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

We'll be making some forward-looking statements during this presentation. So please refer to our filings with the SEC on the risk and uncertainties involved in investing in our securities.

Okay. Great. Yes, I'm Simon Woo. I cover memory chip industry, and I'm going to moderate this session. Actually, I have already received some investors' questions. Thank you very much, Jason.

So number one, what's the Silicon Motion's specific long-term growth catalyst?

Sure. Yes, look, I think it's increasingly important in this industry to -- for memory. NAND flash memory is becoming critical for a wide variety of applications. Smartphones and PCs being the main opportunity right now, but certainly seeing more interest in enterprise, in automotive, industrial, commercial, all of these applications are increasing the amount of memory required and the performance requirements for that memory. As each of these end markets, as the devices in each of these end markets become more complex, the need for memory increases. What we're seeing is as memory makers, expand and improve upon generation after generation of NAND, that creates opportunities for us to deliver additional solutions that help drive better performance, higher density and overall expands our TAM and SAM.

Yes. That's the great long-term trend. How about the second quarter, usually the low season, but any quick recap or update near-term second quarter outlook?

Yes. Look, I think, obviously, the world is -- there's a lot of uncertainty around geopolitics, around tariffs, and it's certainly challenging to forecast around that. However, we are still confident that this will be a growth year for us. We expect to be growing in the second quarter sequentially as new projects begin to ramp, as opportunities begin to scale. We've got a number of new projects that are coming online this quarter, including PCIe 5 for client SSDs on the high end. We've got certainly continuing new customers, new products increasing over the next few months that we anticipate will be -- that will drive significant growth for us later this year.

Yes. The follow-up question is any sort of the pull-in demand from your customers for second quarter?

The conversations that we've been having with our customers would indicate that the demand that they're seeing in the second quarter is real demand and not pull-in demand. Certainly, there is a concern around uncertainty in the back half of the year and whether that would present itself in pull-in demand here in the second quarter. But from our conversations with our customers, our belief is that the demand that we're seeing in the market today is real.

Yes. And then the -- looking at the second half, usually a good season, but lots of uncertain factors there because of some China tariff or U.S. tariff. So any rough idea, second half outlook? And then the management's dream number, $1 billion run rate, could be achieved...

No, I wouldn't classify as a dream number. I think it's something that certainly, we've got a lot of visibility towards and confidence towards. We have a number of new projects that are ramping in the back half of the year that we expect to be additive and help drive that -- our business there. Obviously, to your point, there are things that we can't control. We can't control tariffs. We can't control geopolitics. But what we can control is certainly our ability to win more sockets to expand our customer base, expand our project base. And I would say that in the 15-plus years that I've been with Silicon Motion on and off, we haven't seen a backlog as strong as we have in terms of the breadth of projects, the breadth of customers, the engagements with our flash maker partners, we haven't seen an opportunity this strong in a very long time, if ever. And so what we're seeing is a really broad-based increase in outsourcing by every flash maker across the board, and we are gaining significant share. On the client SSD side, last year, we went from 25% plus to 30% plus. We believe we're -- we can achieve 40% market share over the next few years on client SSDs. You take a look at on the high-end PCIe 5 controllers that we just launched late last year, we believe we'll be able to achieve 50% plus market share in this category. And from a market historically, we haven't played in because the flash makers typically develop those controllers themselves. And they have now outsourced for -- those flash makers have outsourced their entirety of their high-end PCIe 5 client SSD controller business to us. So that's been an opportunity where we see certainly the first of many opportunities beginning to open up.

Yes. Maybe let me double check the number. You are talking about 50% market share for the client SSD?

For the high end, for the high-end segment of the market where historically we haven't played in. So this is incremental opportunities for us to gain share here that historically, flash makers have provided the solution in-house with their in-house controller that now are outsourcing.

Yes, that's a great thing. But this 50% market share mostly based on the PCIe 10?

5. Correct.

Great. Maybe we can continue this discussion because some investors are a little bit confused, the conflict of interest with the NAND flash chip makers. So because memory chip makers, they have in-house controller, Silicon Motion also sells your controller to the chip makers. Would you recap how the business model works for memory chip makers?

