Thank you for joining us here on day two of the 26th Annual Needham Growth Conference. I'm Quinn Bolton, I cover the semiconductor stocks for Needham. It's my pleasure to host this fireside chat with Silicon Motion. Silicon Motion is a global leading supplier of NAND flash controllers for solid-state storage devices, including SSDs and embedded storage.
The company supplies more SSD controllers than any other company in the world for servers, PCs, and other client devices, and is the leading merchant supplier of eMMC and UFS embedded storage controllers used in smartphones and IoT devices. Joining me from the company today is Jason Tsai, VP of Investor Relations and Finance. Jason, thank you for joining us.
Thanks for having me, Quinn. Always love being here.
Yeah. We love having you here as well. Especially since you went a little silent there for a while, and...
We did.
We'll get into that. Now, I'm gonna start with the sort of elephant in the room, and I know this is gonna be something you probably can't say a lot about, but, you know, as many of you know, as the company previously announced last fall, Silicon Motion filed its notice of arbitration against MaxLinear for willful and material breaches of the merger agreement signed in May of 2022. The company is seeking payment of termination fee of a $160 million, further substantial damages, interest, and costs. Jason, I know it's a private and confidential process. What can you say, if anything, about the process?
Yeah, I mean, obviously, you know, we believe strongly that there was no... We didn't do anything wrong. And so, you know, the termination fee is certainly something at the very least, we're plus all the substantial damages because, you know, again, we did nothing wrong. There was a valid signed contract that we had together. This is a process, to your point, it's a confidential process, so there's very limited things that we can say about it. As you know, this through the Singapore arbitration process, typical timeframe is that 12-18 months.
So we filed for arbitration in early October, so the clock starts from then to 12-18 months. You know, at the time, you know, we pick a judge, they pick a judge, and then the two of them pick a judge together, so there's a tribunal.
It does take time for the process to progress. I am trying to provide as much clarity and transparency as I can throughout the process, but given that it is a confidential process, there are going to be limitations in how much and how frequently I can address things, but obviously, only up until what the latest public comments were, which is what you had just summarized.
Got it. Okay. So moving sort of to the business, you know, I guess maybe before I get into, you know, the bulk of my questions, from May of 2022, when you signed the agreement, until the, you know, the deal was terminated in late July...
Mm-hmm.
You guys obviously didn't say a lot to the street. As you've been re-engaging with the street and investors now for the past about four months, what's been your key message to investors? What should we know about SIMO? Did you go to sleep for those twelve months? You know, what can you tell us?
Yeah, I mean, look, you know, our team has been so focused on delivering and executing. You know, we continue to build the, you know, controllers that are better, faster, more cost-effective than anything else out there, right? Some of you guys may know my background. I started with Silicon Motion about 1five years ago.
I left about five years ago, went to a couple of other places, and then I came back last year, initially as a consultant, to kind of help them through the process until they got bought out, and then in July, September, you know, when things blew up, I came back full time in September. So when I came back, I sat down with all of our product managers or business leaders, and to get an update from what I had missed over the last five years.
And the funny thing is, after my meetings, I was sitting down with our CFO, Riyadh, and yeah, he's like: "What did you learn?" I'm like: "Kind of, you know, what I learned was, like, nothing's changed. I'm meeting with the same people, sitting in the same rooms, talking about the same things. The deck is actually the same.
The only thing that's different is, you know, the interface speeds have changed, the number of customers have increased, and we're doing more with each customer than we were doing then, and obviously, revenues are a lot bigger today than they were, you know, five years ago when I left." And that, that goes to the durability of our model, right?
We've always built a business that was predicated on providing controllers that were better, faster, cheaper, than anything else out there that was available, whether that's in-source or outsource. Because we've been able to consistently deliver upon that, we've been able to grow our business consistently over the years. You see higher penetration of SSDs to now where it's near ubiquitous. You have more cell phones today, you have more PCs today than ever before.
