I appreciate it. Okay. So I think in prior conversations, the way that we started is with a larger thematic, which is talking about how does this AI spend, in your vision, look? Is it premature? Is it front-end loaded? Do we see this bubble? I think a better question tailored for Silicon Labs is, one, I'd love to hear your opinion on just where we are in the investment cycle in aggregate. But two, how do you guys intersect this trend? Obviously, you guys are going to be more towards the edge. I'd love to hear about your products and your vision for the AI dollars coming into SLAB.
Yeah. I mean, Tom, we're super early innings. Look, Silicon Labs, we're not data center focused. So we don't really have a great opinion on sort of how the data center sort of rollout is and where they are in their investment cycle. But I look at it as, look, this edge opportunity will ultimately be in addition to what we're seeing at data center. Right now, you're seeing billions and almost trillions of dollars getting sort of invested in the data center. Why? All the data exists there. It's the easiest to train on. It's the easiest to deploy. That's still early innings. I don't think, as a society and sort of technology leap, that we've fully figured out what the data center AI is and where it's going. I think there's several years of runway left just in data center.
Then from there, that's where all investment is going now. From there, what has yet to happen is it goes from data center, most likely into enterprises. Those are the next set of class of people that have large data sets that can learn on them, that can deploy them, that can add intelligence. I think we're super, super early in enterprise-level AI deployment. So I think there's a lot of upside sort of room to grow within enterprise AI deployment. And then finally, what's just now starting as percolating as an idea, not even deployed yet, is at the very edge, like that physical AI. And in that, if you just look at the number of devices that are deployed around the world, you're talking billions and billions of devices. That's added on top.
So I think I look at it as a multi-year journey that's starting with data center that will morph into enterprise that will ultimately become even bigger and expanded as it gets to the sort of far edge physical AI. So I think we're super early, Tom.
Yeah. And when you think about the edge, something that I always kind of go back and forth with is you've got a device today in everyone's pocket that has run 99% of your applications for most of your life. You have this fork in the road, right? Do you think that it's your smartphone that brings you Edge AI, or do you think that there's a variety of other devices that come into play? If you look at where other products have tried to intersect this, it's glasses, it's wearables, it's watches. Do you see future content for SLAB making its way into this device, or do you think that you see this burgeoning of other products at the edge that are AI intelligent?
Yeah. My view is the other products become more intelligent over time. And I think it really is around what is the computational requirement? What is the user, if it's machine to machine or person to machine, what is that user trying to accomplish? So if you're at the data center level, you're trying to accomplish, hey, crunch a lot of information, do a lot of correlation. If you're AI PC, AI PC is sort of new over the last couple of years. A lot of effort's been going there. That's a next level computational. It's not data center, but it's also not phone, right? So you can do a lot more on the laptop at that level. And then phone, right, you're doing a lot of querying. Today, most people on their phone might use it for videos, Google search type applications.
And that has a lower processing power, although it's very powerful. Lastly, when you get out to the physical edge, the Edge AI, that is machine to machine or automating sort of things around you. So I think ultimately all of those links will have these intelligence engines in them. You'll have training at one end of the data center and inference for the most part at that far edge.
Helpful. I want to go into the technology first, then kind of go into the verticals because your technology does cut across both of your verticals. Maybe talk about why the connected MCU is so important when you look at the edge. You have a variety of different radios that you're attaching to your MCU. I think that you were very early in this transition to offering a suite of different connectivity options with your MCU. In fact, we're hearing about more companies coming into the fold today. Maybe spend some time talking about what differentiates you in the connected MCU space. I know you're moving from Series 2 to Series 3, both on the connectivity side and then also on your microprocessor side.
Yeah. I mean, I think one thing that people don't always appreciate is almost everything that we sell is an MCU, right? Most of them have radios attached to them. So MCU plus Radio, they're fully integrated. In some spaces, we'll sell people just a vertical MCU or just a vertical Radio. But the vast majority of our products are the intersection of these two integrated radios plus MCUs. So there's different needs depending on different applications. So to give you an example, if you're running an application today that might be a fairly basic, call it a commercial HVAC controller. While commercial HVAC controllers have sort of always existed and there's always been an MCU in there and that's sort of controlling that, a lot of our market opportunity comes from, hey, we now want to network that. How do we take what we've already got and network it?
