Great, so I'm Brian Gesuale. Really happy to have Cerence here to take us through their story. It's a very exciting AI-driven story. They're a market leader, and we think the time is really interesting to look at it. We're gonna do a fireside chat format, so feel free to get ready and load up some questions for when we do finish, but we have the company's Chief Executive Officer, Brian Krzanich, and Chief Financial Officer, Tony Rodriquez, here to present the story, so welcome, gentlemen.
Thank you.
Let's start maybe with level setting the audience here to the Cerence story. Would you explain kind of the key IP, the competitive moat, really how you size your serviceable addressable market, and maybe talk about your geographic revenue exposure as well as how you report the business?
Sure. I can start and just kinda talk about our IP and, really, our how we think about the business in general. And then Tony can give you a little bit more detail about how we break out the various business segments. But our intellectual property goes very far back into the early days of voice. So you-we spun out of a company called Nuance. Nuance was one of the very first voice, if you ever used Dragon PC, the one where you could put on the headset and talk to your WordPerfect, if you wanna age yourself. That was Nuance. And so our IP spun out of that specifically for automotive. We have a lot of the fundamental IP around voice that came with us, things like text-to-speech, Wake- Up Word. Those are IPs that this company owns, believe it or not.
What's happened is all of that has become very agentic and LLM-based. So almost all of those things that, how you recognize a voice and how you recognize a word has all become a neural, large language model-based system, all the way up through orchestration levels. Our moat is that technology. We're very good at this technology and especially integrating it into a vehicle, which is also a very difficult process to do. The vehicle, the hardware is constantly changing up until shipment and even sometimes after shipment. And integrating and working with the car OEMs is very different than, like, loading software into a data center or a PC or a phone or something like that. So all of that is really how we make our money. We're in about 50% of the cars on the road, roughly.
They have some level of our software in there. And so, in our geographic, we're in almost every single geography in the world, including Chinese OEMs outside of China. Inside of China, they're kinda required to use a Chinese version of us, but they choose to use us, BYD, Great Wall, Zeekr. They use us for our language capabilities and our software capabilities outside of China. So pretty much everywhere those cars are sold outside of China, it's us in there doing the voice and activation. And we do basically. We can do everything. And what's nice is you don't have to use keywords anymore. So this is the cool thing is you can now just say, "Hey, crack the window. Hey, my feet are cold," you know, and you can just talk to your car more and more like a friend, and that's gonna become more and more prevalent.
It's gonna talk back to you, recognize you, welcome you. "Hey, you're stressed. There's a lot of traffic. You wanna take a rest?" Things like that.
Great.
Business segments?
Yeah. So and geographically, from a revenue standpoint, there's about 16% of our revenues from the Americas for 2025. We just ended our year. Our year-end is September 30th. And then the rest evenly split between Europe and Asia. And so that's kind of from a concentration standpoint of where the revenue comes from. From a segment standpoint, we only have really one reportable segment, which is automotive now. Within that, we have revenue streams, our license revenue, which is, you know, software that's within the vehicle, embedded within the vehicle. You have a connected service revenue stream where, with if you have an embedded technology, you could also connect it outside, ask questions outside of it. That revenue stream is built a little bit differently. Embedded, we record that revenue when we actually ship a car.
The connected is over a subscription period, and then we have professional services as well.
Fantastic. Maybe let's move on a little bit to the competitive landscape. It's a question I get. I think people don't appreciate your position in the dominant position you've had in the automotive space. On your call, you mentioned not much has changed from a competitive standpoint most recently. But as I think about this, you seem to have several factors kinda coming together. It's, you know, one, customers are moving more to agentic AI, seems to really align to your product roadmap. Can you square maybe your product roadmap and the xUI platform, the move to agentic AI from a market standpoint, and then how you compete given the vast amounts of capital that are available to private companies, as well as your hyperscaler competitors as well?
Yeah. So if you take a look at the competitive environment right now, it nine times out of 10 usually comes down to the last two players on the playing field for bidding for a deal are us and Google. Occasionally, we see Amazon in there. But it's usually one of the three, mostly two of us, kind of at the last bidding of a deal. That's how it's been at least the year and a half or so that I've been here, and from what I've talked about with my team, it's pretty much been like that for a while. It is becoming more competitive, in that Google is getting better. They used to only sell a cloud version of their software. Now they have, but OEMs always want cloud and embedded because embedded on the vehicle is cheaper, so they've now come up with both.