All that, all of that has led to our business continuing to grow and scale over the last many years, and that's where we continue to execute and deliver. Going forward, I think there's a lot of opportunity here for us to continue to gain share. Cost of development of controllers is getting more expensive.
You know, we talked this last quarter about three new controllers that we're developing, taping out in fourth quarter of last year and first quarter and second quarter of this year. This is two PCIe Gen 5 and a UFS 4 controller. These are all being done at 6 nanometer.
Each controller costs, development cost, is north of $15 million. So, you know, that really changes the math and the calculations, not just for internally developed solutions, but also for a lot of merchant suppliers who don't necessarily have as strong of a balance sheet and P&L to support that type of development, to provide a portfolio of controllers that you need to really, really a robust competitor in this market. So that's really where we continue to drive value. The natural progression of NAND and interface for us has always been the same, right? The interface speeds get faster and faster.
The NAND, in order to get cheaper and cheaper, gets weaker and weaker. So that bridge that we provide the controller for, gets bigger and bigger, and the controller needs to be more and more robust, and that's what the value proposition that we've always brought and will continue to bring.
Perfect. You guys had a nice announcement on 5 January 2024, when you announced the earnings date for the fourth quarter. You also announced that preliminary results for the fourth quarter would be slightly ahead of revenue and gross margin guidance.
Yep.
Maybe spend a minute, if you can, what drove the upside, both on the top line as well as gross margin in the quarter?
Yeah. So, you know, look, I think if you, we haven't provided a whole lot of clarity around that, but I think in general, if you take a look at our gross margins, they're driven by two things: pricing and mix. Our products, if you take a look at, you know, just kind of broad strokes, SSD controllers tend to be a little bit above corporate average. Embedded UFS, eMMC controllers tend to be a little bit below corporate average. So, you know, depending on.
You know, obviously, we've all seen different data points around strength in the fourth quarter. What we had embedded in our original guidance for the fourth quarter was that our smartphone-related business, eMMC, UFS, would grow sequentially, but our SSD would be flat.
And so I think there are certain dynamics throughout the fourth quarter that, you know, would give you data points that said, you know, certain segments performed a little bit better, and that would result in stronger revenue for us and certainly stronger margins, too.
Got it. So sounds like SSD may have had a positive contribution, both in the top line, and the margin line, relative to the initial mix expected in the quarter.
Yep.
Got it. I want to get to some of the near-term dynamics. On the third quarter call, and, and, look, this has been true of pretty much everybody in industry, you talked about inventory levels. I think you're saying inventory levels, you see normalizing, you're seeing sort of, you know, orders and, and the environment, you know, moving in the right direction. I assume that that continued in the fourth quarter, given the, the positive pre-announcement. But, you know, any, anything to call out in terms of order trends or, or inventory levels, you know, in, in the end markets as, as you sit here today?
Nothing to call out specifically, but certainly, you know, as we've been saying for a while, as each day passes, inventory in the channel is going to get a little bit better, right? Our customers, their customers are going to work through more and more of their inventory as each day passes, and that's what we continue to see.
Over the last two quarters, you know, when we reported third quarter results, you know, you saw our inventory levels come down from $300 million to roughly $200 million over two quarters' time. Not only are we seeing inventory improvements in the channel and our customers, but we're also improving our own inventory levels, and that's, you know, something that we expect to continue to work on.
No question. Any, any thoughts when you think that might normalize, at least in the channel, in terms of inventory levels?
You know, I think there's always going to be. There's still some pockets where it's a little abnormal, but I think by and large, I think we're seeing, you know, most folks kind of working through most of that already.
Okay. Okay, perfect. You know, any thoughts on the, on the China smartphone market have been soft? Seems like there have been an uptick in activity. Do you think we're getting back to sort of end consumption now for some of the, the PC and the mobile markets? Do you think there's potentially a rebuild of demand after, you know, a period of, you know, pretty weak consumer demand for the last, you know, 12-18 months?
