Awesome. Good. All right, great. Well, welcome everybody. Welcome back, rather. I'm Marco Lagos. I am the Head of U.S. Semiconductor Investment Banking for Morgan Stanley. I am honored to have with me today Rajesh Vashist and Beth Howe, CEO and CFO of SiTime, with us today. Before we get started, the company's asked me to remind you that today's discussion includes forward-looking statements that involve risks, uncertainties, and assumptions, which are further described in SiTime's SEC filings, including SiTime's most recent Form 10-K, and that actual results could differ materially and adversely from those anticipated or implied. SiTime assumes no obligation and does not intend to update any such forward-looking statements. For more information, please visit SiTime's investor relations page at investor.sitime.com. All right, I'm glad I didn't memorize that.
Rajesh and Beth, thank you for being here. I guess let's start with sort of some of the basics. Just top level timing is often described as a small component, but it's a mission-critical one. If you'll permit me an analogy, and you can tell me if it's a terrible one, in data center. If data center is the concert hall, control and orchestration software are the sheet music, storage is the percussion section, power is the brass section, et cetera, et cetera, timing is the conductor.
Well, I love that.
Yeah. Good.
I should use that.
All right, there you go.
That's pretty good.
You don't even have... I won't even pay you a royalty. That's great. Can you explain why timing, performance, reliability, and programmability matter so much?
I mean, the whole thing with timing at the highest level is that it's been the least understood part of components. Frankly, there hasn't been a timing company. I remember when we went public in 2019, people wanted to know what is timing, because there wasn't one company that was dedicated to it. One of the things, one of the innovations, if you will, of SiTime right at the top has been that we declared that we will deliver all timing. That is hard to do, that's differentiated. We call it precision timing. That's a secular story. That's a secular growth story. SiTime does really think that we're the heartbeat of all electronics. In some cases, it doesn't really matter. My favorite example where it doesn't matter is a remote control.
We may get irritated with the remote control if it doesn't work one or two times, nothing happens. Nothing negative happens. In a satellite, it matters. In ADAS, it does. In data centers, it does. People who build phones are very obsessed with the quality of the timing. The programmability is a natural evolution of being semiconductors versus non-semiconductors. While our innovation is to be focused on timing exclusively as business, on technology, it is replacing non-silicon quartz, which is a passive material, with silicon MEMS, microelectronic mechanical systems, and that's the biggest technical innovation that just changes the game.
Fantastic. We'll get into that in a minute. Put a pin in it and just take a step back, talk about the big picture, right? You've gone through a lot of change since the IPO in 2019. From your perspective, what is the most important investors should understand today about the company, from where it was a few years ago? What's kind of been the biggest surprise to you in the evolution of the company?
Well, qualitatively, no surprises. We always knew timing was critical. We always knew that wherever there was great communications, great compute, and any functionality that was important, timing was gonna be important. What has taken me by surprise, of course, is I think everybody else, is the growth of automated driving, the growth of compute, see AI, the growth of embodied AI, see robots and such. That massive growth has sort of changed, has just changed everything. In general, we always knew that timing is diverse. T iming when I started at SiTime pre-revenue, it was a 30% company, and we were one of six startups, and there were three other larger companies at that time IDT, Silicon Labs, Maxim. It was a crowded field, and all the others bailed by 2011 or 2012.
At that time, the view that semiconductors need to replace non-semiconductors was very clear.
Mm-hmm.
We're the only ones who have successfully done it, so that's not a surprise. The adoption has been a surprise. Some of the use cases have been a surprise. The market has gone from $5 billion in 2005 to $11 billion, and a decade from now, it'll probably double again. SiTime is a small player in this business, but we are an important player in precision timing, which is high-performance timing under tough environmental conditions.
Mm-hmm.
Which is happening, of course, a lot.
Okay. Let's get into that a little bit, your core differentiation and competitive mode, right? From a technology standpoint, what aspects of SiTime's MEMS-based architecture are hardest for competitors to replicate? How does the mode strengthen as systems become more complex?