You know, our technology and their technology, we're constantly battling out. We moved to some of the LLMs on the vehicle first. They come back behind. So it's a good battle. I think that always makes the market better if we do that. Price-wise, it, you know, what usually happens is in the marketplace, we have to compete with a combination of the level of service we provide, the amount of, we're agnostic. So if you go with Google, they want you to put the data in their cloud. They want you to use Gemini. They want you to use Google Search for things like, you know, where's the nearest Starbucks, parking lot or charging station. All of those kinds of things become kind of restrictive. We're agnostic. That's really how part of our sales pitch is. We can connect to DeepSeek. We can connect to, Gemini.
We can connect to Claude. We can connect to Mistral. We can change it as you move across geographies, and vehicles, and shift who, where, and how we connect. They can customize, and that's really how we differentiate. Our software is very good. We can compete very well head-to-head with Google. Pricing, it's, we're, you know, very efficient, and we're able to compete on a pricing level easily with them, and it's really, you get a big tech name if you go with them, but you're fixed in a kind of a silo, or you can go with us, and you're agnostic, and you get, you know, the options to really change, and as they start to see this world and how fast it's changing, that agnostic view is actually becoming more and more valuable, that they want the freedom to be able to adjust.
We can turn switches in the cloud, and they can go from ChatGPT to Claude overnight, and the end user doesn't know it. Those pricings are becoming very competitive.
Very, very interesting. I do think there's a lot of value on that flexibility and then even just your ability from a regional perspective to offer a much more open solution than what those can.
Right, and you saw, like, you know, OpenAI came out with Code Red that, you know, they were Gemini really advanced, and now they are accelerating ChatGPT 5.2, and they wanna go faster. They don't fear being left behind because they can just flip a switch with us and go, whether it's a better price or a better product, they can get it.
I like the sound of that. Maybe let's dig into some of your products and, whatever direction you wanna take this. But I was thinking we go look at the xUI platform. You have a couple of customers in hand, and it sounds like a really healthy pipeline of customers engaged. How should investors think about the production levels of those current xUI customers, their ability to grow, and then the amount of production that the half dozen or so that you articulated on the call would grow or represent?
Sure. So first, we should, for this audience, make sure they understand what xUI does is really takes the end user, the car, the vehicle, and you, the driver, to the next level of really a true agent in your vehicle. So it's now going to - you could be driving down the road, and, you know, it uses the cameras that are already there for parking and sensing to - you can say, "Hey, I think I just passed a billboard." And this is working today. This is not, you know, scripts and slides. "Hey, I passed a billboard. Can who - what? I think it was a concert." And it'll say, "Oh, yeah, it was a concert with so-and-so. It has a phone number. Would you like me to call it?" And it will do all that and help you orchestrate that.
It'll sense, "Hey, I noticed you're driving a little erratic. Would you like to change the temperature, put on some music, stop at a Starbucks? What would you like?" Right? And it will learn from that, as you keep going. We've announced two customers that will be launching roughly mid-next year. I say roughly 'cause they own that schedule, not me: JLR, Jaguar Land Rover, and one of the VW Group. I'm not allowed to say which one, but it's their whole stack, all of the vehicles within that group, right? And Volkswagen has a bunch of them, so you can try and figure out which one. We have, like we've said on the calls, about a half a dozen. So those will launch next mid-next year. They'll ramp. You'll start to see revenue at the very—you gotta remember 'cause Tony said we're, September to September.
So they launch at the end of the summer. We'll get just a little bit in fiscal 2026, but 2027 is really where you'll see the ramp. We have a bunch of other customers. We reviewed one deal on Friday. It's a 2028-2029 model year 'cause that's what they're working on right now. I was working with a U.S. manufacturer. They're working on a 2029 model year right now. That's what people are working on, unfortunately. That's, that's my biggest complaint about this industry is the OEMs are slow. Some of the Asian OEMs are even slower.
Right.
Than the U.S. ones.