It's tough to say. I think right now, I don't think anybody's getting too overly optimistic about where 2024, what 2024 does. I think there's probably a little more pessimism around the smartphones, maybe a little more optimism around the PC side. You've got a number of dynamics on the PC side that create potential tailwind, right? Which is, you know, you got the sunsetting of Windows 10, you got the introduction of Windows 12, you've got, the COVID corporate refresh cycle happening, and, you know, if you believe all the hype around CES, there's this AI PC massive upgrade, you know, coming upon us.
Push of a button.
Push of a button. Our lives getting easier. So, you know, again, I think all that, what that ultimately means is there's a little bit more optimism around PCs, and that's great. But, you know, I think—I don't think anybody is in a position or confident enough at this point, this early juncture in the year, to really call any sort of meaningful recovery yet. I think as we all kind of digest more and more of the year and we see more data points that support that optimism, we can all be a little bit more optimistic. But right now, I think cautiously optimistic is probably the right tone.
Got it. Makes sense. One of the questions we get a lot, especially, is NAND pricing starting to recover? How is SIMO affected by rising NAND prices?
Yeah. So, you know, I think you guys have seen our gross margins. We dipped down to about 42.5% earlier this year. That was the case in the third quarter as well. We guided for gross margin to improve sequentially in the fourth quarter and at the midpoint at 43%. But, to your point, our preliminary results said we came in above our guidance range.
With NAND flash pricing, with the NAND flash market being in such dire situation throughout, you know, most of 2023, you know, a lot of our customers were going through a significant amount of pain and suffering, right? Negative gross margins for extended period of time. We hadn't seen this level of pain and suffering for a lot of these customers.
Last time was really global financial crisis back in 2008, 2009. Coincidentally, that was when the last time our gross margins dipped below our corporate average of 40% to 50%. As pricing has improved, you know, we're gonna be able to gradually improve our pricing to get it back more in line with historical norms.
As we introduce new products, things like the PCIe Gen 5 and the UFS 4 controllers later this year and into next year, that'll be a catalyst for improving gross margins as well. So we're confident that we can get back to that 48% to 50% over time, and certainly provide more color around that when we report fourth quarter results here in a few weeks.
Got it.
So it sounds like, unlike maybe the NAND manufacturers, they can see some pretty sharp moves. In margins, you'll see a more sort of stable to sort of longer term benefit from that price environment.
Okay, got it. I wanted to, to sort of ask a couple of questions, you know, the merchant versus captive question. I'm sure you've gotten this, you know, just a few times over your, your multiple years at, at SIMO. You know, recent concern from investors about Micron's decision to in-source UFS 4.0...
Yep.
How does that, you know, kind of affect, Silicon Motion? Can you talk about your business with them in UFS 3.1?
Sure.
Other, you know, SSD projects, you know, is this something you could absorb? And to the extent that they've in-sourced that, how does that impact your revenue over, you know, and over what period?
Yep. Look, you know, we're a merchant controller supplier. You know, our business is going to be driven off of a combination of wins and losses. You know, we're not a monopoly. We don't own 100% of any customer, so, you know, our business is... and how we've improved our business is, you know, we have more wins than losses, and that still continues to be the case today. With, you know, Micron, sure, it's disappointing that they're insourcing UFS 4, but it's a very broad-based customer.
They buy a number of products, including UFS 3, UFS 2, as well as a number of SSD products. So, you know, it's a very diversified business with this customer. It's not, you know, all or nothing. We see a trend towards more outsourcing.
You know, we talked in the last quarter about winning a large flash maker that historically has not outsourced controllers. So, you know, this is an ability for us to continue to gain share. I think as the market for SSDs, you know, saturated, I think, you know, flash makers are looking for ways of expanding their TAM, expanding their presence, and going after more and more end markets.
We're the most cost-effective way for them to expand their reach, especially with, you know, Gen 5 controllers, you know, north of $15 million a controller to develop, that's, again, all the more reason why our value proposition makes more sense. So, you know, we're never gonna be 100% of the market, but we're, you know, we've got a strong history of being able to win more over time, and that's what we'll continue to do.