Right. We started with the first mode, which you're talking about the MEMS, but we also have significant differentiation in our analog because we're an analog semiconductor company. We have significant differentiation in the putting it together as a system because we take two chips and put them together and make a system out of it. After that, we have several others. Back to the MEMS piece, MEMS is hard to do, and those who get it done right stay in pole position. Look at Bosch in sensors, look at old HP, then Avago, now Broadcom in filters, look at TI in mirrors, and display. What the hard part of MEMS is not just the design, which is important, but also the fact that you have to build the process itself.
In other words, SiTime innovated and built its own process all these years ago, and it's portable, it's RIP, we carry it with us, and we've innovated on six generations with it. Finally, there are no tools to do MEMS design. There are no tools to do this. You can't go to a Synopsys, a Cadence to get these. SiTime, little known fact, rolls its own tools and has a team, a rather large team, of physicists, mathematicians, material science, and software people who build this environment. Think of a mini Ansys inside the organization. Then there's the analog, then there's the system piece. All of these translate to a lot of IP, but also a lot of trade secrets and knowledge.
Got it. A nd this is probably as much a financial as a management question, but how do you prioritize your time and where you put the investment in all these different parts of the development with the complexity of it? Like, how do you puzzle through that?
You wanna take a shot at it?
I think in terms of investing, it really is. For me, investing not only in ensuring that we've got a strong financial foundation, compliance and... but also being helping to grow the business. Thinking about our investment models, our ROI, how do we make trade-offs in terms of our R&D investments and working with our engineering teams to prioritize and to find capacity to invest in the different areas. I think as we will talk about, that diversity of applications gives us lots of areas to invest in. How do we think about platform investments and then derivatives to invest in taking that platform into specific application areas, whether it's AI or ADAS or industrial and robotics.
Thinking about that, and then also investing in our people, because as we're growing very quickly, one of the key things we have to do is really invest in our talent and our people and being able to enable and empower them, because at the end of the day, they're the ones that are, whether it's innovating or making financial decisions, and so that's really important as well.
That's terrific. I think everything starts at the top, both from a management standpoint, but also from a financial standpoint. As you think about pricing power and value capture rate customers increasingly care about the system-level performance rather than component-level cost. How do you think about value-based pricing, in timing devices?
For us, our North Star is differentiation, differentiation. Our North Star is providing a value that customers can't get. We go the opposite of value pricing. We go premium pricing. We basically tell the customer, "If you can find this product somewhere else, you should not buy from us.
Mm.
I know that my sales guys don't actually do that. They shiver when I say all that because they think that's really stupid to say, but that's what I want them to say. They don't really listen to me. That's what I tell them to do. The premium pricing is a gift that keeps on giving because what happens is, to those customers that do want the product, are willing to pay the premium pricing, the product is highly differentiated by definition, otherwise they wouldn't buy it. What that does is it gets their specs, their requirements, get our engineers super excited because they are building something that otherwise can't be built.
We give them challenges, this much phase noise, that much jitter, this size, that power, these functionalities, they get very excited about it, and that's the best thing we can do for an engineer, is to get them excited about building great product that nobody else can build. That circles back to a pole position back at the customer, premium pricing, more investment. It's how we filter our opportunities. Once we find a value proposition that's important, we don't give up on it. In other words, we started on Titan resonators, which are standalone resonators, about 15 years ago, and it was a huge challenge to build something which was 1/25th the size of a crystal, but had equal or better performance and behaved like it.
Taken us almost 15 years, at least six years of concentrated performance to do that, and we've reached the level of performance that quartz crystal is at, but smaller size, lower power, better performance in other ways. We're not done. We're headed to getting an order of magnitude better in performance in the coming years already. Once we find something like that, we just keep on going at it, and that's how we make our investment.
It's a bit of a virtuous cycle, right?
That's right.