And that means sometimes you're augmenting an existing MCU with a radio. Sometimes you're replacing that in its entirety. The biggest inflection we see is more and more move toward networking things. And that's sort of at the very basic level, the investment thesis that Silicon Labs has is, hey, there'll be more connected devices in the world going forward. And that requires both MCUs and radios. To give you an example on sort of we're always out in front of most of the end markets that we service. In Series 3, which is our next generation device coming out, we've designed that to be ready for the future of AI coming to the edge. And that comes in a few flavors, but primarily in processing capability. So our Series 3 device at the higher end will support up to 100x the processing capability of a Series 2 device.
So really sort of extending the processing capability on the next generation. It also adds a lot of features that I think people underappreciate today that we'll learn over time that will become more and more important in our world as there's more connected devices and these devices are doing processing with real-life data. I think security elements come into play as well. I think that's an area that Silicon Labs has done a super job over time is actually adding leading-edge security protocols that whether that's PSA, so PSA is a third-party standard, Silicon Labs was the first to achieve level three. It's first to achieve level four, which is the current existing state of the art. On next generation Series 3, you have all this extra processing power 100x. We're also upping the game in security around that.
So as you're processing, how do you encrypt these communication flows? There's one version that is likely to be post-quantum computing resistant. How do you really sort of spin your ciphers in the most encrypted way possible? And look, not every application is going to require that level of security, but we think the world is underappreciating what security will be needed in this much more processor-intensive future.
Yeah. We're definitely hearing more of that. We had Synaptics up on the stage yesterday, and they talked about a specific AI core that they were using within their next generation processor that's going to be connected as well. They play more in the Wi-Fi space. That's not an area that you spend as much time in. But when you say more processing power, does that just mean dedicating more cores to AI processing? And then maybe what does the Series 3 bring from an ASP uplift perspective? Should we be thinking about this as a tailwind to total revenue on kind of flat units?
Yes. I mean, first, generally, I mean, across almost all semiconductors, you can generally assume that, hey, a next generation comes with higher ASPs. And why is that the case? Because next generations either are faster or more efficient or have more features. So that would also be the case for Series 3. But I think it's important to note that a lot of applications are existing Series 2 is still hands down winning in the marketplace. And I think many semiconductor players, it's all about sort of the next processor that's coming out, the next set of features. Where Silicon Labs, actually, the things that we have in hand actually are servicing most of the market needs today. And as we go into Series 3, it's out ahead of the need. So we think that we're sort of future-proofing for the most part.
When it comes to processing power, what does that mean? I think was kind of your first question, Tom. It's not just that the AI accelerators and NPUs that are onboard the chip. That's not how the processing power is going up. There's different sizes that you need. So depending on the complexity of the algorithm that needs to run, you either need big NPUs or small NPUs. And so there's sort of different flavors on size and capacity. But for the most part, these are multi-core processors that are running on the MCUs. And there's sort of bigger processors that might be used for operating systems. There might be accelerators that are linked in. There might be network co-processors. And then there's a whole array of things that are running on chip and things that are running even off chip.
As computational power goes up, one of the sort of constraining factors is memory. So customers are wanting to run more and more software on chip. And that I don't think is actually going to slow down anytime soon. As AI sort of keeps moving forward at this fast rate, software is going to be more and more important. And the computational power is able to keep track of that.
Super helpful. Last one on the technology side. Series 3, when are we going to see that in the market in more volume?
Yeah. Series 3, the first device is already shipping high volume. Earlier in 2025 is Q1 or Q2, somewhere around that time frame. Our very first Series 3 went into general availability. Therefore, it's already shipping in high volume. The second device is out soon. The third should follow pretty quickly. The first sets of the Series 3 family are now starting to come to market. The first one's already shipping, and it's reasonable volume. I think what you would see is the additive of revenue will continue over time. Actually, Series 2 design wins are right now actually growing even faster than Series 3 design wins, given the recent nature of Series 3 just coming out. That, I think you will continue to see a very long lifespan of revenue generating out of Series 2.