Sure. That's a good question. I mean, look, certainly, the challenge historically has been around how can they -- how do we reliably gain share over time if the flash makers have their own solution. What we're seeing now is certainly a shifting dynamic in the market. As more -- as NAND evolves as more 3D NAND and 3D NAND layers complexity increases, it requires new generation of controllers. We talk about the latest generation of controllers for the latest generation interface like PCIe 5, like UFS 4, we're moving down to 6-nano process geometry, which is incredibly expensive. For every single controller that we're developing at 6-nano, it's $15 million to $20 million in development cost for a single controller. So you're now at a crossroads for the flash makers where you're seeing increasing reliance on outsourcing because the cost of development has grown -- has increased. The cost of -- the range of portfolio solutions that you need to have to serve all the different end markets is continued to increase. And you're starting to balance, you need -- the need to balance between NAND development as well as DRAM and HBM development is creating certainly challenges for resource allocation. And so all of these things put together presents an opportunity for us to continue to gain share. We're the only merchant supplier out there that works with every single flash maker. And with that, we're able to win a disproportionate share of the market in terms of gaining share. And that's what we continue to be able to deliver on, and that's what gives us confidence about the near term and the long term.

Yes, yes. How about the overall competitive landscape with other pure NAND chip controller suppliers, the Taiwan guys or U.S. players?

From the merchant suppliers? Merchant controllers?

Yes, yes.

Yes, so I think in my -- again, when I first started here 15-plus years ago, there were probably half a dozen, 10 different other merchant controller suppliers for NAND memory. I think over the last 15 years, that number has come way down where there's really only 1 or 2. And even with those guys, they're a bit conflicted. There's a bit of more of a conflict of interest. We're really the only leading merchant controller supplier that isn't conflicted. We don't sell -- we don't compete with our customers. We are a pure merchant supplier for the most part, and we are able to work with every flash maker and not compete with them in a lot of the areas that our competitors would.

Yes. Yes. Maybe we can talk about the mobile solution. So any quick update to UFS-driven top line growth versus maybe eMMC type of the demand downside?

Sure, eMMC as part of the smartphone market continues to contract, getting replaced by UFS 2.2 and 3.1. And then UFS 4 historically has been more on the high end and now beginning to migrate into more of the mainstream segment of the market. We developed our UFS 4 controller last year, and that we expect that to start scaling late this year and into 2026. Our solution is a 2-channel controller versus all the other internal solutions out there are a 4-channel controller. And the advantage of a 2-channel controller is that it's a lower cost solution. And it's -- from a performance standpoint, it's pretty comparable to what they developed many years ago. And so we're able to help drive UFS 4 adoption from the high end into more mainstream, utilizing newer generation NAND that's more cost effective, utilizing a 2-channel controller that's more cost effective. So in essence, being able to improve the cost while maintaining the same performance in order to drive adoption into more mainstream handsets. We're working with a number of customers, including flash makers and module makers that are targeting the smartphone market that we expect to see meaningful growth starting to scale next year as the baseband folks start driving -- incorporating UFS 4 capabilities into their mainstream baseband products.

So currently, the UFS already the mainstream mobile...

It is, yes. So you've got UFS 2.2 and 3.1 that's primarily on the mainstream to lower end segments. You've got eMMC that's at the very kind of utilized in the very basic smartphone category and then UFS 4 in the premium and flagship category today.

But a little bit -- we can talk about more details of China or geopolitical risk, but Android phones mostly manufactured by Chinese OEMs and then, of course, Korean company, Samsung, but what's your current business, maybe risk opportunities with the Chinese smartphone makers?

Sure. Look, I think if you take a look at the top 10 smartphone OEMs in the world, right, #1 and #2 is Apple and Samsung. And we have virtually none to -- virtually none, no business with them. We do have a few sockets within Samsung, but it's fairly small. But I think overall -- and certainly, Apple uses their own internal solution, but we support all the other 8 players in the top 10. And none of those other folks really sell a whole lot into the U.S. So we don't see a significant risk in terms of tariffs, et cetera, that would impact our smartphone-centric business. Most of the stuff that we're selling to them are fairly trailing edge and so are not subject to any sort of issues that are present in the market today.

Yes. You mentioned that previously auto industry offers also the new growth opportunity. So what's the management strategies for the auto area?