That's great. Okay, back to sort of structural stuff, so tailwinds. What secular trends are most important to SiTime's long-term growth? How do those trends specifically translate into timing content growth?
Right. In general, we grow when more units are shipped, of course, whatever design in. We also grow because we think that the density of chips, of timing chips per use case increases. In other words, if there was 1 phone, 1 chip, timing chip that we had in a phone, five years ago, now when we have a phone design win, we have 2 chips in it. Density increases as the system gets richer, more complex. The ASP also goes up because the functionality increases. SiTime's ASP trend, we don't talk about it much because it doesn't help us talking about it. The trend has been significantly high over time.
Overall, wherever there's more AI, wherever there's more compute, wherever there's more communication, wherever there's secure communication, wherever there's need for small size, low power, SiTime's timing chips really come into play in precision timing. That's our secular tailwind. Without AI, without big things that are happening in the world, we still commit for the next several years to be a 25%-30% growth company. All of these other factors like AI or a phone or a car, a self-driving car, take it up. Last year we grew at 61%, the year before that we grew at 41%. We're beating our quote-unquote standard growth rate for the last few years, and hopefully that continues for some time.
Yeah, you've definitely inverted that growth curve a little bit the last few years, so that's great. This is gonna shock you, but I wanna talk about AI a little bit more. AI and high-performance data centers are extremely sensitive to synchronization, right? We talked about the conductor.
Yeah.
Latency, reliability matter. Can you walk us through why timing is critical in the data center environment specific and how your products enable higher performance and scalability?
Yeah. I mean, one stat that I got, I don't know if it's still true, maybe a little dated, is that the GPU is idle 30% of the time, and that is because it's not getting fed enough data at the right time. All we are trying to do is to increase the data throughput. When we talk about better jitter, better phase noise, better stability, we're basically saying that the rates are so high that the timing needs get to be more and more tight. Better timing enables better throughput, enables better latency, enables synchronization is another part of it. And that is kind of a bit of an insatiable requirement. 400 gig we had a play, not that much. 800 it went up in terms of needing SiTime.
At 1.6 T, we're talking about higher frequency, lower jitter, to the extent we can be smaller size. All of these things play, exactly along with us.
Mm-hmm
... in favor.
We talked about sort of content expansion as part of the story, right? As things evolve. With accelerators, networking equipment, memory architectures evolving, your dollar content per system, how is it changing over time, and where are you seeing the biggest step-ups as it relates to those?
Three generations ago, we were probably at $200 per rack, fully populated. Three generations later, we're probably $500-$700 per rack. The density is increasing, which means more timing gets to be more localized because if you're sending the timing signal too far in relatively noisy environments, it degrades as , and so it's not very useful. Well, just as people talk about vertical power delivery, there's conversations around vertical timing delivery because there isn't room to be sending it horizontally, so timing would land up being sort of on the other side of the board, as it were. co-packaging is one example of wanting to get into that. There's a lot of new accelerators that are coming on the horizon.
They aren't playing out yet, but I think SiTime gets to win no matter what. W e hear this ding-dong between, oh, is optical happening? Is it not happening? Is copper in trouble? Whatever. On the one hand, there's Credo kicking the ball out of the ballpark, but on the other hand, SiTime gets to play no matter what. The need for more precise timing, lower latency, greater synchronization, greater stability, better phase noise, lower power, smaller size.
Yeah. The beauty of it is, you mentioned Credo, you don't compete with folks like that. You orchestrate. You help them.
Yeah. Well, they're customers. Almost all of them, whether it's people in the optical world or OCS world, we're talking with some of those people earlier, before this conversation, they're all customers of SiTime.
That's wonderful. Without getting into customer specifics, how should investors think about design win cycles and revenue ramps in AI and cloud infrastructure? Are these multi-year type platforms?