So I think, you know, you gave us kinda the back end of this. It, you know, these are 2029 model years, so a few years out. How long are you in that solicitation process to get designed in before you work on something that's three years out?
Yeah. Usually, it takes, you know, that solicitation. We have to sometimes go in and do a proof of concept where we'll—they'll give us a head unit, or we'll bring in somebody else's head unit that already has—the head unit is the cockpit, the, the screens and displays. And we have to show them that. We have to work with their engineers. Our sales process is not a salesman's job. It's an engineering job. It's, it's really engineer to engineer. Then they hand it over to the procurement team to try and beat the heck out of us, right, and really put the bidding process in place. But the engineers usually give them the kinda, "Here's our top choice, and here's our last choice." That whole process takes about six months, six to nine months.
It got a little slow with all the tariffs because those guys went into kinda turmoil of, "We don't even know what and when we're gonna build things and where." But, you know, like, six to nine months is a pretty good thing. And then it's about a year for them to kinda, like, solidify the hardware. And we're actually the last guys to show up and, you know, they're working with Qualcomm and MediaTek and guys like that working on all the hardware. And they're letting us know what it is, and we're kinda planning for that. And then at the last minute, they give us the final configuration, memory size, and everything, and say, "Go, make it work.
It sounds great. Sounds like a lot of activity happening.
The LLMs have made it easier for us to get that done quickly.
Yeah. Maybe, you know, thinking about these timelines and the duration that it takes these OEMs to kinda spin up model years, it sounds like everyone's doing kind of a reboot of their cabin and their AI capabilities. When are you kind of penetrated with this next level, fully penetrated? Is that, you know, sometime in the mid-2030s where these OEMs have kinda come together and maybe just help people understand the magnitude of incremental content per vehicle that you get from this as that happens?
Sure. Like, as you said, the OEMs are really, I think they're starting to think about the cockpit differently, especially as more and more autonomy comes around. You're starting to see screens that go, you know, displays that go all the way across the cabin. You're starting to see things projected up onto the window screen, and those are all targeted like 2028, 2029, 2030. I'd tell you, I think, you know, you'll see, I think we'll see the large percentage of the deals out there in the 2028, 2029 timeframe. That's when they'll start their ramp. Most of those things are five to six-year, you know, deals where they're gonna build that car for five to six years, that hardware stack, and so, you know, I think that's where you'll see it. Then we produce a PPU, we call it, a price per unit.
Tony reports that each quarter. It's a 12-month trailing, because of the way he said, like, some of the revenue comes in when we sell a car, and some of it we have to apportion over a period of time when you buy the car. It's at $5 today. We haven't told anybody what the actual price would be with the next generation, but I can tell you it's significantly higher than that $5 number. Thinking of the numbers we're quoting today for prices are a multiple of that number. So but it's out in time though, right? That's a 2028, 2029. In the meantime, that doesn't mean prices are gonna be flat. We're working on doing, you know, cloud-based features and upgrades to their existing systems that we can turn on over the air today.
Fantastic. That actually dovetails out really nicely into my next question. I thought your, your fourth quarter was exceptional. You really surpassed our expectations across all the major KPIs, and I thought, the guidance looked really healthy as well.
That's Tony's work.
That's, it's spot on on the core tech. It sounds like we're looking at about 8% kind of core growth. I wanna just kinda dissect that across two sector, two kind of vectors, really. Your volume versus price growth, which I think dovetails into what we were just discussing. But also, it looks like IHS production's about 1% next year, so you're 7 points growth over market. So maybe discuss kinda the confidence in that growth outlook as well.
Yeah. So the number you quoted is our core technology. We'll talk a little bit about some outside of automotive patent license ones that we've announced, which are exciting as well. But for our core technology, the license and the connected, we see a roughly high single-digit growth, so you know, 8% or 9% growth in the core business coming from two ways. If you think about price volume, we give a metric out, which is our penetration rate. It's right around 50%, a little over 50%. So meaning every other car that rolls off production has some sort of Cerence technology. So when you think of that volume level, we think of that as being relatively flat, consistent with IHS over time.