These devices are generally a 10- 15 year life device. So we're sort of maybe not even into the sweet spot where Series 2 is, and now Series 3 is already coming out.
Super helpful. Okay. So you've segmented your business into two buckets. They're both growing quite nicely in calendar year 2025. I kind of want to walk through each of them. So why don't we start with Home & L ife? A broader trend that you guys have talked about is CGM. And that's been a big vector for you guys, so continuous glucose monitoring. How big is that business for you now? And then can you describe where you're positioned in that market and what customer engagement has been like so far?
Yeah. Home & Life, just maybe to start off, we have sort of two sides of the business. Home & Life, which is the smaller side, and then Industrial Commercial. And to a first order, it's 45-55, sort of the bigger is on the industrial side, or in some cases 60-40. But it's a pretty balanced portfolio. Within sort of the life side, which I think you were touching on, Tom, that's probably the fastest growing piece of the portfolio. We've spent a lot of time over the last couple of years sort of focused on connected medical devices. There's sort of an expanding need for monitoring, so health monitoring sort of in the world. One of the big areas we found a ton of traction is continuous blood glucose monitors. And for anyone not aware, generally you might see them on the triceps.
It's measuring insulin levels for people that might be diabetic. That's going through a bit of a shift where it used to, if you recall from days of the past, you would hold up your phone to the triceps. It would take the reading. That's generally an NFC communication. That's largely converted over to Bluetooth now. And our Series 2 sets of devices have really hit the sweet spot for a lot of these makers of glucose monitors. And that's where we're seeing sort of the fastest uptick in traction and deployment. We have publicly disclosed that we've won about a dozen of these customers now in their various ramp phases. We think that that gets to about 10% of revenue pretty soon over the next couple of quarters. This is in comparison that if you run the clock in 2024, the revenue from that application was approximately zero.
It's not too often as a company that you come across these zero to one movements. This is an opportunity that we've been able to seize on where we're literally going from zero to very fast revenue and adoption. We said, I think in our last public call that medical has now grown 60% year- on- year for us in 2025. It's been a great business. And look, we're not done. We're highly confident that we can win more customers in this application and more even meaningful customers. We've engaged with 60 programs now. We're very confident that we'll win a lot of these, although we haven't officially announced any new sort of meaningful wins yet.
We had a good conversation earlier this year around the customer set, and if I remember correctly, there's a couple of chunky customers at the top that when one customer converts, that can be really beneficial for you guys from a growth perspective. In terms of your strategy with those customers, is it just a bring out new devices, show that they're very performant, show that they're very cost-effective, and you hope that one day they move over to you guys? Or is there anything more specific that you guys can point us to as reasons why you take down one of those big guys?
Yeah. I think our observation around this set of applications has been our feature set is like a perfect fit for most of them. Like I said, in days past this might have been running NFC and this market sort of transitioned into Bluetooth. What we find is the people that have actually adopted Bluetooth as a technology have put in relatively basic older sort of Bluetooth applications where they have weaker energy sort of focus. And for something that's only going to last a week or two, the energy sort of consumption rate means lower batteries, right? So smaller batteries, so less energy usage using smaller batteries. And these things are disposable. So you don't want a whole bunch of batteries going into landfill. So any extent that you can shrink battery size for things that are disposable is sort of better off for the environment.
In a really big way, actually, the security element that we've added into this lineup of devices has been a meaningful needle mover for most customers. When we approach a new customer in this application and we explain to him, "These are the feature sets that we can bring," it's generally met with open arms, like, "Yes, how do we enable some of these security features?" Yes, we have high interest and energy consumption and reliability of signal around these things.