They typically are multi-year platforms, at least two to three years. Of course, what NVIDIA has achieved is cutting development times in half. What we thought we had two years to do something, Jensen tells us, "No, we have to do it in 1 year." We're on that cycle. Maybe it gets even shorter. We are continuing to use AI for some of the development work. Not a lot, but some. We'll probably use a lot more in the future. Our problem has been that we haven't paid as much attention to some of the customers as we probably should because we're still relatively a small company. Last year, we were only $330 million.
Mm-hmm.
The basic consensus has us at only $460 million-$470 million, sub $500 million. Of course, we always hope to do better, but that's kind of where we are. We're investing in sales people, business people, as much as we're investing in engineering.
That's terrific. I guess you've talked a little bit about the resource decision, the opportunity cost sort of decisions that you have to make about markets, about customers in the market. What can you touch on about other end markets like auto, military or aerospace, defense, industrial? What momentum are you excited about beyond?
Yeah. O ne of the lesser-known ideas, like I've said before, is an understanding that SiTime is a very broad-based company. We have anywhere from 400-500 applications at any one point in time. Most of them grow at 15%, 10%, 20%, and then there is about 5% that grow at 30%, 40%, 50%, 60%. For example, right now we really are looking at ADAS, robotaxis. A level three uses car, uses timing to the tune of $5, $6, $7. A level 4 gets to use it to the tune of $30-$35 or even more. Again, because of autonomous driving. That's one. I think embodied AI or embodied LLMs, robots, humanoid robots. There's not many shipping, but we think we have content in that.
Clearly, mil aerospace defense, with the geopolitics as we see them, whether it is smart fuses, smart munitions, communications, missiles, drones, we're in a bunch of those. That's very exciting. We do in fact have, as I said, 400 to 500 applications, so there's always something happening that's very exciting. Of course, there's always the phone, which is a big booster of revenue when it hits, whenever that happens.
Right. I guess with regards to that, right, proof is in the pudding. You've seen your pricing is premium pricing. Your revenue growth curve is bizarrely inverted in that you're growing, your velocity is accelerating every year. Those are great things. There's also been a transformation in the revenue composition. Why don't you talk about that a little bit, Beth, and kind of what that business looked like a few years ago versus the business mix today by end market?
Yeah. We're really excited about the transformation. At IPO, we were about 65% consumer mobile business and about 12% in we call our communications enterprise and data center. We fast-forward to 2024, they were all about equal kind of a third of the business each. Here in 2025, now 53% of the business or just north of half is our communications enterprise and data center driven by AI. We're really excited about that, in terms of both the opportunities to invest with some of those waves in terms of the technology, but also in terms of the value we provide, whether it's ASPs or our margins. Those are all accretive businesses, we're moving in the direction of those more accretive businesses, which is really exciting for us.
In addition, from an investment perspective, we think about platforms. We invested several years ago in these products that are now what is really generating that revenue in AI and data center and are able to take those platforms and also have derivatives for some of these other application areas, whether it's automotive or industrial or aerospace as well. Really being able to leverage those investments into multiple opportunities, that are both high growth as well as high value for us.
Got it. Well, look, this is the end of a very exciting month for you folks. Great earnings call, then obviously right on top of that, the Renesas timing division acquisition. Why don't we spend just a minute or two on your recently announced acquisition? Why was that acquisition so critical to your strategy? How does it change your long-term trajectory?
Well, pretty much when we went public in 2019 November, somebody said, "If you did M&A, what would you do M&A in?" I said I'd acquire a clocking timing division of one of these four large companies, and I rattled off the four large names. Clearly at that time and now, the biggest, the brightest, the most gold-plated version of that business is, in fact, the Renesas timing business at scale, in gross margin, in the quality of the customers. I always like to remind people that this is the IDT business that they acquired a certain company called ICS, which was run by one Hock Tan, and he took that public and sold it to IDT at that time for the princely sum of around $1.2 billion-$1.3 billion. It has very great antecedents.