So then you think of well, how does that then, if it's flat volume, translate into those two components, license and connected? On the license side, what we're seeing is that more of the shipments of that volume are going to make it, as in, your revenue, primarily because in previous years the company took advantage of these relationships where we would negotiate with our customers to have them pay us upfront for the volume that's gonna ship in the future, and it was non-cancelable. So they were able to accelerate that revenue, and so we've done less of those deals 'cause what we had to give up for that was discounts on the kind of contracted price. So we've said, "Listen, we would, we know that the volume's gonna come in.
We'd rather not give you that discount and just take it, in here." So what we're seeing in the license side is that 8 or 9% growth on the license side is really coming from more of those shipments in year are becoming in- year revenue. And the connected side is growing on volume 'cause you can see that we report what's an attach rate. So of that volume, how much, how many of those are connected cars? And that connected attach rate is increasing. So the growth on the connected side is really that growth in connected cars, the growth in the billings. And then, as Brian mentioned, we amortize that over time. So those are historically, over the last 18 months, we've seen billings increase.
With those billings increasing, going to deferred revenue, we're amortizing that. So it is really connected volume and then more in-period revenue recognition for the licensed revenue.
Yeah. I thought we saw a really nice jump up sequentially on the connected, which as that water falls through, it should continue.
Right. Exactly. It's a kind of a compounding effect, right? And I think we get kind of jaded in the U.S. We think, "Well, every car must be connected." But I think if you look at worldwide, the estimates are only about 40% of the vehicles are connected. And the estimates I've seen are 20%, 30%, maybe 70%-90% are connected. So there's still a lot of growth in connected vehicles in time, to go.
Yeah. Absolutely. I think you alluded to this a couple moments ago, but it seems like your IP monetization strategy's really coming into focus. You're fresh off of a really big win with Samsung for $50 million, and you talked about having other active litigations with Apple, TCL, Sony, and a few others. Obviously, the right path, the best path is for them to pay license fees upfront. But I like this IP monetization strategy 'cause you have a lot of great IP. Help us understand maybe the size of this opportunity for you and discuss kind of where these large companies are having the most difficulty creating your IP. So maybe it's a you guys will double team this question from that.
Yeah. I think so what we're seeing is that. Yeah, we announced that we had our first patent license agreement 'cause we don't really call it selling. Yes, it did go through, potentially through the courts, but we got to an agreement on a patent license outside of automotive. So this is with Samsung, for non-automotive, obviously non-automotive. And it was roughly $50 million of a one-time patent license. And we've disclosed as well. We've got a few more of these cases coming down the pike, within the television side, really Sony and TCL, and then Apple as well. And so when you think about, and then we've got, as Brian mentioned, this is foundational IP. So if you think of anything that has a wake-up word or has a kind of a text-to-speech, we can go after.
There is a good pipeline behind these ones that we've announced. What we've said is, "Look, this is $50 million this year. Well, that's the only thing we've projected in our 2026 guidance. Could we settle and come to an agreement on a patent license or running royalty license for these other ones in 2026?" We could, but it likely may, you know, we position that it might be in 2027. The way we look at it is that pipeline is. We gotta think of it as over five years. If you think about that, average that over the five years, you're probably gonna come up with a very similar number on average 'cause some will be smaller, some will be bigger. That portion of our business is gonna be a little lumpy.
Our core technology business within automotive will be, you know, we'll be able to forecast that very well like we have. But this will be a little bit lumpy. Some may come in a little bit less, some years a little bit more, but it is transformational on a couple things. One is the magnitude of the dollars that we're seeing, but that we're monetizing outside of automotive. And those efforts will dovetail with business development. Look, we're gonna go into these areas outside of automotive and look to gain running royalty rates for what we believe our technology is doing. But if they don't follow along, we have the hammer as well.
And our process is, we send out a letter and say, "Look, we believe these products infringe on these patents, and we really think you should have done a license in the past, and we'd like to come and talk to you for the business agreement." And we wanna, we always want to not sue and avoid that, but just actually work. If they're unwilling to work with us, then we'll sue. But I think in the past, they didn't believe we were gonna go take that next step. And so they just kind of ignored everything. And we're like, "No, we're going for it. This is ours. And we invested, and you guys as investors paid for this, and we're gonna get the value for that." And so that's why we've been much more aggressive over the last year to go do this.