Sure. Automotive is an opportunity where we see significant growth longer term. It's already more than 5% of our revenue. We believe it will scale to 10% plus of our revenue in the next couple of years. And yes, certainly, design wins today will lead to revenue in 2, 3 years' time. And that tends to be very sticky. That tends to have a lot of visibility towards that kind of long-term growth trajectory. Automotive, while certainly the automotive market itself isn't growing, the number of storage devices in a car is certainly growing very quickly. From cameras to sensors to LiDAR to ADAS, infotainment, central processing, all these things require memory, typically a discrete piece of memory. And so that increases the number of memory components that are going into a car, where historically, you had one for the infotainment, but now you're getting upwards of 10 or more devices depending on the complexity of the car that it's going into. So it creates an interesting opportunity for us. And it's not about just the very high end. It's -- we've got opportunities with Samsung for eMMC going into automotive. We've got a number of engagements for UFS going into automotive. SATA, PCIe 4, PCIe 5 wins that we're working on for long-term growth. So there is a number of opportunities here that we're really excited about that I think as cars become increasingly more complex and the capabilities become higher with ADAS and autonomous vehicles, the opportunity for us will continue to grow nicely.

Yes. Here in San Fran, we do see a lot of autonomous driving cars. So any kind of the Silicon Motion specific solutions or controller for autonomous driving cars?

Yes. So we've talked about -- we've named off a number of customers, including Waymo, Tesla, GM, Mercedes-Benz, Toyota, Honda, again, a wide range of cars that we are already supporting with our -- whether it's controllers or SSD solutions, but providing memory capabilities to a lot of these cars. And so we're excited by all of this. And I think certainly, with Waymo being the leader, at least here in the U.S. for autonomous vehicles and everybody trying to figure out kind of how to catch up to that I think the opportunity continues to grow very nicely for us long term.

Yes. So this one, for example, Waymo here in San Fran getting more popular, lots of people are using this. But would you specify Silicon Motion's, for example, controller only for Waymo or you purchase NAND chip from the chip makers and then you provide a solution to each?

We haven't commented specifically on the solution for each automaker. And certainly, each automaker, we could have one solution, we could have multiple solutions. So it does vary quite a bit, but we haven't specifically called out what we're providing for each customer.

Yes. And then the -- yes, let's talk about the data center, AI server requires some extra maybe NAND storage. So would you recap all the boot drive, et cetera?

Yes. So our enterprise strategy is twofold, right? We talked about on this last earnings call, we talked about a new opportunity with the BlueField-3 DPUs from NVIDIA, where we're providing the boot drive for that. That's a great opportunity. It gives us certainly a foot in the door into the NVIDIA ecosystem, having a better understanding of the enterprise SSD requirements and the opportunities there. But where we see a bigger opportunity is what we've been talking about historically, which is our MonTitan enterprise SSDs, right? This is a TLC, QLC controller supporting densities upwards of 128 terabytes, where we are starting to -- we won 2 Tier 1s and 4 additional customers last year, and we expect those customers to continue -- begin to scale throughout this year and then generate 5% to 10% of our revenue by the '26-'27 time frame. And we're on track for that. We anticipate that with the increasing focus on AI, delivering solutions that combine both high performance and high density at a TCO that's very compelling and very comparable to hard disk drives creates a new opportunity, a new window of growth, a greenfield opportunity for us longer term. We have more experience around managing QLC than anybody else out there. We've managed more QLC than anybody, flash makers included. And so being able to bring to bear that level of experience to this product category, I think, really sets us apart and creates a pretty meaningful moat around our capabilities.

Yes, yes. Maybe a little bit more details. For example, when we look at the AI server, so we can see first GPU, graphic processor unit and HBM, high-bandwidth memory, okay. So one combo. And then we can see the CPU, for example, Grace CPU and low-power DDR5, that's the combo. And then we can see the storage called eSSD, whether you are maybe MonTitan or a chip makers' controller.

Yes.

But the question is, would you recap your new solution, the boot drive, how it works and within the AI server with the GPU, CPU, a lot of chips are there?

So the BlueField DPU is attached to the Grace Blackwell configuration. And it helps offload some of the processing onto the DPU, the data processing unit, and that DPU connects to up to 15 enterprise SSDs. And so the opportunity that we've talked about here, the win that we talked about here in the last earnings call is around the boot drive for that BlueField-3 DPU. And we expect to start seeing revenue late this year and that's scaling more meaningfully into '26 and beyond.

It is possible to expect a multiple, more than one customers rather than one single NVIDIA dominant demand?

So we already provide boot drives. We've been providing boot drives, for example, for Google servers for a very long time, right? For the last many years. So there are multiple opportunities here, but certainly, obviously, I think in this day and age, being part of the NVIDIA ecosystem is very beneficial.

Yes. And then maybe we can continue the MonTitan discussion. When do you expect any meaningful revenue contribution in MonTitan?