For us, as a pure play timing company, as a company that plays just one song, it's very important to play it very well. Having a customer, especially the CED customers, especially the industrial, the automotive customers, they all consume resonators and oscillators, which is the bulk of our revenue, but they also consume clocks. In fact, they're part of complementary set. SiTime has very little clocks, and it takes a long time to build a clock, timing, portfolio. The old IDT Renesas business has a very large business. Last year, they did close to $200 million, a little bit more than $200 million, and perhaps they'll do close to something significantly, maybe closer to $300 million, or if not that, somewhere in that ZIP code. The customer consumes both of them.
They need the frequency production with the oscillator, and they need to disseminate with the clock. Therefore, it makes complete sense as a complementary product for a company which has very little clocking business to acquire a business which has very little oscillator business and put them together in one play. What also makes it really interesting for us is that 70-75% is, in fact, in the CED business. They're quite strong in communications, which is with Ericsson, Nokia, and others. They're quite strong with the enterprise, with Cisco and Juniper and Arista. They are also strong in data center, not so much in opticals and connectivity because that's not a clock play, but in all the rest of it. It fits in really well.
It gives us a very good profile, allows us to sell to those customers where we are not strong, we talked about that a little bit earlier, and allows us to take their products to the customers where we are strong with. It's a North America team, mostly. That's good between Ottawa, San Jose, and Tempe, Arizona, so that's good. I mean, it's a team that was there with IDT mostly. They have longevity and as in analog design, you need that kind of long understanding of that technology. I've met some of the people. I think they're an exceptional team. We're very excited.
In the spirit of sort of preserving brand equity with an acquisition, it's difficult to do. On e thing that's always differentiated you is the focus on MEMS, right? It's part of the precision timing. You look at this business, it is traditionally quartz-based with some MEMS oscillators. H ow do you see sort of that fitting? What was kind of the cost-benefit analysis of saying, "Hey, I'm buying something that's quartz. It's a little different than what I have today, but it's worth it?
Right.
Yeah.
It's important to note that out of the $200 odd million they did last year, only $10 million used any frequency. In other words, all they did was $10 or so million of quartz.
Right.
Which is a very small amount, right, out of all of that. They actually have no need for MEMS or quartz. That is why the problem is shifted on to the customer.
Mm.
What we would do is we would make that customer decision so much easier to do in this. Financially, of course, clocks are high gross margin. We're talking 70%-75% gross margin, and a very relatively high net, well, actually very high net operating margin. I think this would make us not only strategically viable in terms of having the clock piece along with resonators and oscillators, but it would also change the profile of a margin, take us close to our highest level gross margin aspiration of 65%, and take us significantly higher than we've always said we want to be in the 30% net profit, net operating margin. It'll take us to that and maybe even beyond.
That's great. Look, I think this acquisition does change the competitive landscape and timing. Can you talk a little bit about how it does that, especially when you look at larger but legacy incumbents and any emerging competitors? How does this deal post-closing change that landscape?
It's important to note that no matter what, even after the scale, there are $300-$450 and change. Their $250 or so is still about a $700 million business and an $11 billion business. It's a small portion in terms of numbers, but there are some pretty formidable competitors out there in clocking. We're talking about Texas Instruments. We're talking about Skyworks. We're talking about Microchip. Three big names come to mind. In the oscillator market, we've always competed with some big names like Murata and Cook, and Kyocera, and some of the other guys out of Japan and Taiwan. It's a hyper-competitive market. There's 40 purveyors of oscillators and resonators. There's about 10 purveyors of clocks. I think the market dynamics don't change much.
Where it changes is where people are interested in precision timing, which is such a small portion of the market. There, it allows us to bring clocks to them, which always have our precision oscillators. As of now, we don't see any credible companies that are showing up. There's a startup that's been trying to do that. There's some of the larger companies in Japan that have stated that they want to do it. We always welcome co-competition because one of the ways to make it easy for investors is to show us competing with a big competitor so that they can see how well we stack up. There aren't too many people in this space in that precision timing space, which is, of course, a category we created. It didn't exist before SiTime showed up.