Okay. Let's go to the other side. So Industrial and Commercial. The two big trends you've talked about there are ESL and smart metering. ESLs, I think, are a really interesting place to start. In terms of your ability to also gain chunky customers there, so why don't you start with what's the transition we're seeing in stores and in labeling in general and what value do you bring there? And then I think a similar trend, honestly, where you've done a very good job of getting into the market and seeing some really strong growth, but it's kind of the next big chunky customer is the next step forward for you guys. So what gets you to that next customer as well?
Yeah. That one, this one's really more about the market. ESL as a market rather than, "Hey, you got to win some really big customer.
Adoption. Yeah.
So that market is relatively early in their adoption, right? And I think if you wind the clock, just like three years, you would find, "Hey, a lot of people in testing trial phases." Now we're starting to see some pretty big retailers take substantial steps forward in sort of deploying this stuff sort of really broadly across an entire geography or entire set of stores. So that market looks really healthy. And we're sort of seemingly out of the trial phases, if you will. And trial phases for us, just called the last few years, we shipped 300 million devices. So this has a huge, huge quantity potential.
We've heard from some retailers that have sort of published articles around this that says, "In some stores, you might find up to 120,000 SKUs in just one big box store, so if you say, 'Wow, if you put a connected price label at every SKU on every shelf in a store, one store is 120,000 units times number of stores times number of brands kind of worldwide, you quickly get a huge amount of volume,' and I think there's just a lot of room to run globally for the deployment of these things. In the US, you'll see retail kind of deploying this. In Europe, you see grocers deploying it, mixed bag in some of the APAC countries, Korea, Japan, etc., so we're early on, and this ends up being. It comes actually back down to battery consumption.
So we can make a shelf label tag last seven years on a coin cell battery. So that's pretty amazing. No one else would walk around and change batteries. And interference. So think about if you're in one of these stores. It's got 120,000 of these radios sort of simultaneously communicating. There's a big sort of interference engineering challenge to solve. And we're super good at that. So this sort of fits right into our wheelhouse.
A longer-term trend that we've heard a lot more about is smart metering. You had some geographies that were rolling out, including India, that had been pretty successful for you guys. Where are we in that rollout, and do you think that that continues to be a major driver of growth?
Yeah. Smart metering, I do think a lot of people pick up India because it's a new driver. It's a geography that's deploying really for the first time.
A lot of people.
Massive volume. There's a lot of people in India. But what we're seeing actually is a broader worldwide set of strengths that if you asked me two years ago, will there be sort of a worldwide push for more and more of these meters? I probably would have said, "Hey, it's slow and steady, typical industrial." But what we've started to see is more and more upsides actually broadly in that market on people focusing on energy usage from electricity, resource monitoring for water, resource monitoring for natural gas, and we shipped all three: electric, water, gas in just about every country in the world. India has been much faster in deployment than we've seen, so we've been in this business for 15+ years supporting these smart meter deployments. We are probably market leader in most of the communication links for these meters.
And India has just come to market a lot faster than we've seen others. And this is a 250 million unit deployment that we're still relatively early. We started deploying these in kind of mid-2024. And the ramp sort of keeps going. It's not yet at an asymptotic sort of normal run rate. It's actually still ramping up. So that's a good sign.
I want to pivot to gross margins. When I think of gross margins, there's multiple factors. And you guys have shown some really good results lately. I think one that you get asked on a lot is pricing and how pricing is trending as a business year- over- year. Two is really thinking about product differentiation. So moving from more Series 2 to eventually more Series 3. And then three is disti versus end customer mix or sell-in or sell-through, essentially. Where is your revenue exposure going? And I think those three have kind of combined to result in some gross margins that are better than your long-term target. And so just talking about the sustainability of those gross margins, what are the pieces that are most impactful and where do you kind of see those margins trending in the near term?
Yeah. I mean, look, I think people wrongly look at Silicon Labs and think, "Wow, that's sort of a consumer-focused business." It's really not. Actually, it's largely an industrial-focused business. Even though we may say Industrial Commercial is 55%-60% of the portfolio, most people will call it medical industrial if you look across our peers. So we're highly an industrial sort of focused company, more so the consumer. And hence, our gross margins are in the high 50s. So the longer-term model that we've communicated is 56%-58%. We continue to believe this is a high 50s sort of gross margin business. In the near term, like the last couple of years, this year we'll probably shake out 2025 and kind of what's looking for 2026. We're probably at that kind of high 50s, low 60s. And industrial has been doing super well.