And so, you know, we're, as, you know, I mentioned, we're taping our UFS 4 controller. That'll come to market later this year and ramp into 2025 when UFS 4 becomes more mainstream, moves from the flagship category to more mainstream products. We've got a new customer that will—several new customers that'll launch with that product, flash maker as well as module makers, and we believe we can continue to grow our eMMC and UFS business over time. You know, eMMC to UFS transition happened, you know, probably eight years ago, and eMMC is still probably 20% of the market today.
So, you know, the fear that, you know, UFS 4 is all of a sudden magically going to be 100% of the market overnight, and any sort of business we had with Micron in the past would go away overnight, I think is wrong, and this is, these, these technologies, these interfaces, typically last for a very long time.
I assume you get some kind of ASP lift as you go from UFS 3.1 to 4, and so as your new 4 controller comes out late this year and ramps to 2025, there may be some ASP lift there that probably could be a tailwind that might help offset the Micron situation. The other thing I want to ask you, you'd mentioned this new NAND OEM that historically had not outsourced controllers, where you've got a win. Do you expect that to be meaningful? Is it kind of, "Hey, it's our a toe in the water, we'll see how we execute, and then potentially opportunity to expand further at that customer?
I think there's an opportunity to expand further with that customer. We don't believe this is kind of a one and done. So we see the ability to work with them on subsequent generations of the product and new products, et cetera. So we're confident on that. In terms of ASP increases, you know, the. For example, on the PCIe Gen 5, the first one we taped out, it's an 8-channel high performance PCIe Gen 5. If you take a look at a comparable Gen 4 controller, you're looking at roughly double the ASP.
That's probably on the higher end of the premium that you get for a new controller, but, you know, you're looking at $5 to 7 for a Gen 4 controller, moving to $10 to 15 for a Gen 5 controller, for that high-performance eight-channel product.
So, as the industry continues to evolve over time, those ASPs will get better because it's a more complex product that we're designing. It's a bigger chip. Again, just a more robust solution, so naturally, that ASP gets higher and higher over time.
Got it. That was sort of my next set of questions on the SSD controller side. Maybe first, you know, what's the mix today of that SSD controller? How much is PCIe Gen 4 versus SATA, you know, in the mix?
The majority of our business here is Gen 4, because that's where the majority of the market is, right? We still sell a SATA because there's still a little bit of SATA in the market. We still sell Gen 3, because there's still some Gen 3 in the market. But the majority of the market is Gen 4, and that's, you can expect that that's the majority of our business as well.
Okay. And when do you see the Gen 5 impacting the client market?
So, you know, Intel and AMD are expected to release their notebook chipsets that support Gen 5 late this year and ramping in 2025, and so that aligns with the timing of when our products will come out and scale with that broader adoption.
Okay, perfect. And then can you update us on your efforts to penetrate the enterprise SSD segment? Because I'm sure the pricing is even better in that segment.
It is even better. You know, we're sampling today. This is our MonTitan PCIe Gen 5 product. This is a controller that we designed from the ground up specifically for the enterprise market. You know, we talked to a variety of different hyperscalers and server companies, and enterprise guys, and to understand what it is that they were looking for, how do we differentiate? How do we build something that was compelling for them?
And we took that information, built a new solution from the ground up specifically for that market, and that's our MonTitan product. So as I said, we're sampling that today. We expect to start seeing early revenue late 2024 and then ramping more meaningfully into 2025. As we get closer to that, as we have better clarity and visibility on the cadence of that ramp, we'll provide more color.
Perfect. How important is QLC NAND to you, especially in the client device? You know, I think it's gonna be more than 80% of client SSDs in next year, 2025. And data center market adoption seems to, at least according to Gartner, you know, follow in 2026.
Yeah.
You know, where are you in QLC NAND? Do you feel like you have a leadership position there?
We definitely do. I mean, you know, we've always been able to manage that leading-edge NAND more successfully. And QLC, as you guys know, is a lot weaker, right? You know, as you're trying to stuff more bits into a finite amount of area, you're gonna create weaknesses, you're gonna create, you know, signal integrity issues, you're gonna have data protection, data loss issues. So there needs to be a lot more error correction. There needs to be a lot more data buffer, data, data protection capabilities built into this.
