of this next year, because of adoption, because of our NAND partners' top-notch products, including their very high-performance DDR5 mobile part that's going into the UMCP versions of these products, targeting the mainstream phone market.

Okay. Great. Yes, Riyadh, I think with that, we've run out of time. But I want to thank you for joining the conference. And thank you everyone else too as well for listening in. Thanks so much.

Great. Great, Randy, pleasure being here. Thank you.

Okay. Thank you, Riyadh.

Thank you. Bye.