Distribution channel has been super well. Those are all sort of additive to gross margins for the company. This is something I think is sort of underappreciated from the marketplaces, just the sustainability of these high gross margins relative to a lot of our peers, and I think the biggest difference is many of our peers are more consumer-focused than we are.
When you think about two trends that people worry about broadly in the space, one is exposure to China just in general across semiconductors. And then two, we're hearing about potentially memory shortages impacting consumer demand, which you just described was really a minority of your total business. And you look out into 2026, are either of those a concern for you in terms of either lumpiness in revenue trajectory and/or limitation of end device sales?
Yeah. I mean, we haven't seen anything that's materially affecting our customers from a memory perspective. For the most part, most of our customers don't have huge memory requirements. I mean, of course, they all buy different memory products. But you're not talking really high bandwidth, like Gb of memory generally. So I just think of our class of customers, that's not a huge concern from them, at least not what we've heard. So I don't think that's really a sort of dynamic happening right now, Tom.
So if I look into calendar year 2026, I think you guys have done a really good job of verticalizing your opportunities, showing some growth trajectories, a lot of transition from a technology perspective. I think it's been a lot of a show-me story after a recovery post-pandemic where you guys have continued to build the revenue base. Gross margins have been better. I think that the question in many people's minds is just, how do you get comfort around these opportunities when it's very customer-driven? And it's not really easy to do a ton of bottoms-up work on these verticals. So I know you haven't guided 2026, but I guess what are you most excited about in this coming year? And where should investors be paying attention as we look to learn a little bit more about Silicon Labs' trajectory?
Yeah. One, the verticals that we talk about. I mean, people like to do a lot of homework around sort of the verticals that we talk about. And look, those are design win stories, right? These are like, "Hey, where we're focused," and you have a bunch of design wins. But I think what's often missed is when people go, "Hey, show me, I'm waiting to see how the growth sort of materializes." People discount the history of this company. And you can do the start and endpoint math differently. But if you look over the past 10 or 15 years, this company has consistently grown 15%-18% compounding annual growth rate. So it's actually not a huge leap for us to say, "Yeah, we think we can sustain 20%." And I think people sort of forget the history of sort of where this company's been.
Look, the supply chain over and under has created some noise around that. How do you get comfortable with it? The base investment thesis is very simple, at least in my eyes, Tom. It's, "Look, there's going to be more and more connected devices around us in the future than there are today. Those connections will require a wireless radio. They will require an MCU." I think at the end of the day, that's the bet that actually I think we're all making. I feel very comfortable that that is a winning bet. I mean, I just don't know anyone that would debate and say, "Hey, that will not be true in the future." I think that's at the high level, that's what we're anchored to, Tom.
Super helpful. Last one, capital allocation. You've seen a recovery in both revenue and in the stock. Stock buybacks, capital returns, M&A, anything worth commenting on that you guys see as a priority?
I mean, nothing new. We haven't really changed our tone and position around this, which is, sure, we'd like to do some M&A activity, but we have a really fine filter on what would make sense for us, which is, "Hey, we like our wireless end application or our IoT end applications. We like wireless technologies. We like MCU technologies. We are really focused as a company on that." And I think that focus really is the magic that sort of what Silicon Labs is, is that focus on executing. So if we were to go off and pursue M&A activity, we would want to maintain sort of that focus, which means, "Hey, as cash flow continues to go up and as sort of growth comes through to the bottom line, most likely this will end up getting deployed through buyback means rather than M&A means.
Very helpful. Thank you for being here. Really appreciate the time. And it sounds like exciting times ahead for SLAB.
Yeah. Thank you. Thanks, Tom.
Yeah. Appreciate it.
Thanks, Tom.
Thank you, guys. Thanks.