So this is something that we've been working on for a very long time. We have a leadership position in QLC, and we expect to continue to maintain that. The increasing proliferation and adoption of QLC is only beneficial to us.
So you go to QLC, as you go to UFS 4, PCIe Gen 5, does the competitive landscape become narrower, or, or is it the sort of same cast of characters, you know, everybody making those, you know, sort of leaps to the next version of the standard?
I think it becomes smaller, right? You know, when I started here 1five years ago, there was, you know, half a dozen, 10 different controller companies that I would get asked about regularly. Now, there's really us and one other left. I think, you know, when it costs more than $15 million to tape out a controller, there aren't a whole lot of companies out there that have that balance sheet and P&L that can support that type of investment, especially to build multiple products to fulfill a portfolio of solutions that they can offer their customers.
So we think that the landscape will continue to get smaller, because it's harder to do, not just from a cost perspective, but to be able to build controllers with the data protection and the RAID, and all that stuff. It's very complex technology that, you know, it's not readily available to anybody out there. And so, as the NAND gets weaker, coupled with the performance getting greater, there's just more and more technology needs to be stuffed into a controller that fewer and fewer people are gonna be able to do.
I wanted to ask, in December, the company established two new business units, Client and Auto Storage and Enterprise Storage and Display Interface Solutions. What drove the decision to re-segment the business?
Look, I think, you know, more so than it was, you know, a focus on the enterprise side, right? And, you know, that coincides with the opportunity that we see, the ground-up development on MonTitan and, you know, where, and the confidence we have about scaling that business. We wanted to have a dedicated executive responsible for those businesses, and so having carved that out to a dedicated Alex Chou, who's joined us recently as well, I think made a lot of sense to demonstrate, and not just for investors but also for customers internally, that this is something that we're highly focused on.
Okay. Will you re-segment revenues according to those buckets, or will you continue to give the straight, the eMMC and UFS grew X% and SSD controllers grew Y?
Yeah. I mean, today, you know, obviously, what the division that Alex is responsible for is relatively small...
Okay. Yeah
Compared to everything else, right? So it really doesn't make sense to segment that out at this point. Certainly, down the road, as enterprise becomes bigger, we will consider more transparency and more disclosure, but right now, it doesn't make sense.
Got it. A few other questions just, you know, outside of PCs and, phones, can you talk about the company's traction in automotive, IoT...
Yeah
C ommercial, industrial? You know, how do those markets, how are they looking?
Yeah. You know, we've spent a fair bit of time on the last earnings call talking about the automotive side, right? So, you know, we see wins in automotive across all of our business groups, in eMMC and, and UFS, to SSD controllers, to SSD solutions.
As you guys know, obviously, cars are becoming more intelligent, not just stronger infotainment systems, but, you know, whether it's autonomous driving, EV vehicles, just a lot more data processing, a lot more need for storage, and so that increases the number of storage devices, and that's a huge opportunity for us to continue to grow. We have secured a number of wins already. We have begun shipping to a number of customers already.
But a lot of these wins that we've been amassing over the last couple of years will start coming to market over the next year or two, and as those scale, you know, obviously, that creates a lot more stickiness, a lot more visibility into those wins. And, you know, as it makes sense, you know, we'll look to continue to provide more color.
But beyond the automotive, we've seen, you know, a continuing increasing penetration in industrial and commercial applications, in server, telecom, boot drives in point-of-sale terminals, all of these things, right? So I think, you know, certainly as the market continues to move towards a more connected world, SSDs, eMMC, UFS all have a place in this market to do well.
Similar question just about your, the Ferri embedded storage business or the SSD solutions business. Yeah, how is that going? You know, what might we look for in 2024 in terms of, you know, growth drivers?
Yep. Ferri has always been, you know, a strong business for us. It's been fairly consistently around 10% of our overall revenue. You know, we are seeing strong traction here with, as I said, automotive, but also with a lot of these other applications in industrial and commercial. You know, this is a business where, you know, we view it very complementary to what we already do on the eMMC, UFS, and SSD side.
