Great. Good afternoon, everybody. My name is John Vinh, and I cover SMEs here at KeyBanc Capital Markets. We're pleased to have Silicon Motion here, Jason Tsai, CFO, welcome, Jason.
Thanks, John. Appreciate the opportunity. Just to kick things off, we'll be making forward-looking statements, so please refer to our filings with the SEC on risk and uncertainties involved in investing in our securities.
Great. Maybe just to set the stage, Jason, maybe just give us a brief overview on SIMO and your major products and markets. What are you most excited about?
Yeah. As some of you may know, I first started at SIMO 17 years ago, and I've left and come back since. Over the last 17 years, we've grown incredibly well from cards and flash drives initially to eMMC and UFS for smartphones, SSDs for PCs, and now going into enterprise SSDs with our new MonTitan platform. Along the way, we've expanded into automotive. We've expanded into IoT and connected devices. We've got a very broad range of portfolio solutions across really any device out there that requires embedded memory, solid state memory, and that's pretty much everything in our pockets, everything we carry around, and everything that drives our everyday lives. Over the past year or two years, what we've seen is a really major shift from the flash makers to outsource more.
What we're seeing is certainly, especially with AI coming or here, the focus from memory providers is to balance their investments across a wide range of technologies, including NAND, DRAM, HBM, etc. It becomes increasingly difficult for folks to be able to manage those investments across the board. They're relying more on merchant suppliers. One of the things that we've done a good job building over the last many years is developing arguably the broadest and the deepest relationships with every single flash maker there is. I don't think any other module maker can say that. While historically we have had projects with all flash makers, I think what we're seeing now is more of a categorical change in their strategy to incorporate merchant suppliers as part of their long-term strategy. That's where we're starting to benefit tremendously.
We expect that cycle to be driving significant growth for our business longer term as we cycle in new products. I guess a long-winded answer to your question, John, is we're excited by the core business through share gains and increasing opportunities. We're also excited by new markets like enterprise that we're starting to see a tremendous amount of traction with, especially with our QLC capabilities to manage QLC NAND.
Great. The memory market is pretty, pretty dynamic. You saw Micron positively pre-announce this morning. How does this benefit SIMO?
Yeah, I mean, at the end of the day, we're all looking for more density, better performance at a lower cost, right? Whether you're looking at DRAM or you're looking at NAND, with NAND, what we're seeing is certainly that continuing progress forward towards more layer count, lower dollar cost per bit. The trade with the devil here is that as you add more layers, as you go from, you know, three bit per cell to four bit per cell, etc., the endurance, the reliability, performance of NAND becomes weaker and weaker. You need a new generation of controllers to support that new generation of NAND. That's where the opportunity for us really opens up. As the end markets for NAND use become broader and broader, flash makers, everybody needs a new portfolio of controllers to manage that new generation of NAND.
As I said earlier, the incumbents, the internal solutions, the internal development platforms become a lot more taxed. They look towards additional opportunities, additional capabilities to help supplement and augment what they do internally. That's where the opportunity for us is.
Great. You talked about gaining share in each of your end markets, but much of your end markets are being supplied by the NAND makers themselves. How are you able to gain share from them?
Yeah, it's a good question. I think historically that's always been, I think, a challenge that investors have had with us: how do we sustainably gain share? I think what we've been able to demonstrate here, especially with the latest generation PC SSD controllers, our PCIe Gen5 eight-channel controller, our solution is hands down the best one in the market. Whether it's an internally developed solution or externally developed merchant supplier, there's nobody else out there who has developed a more powerful, more power-efficient controller than we have. You go to any of the third-party review sites, all the best controllers, all the best SSDs are using our controller. Some of our customers have built their own custom firmware that allows them to differentiate a little bit more.
Some of our customers are using turnkey solutions that we provide, and that enables them to bring to market very quickly and very cost-effectively, very high-performance, one of the highest performing SSDs in the market. Now we're able to demonstrate, I think, I mean, we were confident about this in the past, but categorically demonstrate that outsourcing is not an inferior solution. If you compare what we've been able to deliver versus the two flash makers who have developed their own internal solutions, our solution, or SSDs based on our solution, perform better at a much, much lower power consumption level. You're looking at sub 7 W of power consumption versus 9, 10, 11 W of power consumption with some of these other solutions. Lower power consumption isn't just a PowerPoint slide, it's actual savings for bill of materials. You have fewer passive components in a PC to dissipate heat.