Yeah. It says something to the staying power of the company, the amount of time you've been able to compete. Beth, back to you just to close out on the acquisition. From your perspective, the CFO perspective, how should investors think about margin profile, integration costs, the path to value creation from this deal? What does a successful deal look like to you in 24 months or 36 months?
As Rajesh talked about, we're really excited about this deal. Clearly, the strategic rationale, as he's articulated, but also the financial rationale. This is a business that's $200+ million in revenue with customers that many of them. Business has been in clocking for roughly 30 years. It's a 70%+ gross margin with very attractive operating margins. It is very accretive to our overall business model. We're excited about that. We're also excited about the opportunity to invest in this business. It's a small division within a very large company. there's they get a specific amount of budget every year, and that's kinda where they're constrained to.
We're excited to bring them into the SiTime portfolio and be able to invest in this business, invest for growth, invest in their product portfolios. Again, modestly, I think we can get great returns to drive that growth of this $200+ million at 70% gross margins, driving strong cash flow for the overall business. An exciting business model from that perspective.
That's great. You touched on growth, and earlier we talked about sort of the, this great historical growth curve that you've had. What is the long-term growth trajectory look like from here? How do investors think beyond near-term cycles? What are the key drivers of sustainable revenue and earnings growth? The growth.
Yeah. W e're still committed to that 25%-30% growth because frankly, the more we dig, the more we find, the more we see the need for precision-based timing. We continue to be focused on timing and timing alone, which is a strength of the company. There are timing pieces, which have nothing to do with components, may have to do with timing modules, may have to do with timing systems, perhaps lower-end atomic clocks, perhaps middle-range to higher-range atomic clocks. We could be in timing IP. We could be in timing software. We just see that the customer's challenge around high-end timing is underserved and that nobody has seen this opportunity the way SiTime has seen it. We intend to keep on moving in that direction, investing in it.
We think that our base market, our resonators, oscillators, and clock market doubles in a decade. If I look at it and I look at some of the great analog companies, I think SiTime has potential to be one of those in a few decades.
It'll be great to see that. Before we just one more question here, but why don't we open it up for questions from the audience? Okay. All right. As we're talking about sort of the future here, what do you think the biggest risks are to the story? technology, competition, macro, and how are you sort of set up to manage through those things?
I think macro in today's world, clearly the biggest problem. When we went public in 2019, there was no COVID, and it's been one thing after another between COVID, the Russian war, geopolitics, China, the threat to Taiwan. Supply chains were never talked about. I've been in this business now for 43 years. Supply chains becoming so critical, where you can't take them for granted anymore the way we did. We thought they were automatic. I think that's the case. Now, in general, the world is in a divided state, and I think servicing all of that becomes the biggest challenge. To me, that's the biggest challenge. As far as SiTime is concerned, we see a very clear path to success. We see a very clear path to growth.
We see a very clear path to differentiation, regardless of whatever competition comes in, simply because we have a game plan, and the amount of focus we put into it has been rewarded. I t's sort of good to look back and see that from 2007, when I arrived, we were Sand Hill Road invested company. In 2014, we sold ourselves to a Japanese company called MegaChips, and we were part of them for about five or six years. Improbably, we spun ourselves out of MegaChips and went public again. How often does that happen? We know what our future is, we know what our destiny is, and we're very focused on that.
That's wonderful. Just last takeaway for an investor. Say one of these folks today, aren't in today, decide to come in today and invest in the company, what do you hope they'll say that the company got right in five years?
I think what they would say or should say or should see or experience is that the company fulfilled its promise of being this highly differentiated timing solution for highly differentiated markets. I think if they did that, they'd be astonished by the breadth of our diversity of our business, because one of the things we see today in semiconductors, very verticalized companies, and my vision has always been for SiTime from the get-go, of building a highly diverse business, very profitable, high growth, very predictable revenue.
That sounds pretty good. All right. With that, we'll leave it there. Thanks very much, everybody.
Thank you.