Yeah. Absolutely. And I think having a win under your belt is probably.
It sets a benchmark. You can now, if you guys wanna figure out, you can do your own math. We don't project this, but it is volume-based. And so you can take a look at its portfolio for these right now that are announced, their phones, its phones and TVs. So you can look at Sony, not very many phones in the, and it's U.S.-based 'cause the patents are most protected in the U.S. 'cause of our laws. But there, there's some in Europe too that we could go after. You could just look at Sony, not many phones, but a lot of TVs, but, you know, probably not as big as Samsung, TCL, no phones, some TVs, Apple, a lot of phones, a lot of phones, not many TVs, but a lot of phones.
So you could kind of scale relative to Samsung the size of these deals yourselves. You don't need a forecast for this kind of.
Yeah. It's gonna be interesting to see how this one plays out over the next couple years. Let's talk a little bit about margins here. Looks like the gross margins are tracking really healthy in the 80% range in 2026. And EBITDA, excluding the IP monetization, looks like kind of low- to mid-teens. Can you help us understand the cost reduction activities that you're undertaking, how much savings that can accrue on an annual basis, and how the costs and benefits flow through the P&L?
Yeah. So it levels out a little bit. So we did some restructuring of the business in the fourth quarter of 2024, and we said, "Look, we're gonna cut these costs out given where we think the run rate of the core business is and make sure that 2025 was profitable from an EBITDA standpoint and, more importantly, cash flow positive," because we had some debt coming due in 2025 as well. So we going into 2025, those were all to be determined. We executed 2025. We had a growth in our core business. We had good margins both on the gross margins and EBITDA margins, and we produced a significant amount of cash, $48 million, double what we had anticipated going into the year. So in 2025, $250 million plus on the top line, we're gonna grow revenue in 2026.
Mid-guidance, it would be about $310, so about 23% growth, 80% margin overall. But remember that benefits from this IP, that growth, benefits a little bit from this IP as well. And the margin, the gross margin benefits from the IP. If we just strip that out and look at our core business, as we said, our core technologies are growing in high single digits. Our professional services is actually gonna be down year over year, but it's our lowest margin business. And so we're somewhat okay with that. And it's become because our technologies are becoming easier to implement. And so it takes less NRE to do that effort. When we think about that business, what we've said for 2026 is we're gonna keep our really our hardcore costs flat.
So, you know, what are adding to costs this year outside of automotive? Well, the legal costs associated with the Samsung patent license agreement will be in Q1. Throughout the year, we've added about $8 million of legal costs throughout the whole year, without projecting any additional IP revenue in that year. So it's kind of front-loading the cost associated with that. Overall, our hard costs, like I said, have been flat when you have some geographic issues. So in our R&D, you know, within the R&D effort, some of it gets capitalized and some goes to OpEx. The amount that's capitalized last year, this year will be lower. So more of it will go to OpEx. So it negatively impacts. That's about $8 million [that] negatively impacts EBITDA, but from a cash standpoint is neutral.
So the whole point is that can we be, and if you scratch all that, we were about 20% EBITDA positive last year. But if you take out some one-time benefits that we had, it's about 17%. We'll be about 20% positive this year when you take out all that noise. So we are growing it. We'll grow EBITDA overall by 30%. So we're mid-guidance around $60 million of EBITDA. So the goal for us is to run a healthy, growing, profitable core business, and then you add on the outside of automotive IP business, and it adds to the profitability. Hopefully that helps.
Yeah, it does. I wanna maybe just, you know, kind of tease out a few other things here. I, I guess one, what's the theoretical limits of the margins, on this business as your professional service business shrinks over time? And, and maybe two, just strategically, is that a business that you need to be in? Why, why wouldn't a Tier 1 just kind of integrate that into a automobile over a longer period of time?
The professional services?
Correct.
There will always be a certain amount of need for our professional services because they'll want some customization. I talked about this orchestration level. They'll want us to customize, like they've got their own parking app, you know, or their own charging stations, deals, like a lot of the European OEMs are. And they'll want you to not go through Google search first. They'll want you to come to their search, list theirs first, then Google. That kind of customization will always have some of that. There'll be some of it around, but it's not gonna be the, you know, it was, you know, $20 million a quarter at one point. It doesn't need to be that, whether it'll hover around $8-$10 or whatever, I'm not sure, but there'll be some around. What's our maximum?