You have better battery life. You can potentially have a smaller battery. All of these things are real tangible cost improvements because of that power savings. Being able to deliver these types of solutions that are not just enabling them to get into market, but also enabling our customers to lead the market with a third-party solution has really changed the narrative with the flash makers. Given my long history here, I would say our pipeline of design wins and opportunities that we have today have far exceeded anything we've seen in the past. We're really confident and excited about where we are today and continuing to drive share. I think that you'll see whether it's in client SSDs, whether it's in smartphones across the board, the opportunities for us are growing pretty meaningfully.
Okay. You mentioned client SSD controllers. You are the market leader. How do you grow this business when the PC market really isn't growing anymore?
Yeah. You know, these PCIe Gen5 SSDs, I think if you take a look at our SSD market share, you know, historically last year we were about 30%. We believe we can get to 40% over the next few years. The reason for the confidence around that is this latest generation of PCIe Gen5. We've run with the eight-channel high-end controller. We've won four flash makers and virtually every module maker. On the four-channel mainstream one that'll come to market late this year, we've also won four flash makers and virtually every module maker. We're confident that with the PCIe Gen5 platform, we're going to be able to win more than 50% share of the overall market. You balance that, you average that out with the 30% that we've had in existing markets, and you can get to that 40% share over the next few years pretty easily.
I think the challenge is also that as you get to higher performance with PCIe Gen5, with a tighter power envelope, you have no choice but to go to the smaller power process geometries. Now these controllers that we're building are built on 6 nm. 6 nm is expensive. I mean, historically, most of our products were at 12, 20, 28, or higher nm. At 6 nm, it's $15 million- $20 million per controller to develop. There aren't a whole lot of merchant suppliers out there that have the balance sheet and P&L that we have to support development at 6 nm, not just for one product, but for a portfolio of products that we're bringing to market.
That's been one of the things that I think has also set us apart and has also increased the need for outsourcing because whether we've spent that $15 million- $20 million or somebody else spends it, somebody's got to have to build out that controller. Might as well leverage the economies of scale that we've already developed. Going forward, you know, gaining more share, we generate alpha in our business, alpha growth. We're also seeing higher ASPs for PCIe Gen5. On average, on mainstream, you're looking at 40%- 50% higher ASPs. On the high end, almost double the ASPs compared to a high-end four-channel, excuse me, PCIe Gen4 controller. There's opportunities to gain additional growth vectors through market share gains and higher ASPs. That's what we believe will continue to drive our business over the next few years as we continue to drive stronger business here.
Maybe a similar question on mobile controllers. You know, what's going on with eMMC and UFS? You know, this market's also ex-growth. How do you grow there?
Yeah, it's a good question. I mean, look, smartphones, it's like PCs, right? I think best case scenario, we're looking at an industry that grows low single digits for the next many, many years. EMMC historically served the low-end segments of the smartphone market. Now that's getting displaced by UFS, but you're seeing other applications in EMMC becoming more relevant. Smart glasses, AI goggles, all these things, IoT connected devices, even automotive. We're taping out a new EMMC controller for a large South Korean flash maker for EMMC for the automotive market. Even though the smartphone market transitioned to UFS seven years ago, EMMC is still an incredibly important part of the market and will continue to be for a very long time to support all these additional opportunities. We're gaining share with these customers as they continue to scale.
On the smartphone side, historically what you've seen is the memory makers that have both DRAM and NAND typically own a large chunk of this market because they can combine the two into a multi-chip package, so EMCP or UMCP. Now that you're starting to see readily available cheap mobile DRAM coming to market, the value proposition of having a combined solution is no longer there. It's more cost-effective to have a discrete mobile DRAM and discrete EMMC or UFS. That opens up a significant opportunity with a lot of the module makers that we've historically had great relationships with to be part of that supply chain, to be buying raw NAND from the flash makers, pairing it with our controller, and selling it to a handset OEM. We're seeing that becoming a big driver to share gains for us.