This is not a business where, you know, we expect to compete against our customers. This is a market where, our customers in this area are typically smaller, smaller scale, and don't necessarily have either the size, scale, or ability to work directly with the flash maker.
So it would make sense for us to work with some of these guys, provide, you know, a full solution and service that market. So, you know, our focus is always gonna be on the controller side, but, you know, there are pockets of opportunity like, you know, things on Ferri that we're addressing that, you know, we see as a good long-term complement to what we do.
Perfect. I wanna ask a question. I believe you said, and I think this was on the call, that your three new products, I believe, are all 6 nanometer.
Yep.
What's, you know, the mainstream products today? What node are you manufacturing, and, you know, how quickly do those need to migrate to, you know, six or potentially lower process nodes?
Yeah, I think most of our stuff, most of the higher vol- higher volume stuff are, you know, that 12-20 nano. Some of the more legacy stuff are certainly higher process geometry than that. You know, we don't typically respin a product at a lower process node just because, you know, whether it's a new generation nano we're trying to support or more cost effectiveness, there are a number of factors that go into it.
But what's important for the Gen 5 PCIe, for UFS 4, was the, the combination of that much higher performance required in these next-generation interfaces while maintaining the same power envelope, right? Because, you know, PCs still wanna be thin and light. Phones, you still expect to have a day's worth of use out of it, right?
You can't, you can't squeeze that doubling the performance or, you know, significantly improving the performance and maintaining the same power envelope without shrinkage, and that's the only way, way, way we get there.
What are you seeing on the foundry wafer pricing side? I know 2022, 2023, we saw some price increases, especially from, I think, your, your large foundry customer. Do you expect that to stabilize in, you know, 2024? Is it up? Is it down? What, what are your thoughts around it?
Yeah, we'll provide more color around that, but I think, you know, certainly, I think wafer... you know, the foundry partners are still, it's still tough working with them, right? I mean, they have a lot of demand, and we understand that. And so we'll, but we are also a large customer there. We've been a large customer of theirs for a very long time, and so there's benefits of scale, benefits of that long-term relationship, and we'll continue to leverage that.
Where we see, you know, we believe we can extract some cost benefits more on the back-end packaging and testing, which is what we said before. But, you know, I think there's, you know... we'll continue to work with our manufacturing partners to help drive better cost for ourselves.
Got it. Can you remind us, because you may have been restricted in what you could say in 2022 and 2023, as wafer pricing increased, a lot of semiconductor manufacturers passed along those price...
Yeah
Input price increases to customers. Obviously, we know what happened with SIMO's margins, so it looks like maybe you didn't pass along as much. Maybe you did, but what was your policy back in 2022 and 2023 in terms of passing along input price increases?
Again, I think we, we've had a good relationship, and we've been able to maintain that. You know, our margins stayed pretty healthy, even as we absorb some of those, those higher costs.
Okay. I wanted to ask, 'cause we've gotten this from some of our other companies. I believe TSMC is your primary foundry, but, but SMIC is also a foundry. Are you starting to see any pressure from companies to move manufacturing first out of China, but more, you know, interestingly, are you seeing any pressure to move off the island of Taiwan?
No.
Okay. And then lastly, you know, margins, as we discussed, is kind of currently in the low to mid-40s, 43, 40, 43.5, kind of, or slightly above. Your target's sort of in the 50, you know, ish range. What are the biggest puts and takes? You talked about being able to raise pricing with new products. You know, what's a reasonable timeline for investors to think about, you know, returning to a gross margin around 50%?
Yeah, you know, our normalized gross margin is in that 48% to 50%, and I think, you know, looking at 2025, sometime in 2025, is probably a reasonable expectation.
Got it. Okay, well, that ends my questions. If anyone in the audience has questions, I think we've got a couple of minutes that we can take them. Well, all right. Great. Jason, thank you very much.
Thank you, Quinn.