Yeah. And if you think about, so our margins last year without the IP business, we were about, say, so call that all core 74% GAAP gross margins. This year if you strip out the IP, it's about 77%. So we are growing that. Part of that is that mix, the higher technology, less ASP. If you just look at our technology margins, they'll grow from last year about 77% to closer to 80% this year. I think we exit the year in even in this low 80s. So I think if you think the upper margin, we can get to 85, 80, 80. I was gonna say 83, but I think we could get to that 85% margin.
Somewhere there. 83- 85.
Either would be a fantastic outcome. That's great. Let's maybe then move on to the cash flow and balance sheet.
Mm-hmm.
You paid down a significant amount of debt last year or this year, really. You have cash on the balance sheet still, and it looks like you're expecting to generate close to $60 million of free cash next year. Remind the audience some of your debt maturities, how you expect to handle those maturities as they come due. And then maybe how do you think of the optimal capital structure and what kind of longer-term structural things might you wanna put in place over time here?
Yep. Yeah. So, in, we had two tranches of debt, the 2025 convertible debt, which we paid off $87 million last year from cash on hand or cash through operations. That leaves us with our 2028 convertible debt. It's 1.5% coupon rate is convertible, but the conversion is about $47. So it's really out of the money from a conversion standpoint. So it's good paper for us right now. That said, the fact that we have, you know, it will end the quarter over, you know, close to $100 million in cash, gonna produce the rest of the year. It gives us some flexibility to look at effectively a capital structure that is lower leveraged, and can we opportunistically buy some of the convertible debt? It's trading below par because of those parameters right now.
Could we pay some of that off earlier in the open market? Yes. And that would be our goal is to opportunistically pay down some of the debt ahead of that 2028 maturity, and provide a more simplistic, lower level, lower leverage balance sheet.
Makes a lot of sense. My last one here, Brian, I wanna give you, you know, a mic drop moment to follow with no follow-up from me. You've been at Cerence for a little over a year. You've accomplished a lot of things so far. In that year that you've been with the company, what's exceeded your expectations? What do you wanna message directly to investors from our chat today? And what are you most excited about over the next 12 months?
Yeah, those are all somewhat interrelated. I tell you for the exact same reason I came to this company, right? I was literally sitting on a beach and when the board called me and said, "Hey, would you come look at this?" there's a lot of really good IP in this company and some really good engineering teams, right? The majority of this company, there's roughly 1,300 employees, almost 800 of them are engineers, right? This is an engineering-focused company with a lot of IP. And what surprised me is probably just how good that IP is and how valuable it is. And the win with Samsung is just the first stepping stone of a long path, right? And what am I looking forward to?
I actually don't look forward to all those lawsuits and things, but what I really look forward to is what this technology can do. It really can become, when you get into your car, it's a pleasant, enjoyable experience. It greets you, "Hey, how's it going?" It knows you versus somebody else who has your key and doesn't necessarily do what you do then. So it's not key related. It actually sees your face. It talks to you. It talks about, "Hey, it's 8 o'clock. Are we going to the office?" and or are we, you know, "Would you like to stop at your usual Starbucks on the way?" It's gonna have that conversation. It's gonna do emergency vehicle detection. "Hey, I hear an ambulance behind us. Would you like me to pull over right now?
Or would you like me to help you pull over and understand that?" What it was that sign I just passed, right? It's gonna add all kinds of enjoyment and safety to your driving experience. As cars become more autonomous, this becomes even more important because the last thing you want to do is get on the freeway, get on an autonomous path and have to start touching things and all versus just talk to your vehicle and say, "Hey, I wanna get off at the next stop instead of the stop I programmed in. I wanna change the channel?" You know, "Hey, can you play this music and put up my last game or whatever," right?
It's gonna become more important for you to talk to your vehicle than touch it, as things become more autonomous. I look forward to that world, and what this thing can deliver.
Sounds like an exciting time. Brian, Tony, thanks so much for joining us.
Thank you. Appreciate it. Thank you guys for coming.