Oftentimes in mainstream and even low-end handsets, flash makers don't necessarily make a whole lot of business when they sell EMMC and UFS into those markets. They would oftentimes prefer to sell just raw NAND, and they can probably make a little bit more money just on the raw NAND itself than trying to develop a packaged solution for those markets. That works, again, aligns well with the ability to gain share as they outsource, as they look to turn towards module makers more, and we gain share there. Historically, we've been about a 25% share here. We believe we can scale that to 30%+ over the next few years as well. There's still an opportunity for us to gain share longer term.
Okay. You mentioned automotive. What's happening in this segment, and you know, how big could this be for you?
Yeah, I mean, look, you know, we all know that the automotive market itself, the number of cars that are getting built each year isn't really growing by leaps and bounds, right? It's a difficult market. The opportunity that we see is that the number of storage devices inside a car is growing very rapidly. To support the increasing number of cameras, sensors, ADAS, emergency, you know, capabilities within a car, infotainment, battery electric vehicles, power management, all these things require storage ranging from eMMC and UFS to PCIe 3, 4, 5. The opportunities here continue to grow. We've been fostering, nurturing this business for the last many years. It's already 5% of our business. We believe we can scale it to 10%+ over the next couple of years.
As you know, with a lot of these automotive wins, a win today doesn't really materialize to revenue for three years or more. The growth, our ability to grow this business over the next few years is really the confidence comes from a lot of wins that we already have, that we are starting to see getting close to that point where they're going to start scaling soon. We're supporting virtually every large major automaker in the U.S., in Europe, in China, whether it's internal combustion, whether it's battery electric vehicles, whether it's autonomous, all of them, where we have relationships, whether direct or indirect, with them already.
As these vehicles become, as our share within these vehicles continues to grow, we believe we can scale this business pretty significantly over the next few years from, again, 5% of our revenue to, you know, more than 10% over the next couple of years.
Right. Maybe we can shift over and talk about the enterprise segment. How is this coming along and why now? What's the differentiating factor that kind of allowed you to break into this segment of the market?
Yeah, you know, enterprise has always been kind of the holy grail for us, right? Something that we've been trying for a while. I think right now is a unique point in time where we're starting to see affordability of NAND continue to come down quickly, especially with the introduction of two TB QLC dies, being able to support enterprise SSDs upwards of 128 TB in density. The reason why this is important, the reason why this is really kind of the crossover point for SSDs is that at 128 TB SSDs, you're getting to a TCO that becomes fairly comparable to hard disk drives. In the world of AI, where these LLMs, whether it's for training or inference, just really want to consume more data, input more data more quickly.
The way the enterprise storage infrastructure today is, you've got a lot of high-performance compute SSDs in there already, but those are expensive. Dollar per bit is still very expensive, and they're used for certain applications that that higher dollar cost per bit makes sense. When you look at more density, you're looking at hard disk drives, and they can store a lot of density cheaply, and that's great, but it doesn't enable the AI algorithms to be able to pull that data, process that data quickly. You can use very expensive compute SSDs, or you can use lower performance hard disk drives. Now with QLC, now you're able to bridge this gap with a high performance, high capacity solution where QLC NAND is almost tailor-made for these types of use cases. You're able to store a lot of data cheaply.
You're able to pull that, read that data very quickly, and you're able to do it in a manner, in a use case where you don't erode the endurance and reliability of that NAND because it's QLC. Because you're doing a lot more sequential read writes versus a lot of random read writes, you start developing solutions that are much better endurance. As AI infrastructures continue to get built out, as flash makers continue to move to next generation NAND, that will start driving higher, more production, more availability of the 2 TB dies and 4 TB dies and so on and so forth longer term. You start getting to that crossover point of SSDs becoming more cost-effective or as cost-effective and eventually more cost-effective.
Because the world is generating more data each day than it did the day before, there's still a tremendous opportunity for hard disk drives as it moves further down into more cold storage. The amount of data the world is creating will continue to drive additional use cases, additional need for more storage across the board. I think it presents a unique opportunity for us. Given our understanding and the fact that we've managed more QLC than anybody else out there puts us in a really unique opportunity to deliver a really differentiated solution for our customers. We've talked about winning two Tier 1 customers and four additional customers and seeing this beginning to scale over the next 12, 18 months. As we exit 2026 into 2027, gain that 5%- 10% of revenue. This is an opportunity where we're starting to see a lot of momentum, a lot of traction around that. We're really excited for it.
Just to follow up on enterprise, what are your revenue and margin targets, and how are you guys tracking towards those targets?