Yes. So we started seeing early revenue late last year. We'll see -- continue to see modest revenue throughout this year and then scaling more meaningfully into '26. As you said, scaling to that 5% to 10% of our total revenue as we exit '26 into '27.

So towards the end of 2026 or sometime 2027, 5% or 10% of your revenue...

Would come from MonTitan.

So good, yes. All right. We have 5 or 6 minutes left. Maybe let's talk about the overall macro or industry uncertainties. So any impact, U.S. tariff, so far or you don't feel any big, big impact?

I think certainly, the uncertainty around tariffs and the geopolitics does create some consternation amongst our customers and end customers, et cetera. We haven't -- if you take a look at the overall expectations for PCs and smartphones this year, it remains relatively unchanged. The industry is still expecting growth of low single digits for both PCs and smartphones, and we haven't seen that really change much since the beginning of this year, even with the implementation of tariffs and kind of all the volatility around that. I think certainly, the hope is that we'll see the rest of the year play out. But what we do focus on is what we can control, right? And what we can control is our ability to win new sockets, our ability to scale with new customers and existing customers on additional sockets. And that's really what's been important for us. And so we continue to focus on winning more share with our flash makers, ramping up new projects with our customers. Those are the things that we can control. And that's what gives us confidence about how we're going to scale in the back half of the year and how we're going to continue to scale longer term that's somewhat independent of the macro. Certainly, if the end markets become weaker, while that would create pressure for everybody in the industry, given that the increasing amount of wins that we have, should help bolster or help buffer some of our performance.

Yes. Maybe -- just before the U.S. tariff implementation, do you see a little bit higher-than-normal level of channel inventories these days? Or it looks broadly normal?

Inventory levels look broadly normal. I think, obviously, we work very closely with our customers to understand their demand, their end market demands and the amount of products that we're shipping to them largely reflects end market demand. So we believe that inventory in the channel remains fairly stable.

Yes. And then maybe we can talk about Singapore. So any update or comment regarding the litigation thing?

No. It's -- the litigation is ongoing. The arbitration is ongoing. We expect Singapore arbitration to convene in October this year. And we expect a resolution decision by the end of this year. So nothing has changed. There's no updates to that. Obviously, legal updates are -- it's a confidential process, so it's going to be difficult for us to provide kind of the day-to-day updates on the process. But overall, we're still on -- the Singapore arbitration is still on track for October.

So Singapore arbitration, their review will start in October?

Correct.

And then they will -- they should make the decision by the end of this year?

Correct.

Okay. And then after that, so no -- any particular guidance?

Yes, we're not commenting on it.

Yes. All right. We discussed the mostly business and then the almost -- also a very important topic is shareholder returns. So would you recap your balance sheet status and then how you want to deploy continuously the shareholder return?

Yes. I think if you -- our share -- our capital allocation policy has been pretty consistent over the last many years. The priority for us has been over the last 10 years is the dividend. We've been paying a dividend for the better part of 10 years. We started off at $0.60 per ADS. Now we're up to $2 per ADS, and we've increased that over time gradually. The second part of our capital allocation strategy is share repurchase. We have share repurchase programs periodically throughout our history. And if you take a look at, for example, over the last 5 years, roughly half of our free cash flow has gone to share repurchase. And the third part of it is M&A. We always look at potential to find things that are additive. I think certainly going forward with technology becoming more complex, performance requirements increasing, there's going to be more synergies between the controller and the interconnects. And so we're continuing to look at how do we continue to develop solutions that are leading edge and that can take advantage of all the processing requirements, processing capabilities that are available.

Yes, yes. And then the -- these days, there are many talented engineers, Taiwan, U.S. So what's the overall of your R&D activities with very professional engineers for AI [indiscernible] growth opportunities?

Yes, we've been increasing our headcount pretty meaningfully over the last 1.5 years, partially to drive wins and development on the MonTitan side, partially to develop new solutions at 6-nanometer for PCIe 5 and UFS 4. So there's a number of new projects and opportunities that we continue to engage in that we'll continue to hire. But to your point, there's a lot of great engineers in Taiwan and in China, in the U.S., and we'll continue to find the best talent out there.

Great. Yes. We covered all the details, industry trends and company specific. Again, if you guys have any questions, please feel free to contact me. Simon Woo at Bank of America Global Tech Research. And then thank you very much, Jason. This is a great session. We actually conclude this session. Thank you, guys.

Thank you, Simon.