Yeah, so you know, the first two Tier 1 wins that we've talked about, plus the four additional wins, gives us a lot of confidence in getting to that 5%- 10% of revenue contribution by the time we're in the kind of 2026, 2027 timeframe. We've said that these Tier 1 customers, one of them could get us to 5%- 10% pretty quickly. Certainly, having more gives us a lot more confidence around getting to that number and scaling beyond that, right? 5%- 10% is the first waypoint. It's not a final destination. That's not the most we believe this business can contribute. This is kind of the first waypoint for us for investors to gauge our progress.
At FMS last week, Future of Memory Storage conference last week, we demoed a product enterprise solution using on the VAST network, on a VAST platform with a Unigen SSD using our controller to showcase the high-performance capability of a high-density 128 TB QLC SSD. It was very well received by the press and potential customers, etc. . We believe that we can continue to see increasing opportunity share. The two Tier 1 customers that we won, one of them is in the U.S., one of them in China. The way we classify Tier 1 is either you're a flash maker, a storage solution provider, or a Tier 1 enterprise or CSP. We've won one in the U.S., one in China, and along those lines. The four additional ones will give us, they have significant reach in existing business with a number of Tier 1, Tier 2 enterprises and CSPs globally.
We're very excited about the way this scales from a margin perspective. As you would expect, enterprise SSDs are margin accretive compared to our overall business. We haven't provided direct guidance on what that looks like, but certainly margin accretive relative to the overall gross margins of the business. We expect as that continues to scale, that'll have a positive uplift to our overall corporate gross margins long term.
Speaking of gross margins, you know, you talked about getting back to the historical range of 48%- 50%. Looks like you're going to be at the lower end of that this quarter. Your long-term operating margin targets are 20%- 25%. Seems like you're a little bit further away on that. How do you get there?
Yeah, over the last year and a half, as you guys may recall, a year and a half ago when the world was kind of in more difficult places, the NAND makers were negative gross margins. We dropped our pricing a lot in order to support our customers. Over the last year and a half, we've steadily improved our gross margins. Now this quarter, third quarter, we're guiding to 48%- 49% gross margins, back to that historical range. Exiting this year, we anticipate being in the upper end of that 48%- 50% range as we exit this year. On an operating margin basis, one of the things that we've been doing over the last two years is we've been investing a lot, right? We've taped out three 6 nm products, each at, let's call it $15 million- $20 million each.
We've been investing in MonTitan, our enterprise platform. These have all been high dollar cost investments on the OpEx line that we have just really begun seeing the first early revenue contributions, first return on investments that we've made, especially on the first product being the eight-channel PCIe Gen5 controller. As we start seeing the other products begin to generate ROI, as we start seeing enterprise begin to scale, we expect to see that really benefiting our operating margin line. We will continue to invest. We will continue to build out advanced geometry products. We're slated to tape out a 4 nm product, PCIe Gen6 controller next year for our enterprise customers as they transition, as the industry transitions to PCIe Gen6 in the 2027, 2028 timeframe. We want to make sure that we're there for that as well.
We will continue to invest, but I think over the next 12- 18 months, we'll start seeing a lot of those investments we've made over the last two years really coming back to helping our return on investments. Over time, we do expect to be able to get back to our historical operating margin range. We're looking at revenue to scale in order to get there, but we're confident that we can.
Okay. You guys have a lot of cash. Maybe just talk about use of cash going forward.
Yeah. The way we think about capital allocation is it's a three-prong strategy for us. We've consistently paid a dividend for the last, you know, better part of 10 years. Started off at $0.60 a year. Now we're up to $2 a year. The only time we didn't pay was when we were in the middle of the acquisition process. The second part of our capital allocation strategy is share buyback. While we do not have a consistent share buyback strategy, we have consistently bought back shares periodically. If you take a look at the last three to five years, we've actually spent about half of our free cash flow on share buybacks. Thirdly is acquisitions. We don't tend to do a lot of acquisitions, but we tend to always want to look and see what's out there.
Those are the ways we think about capital allocation and use of cash. Certainly, depending on kind of what we're doing, there'll be different emphasis, but the expectation is that, you know, the dividend will be something that you'll continue to expect from us. You know, we'll continue to layer on other uses of cash as appropriate.
Right. Looks like we're out of time. Thanks, Jason.
Thank you.