Cerence Inc. (CRNC) Earnings Call Transcript & Summary

Good afternoon, everyone. I'm Vineet Chhangani, Morgan Stanley, Tech Banking. Today, we're here with Stefan, CEO, Cerence; and Tom, CFO, Cerence. Thank you both for joining us.

Maybe just to kick things off, Stefan, why don't we talk a little bit about what is Cerence about? What's the genesis of the company? Maybe give the audience a little bit on the company and help them understand?

Thanks for having us here, Vineet. Good morning again. Yes, Cerence is a voice AI company spin-off from Nuance in 2019. Our focus is on conversational AI interaction, especially in the transportation market, cars, trucks, 2 wheelers. Literally, we're working with more or less all OEMs across the globe here in the United States, in Europe, in Japan, China, actually everywhere. We have a huge penetration rate of 54%. We see also a strong adoption to our new technologies. Monthly active users are growing year-over-year by 30-plus percent, which is phenomenal. Yes. And now with Generative AI, large language models will become the new AI agent also in cars. That means we have developed one conversational AI interface across all applications for completing tasks based on user preferences, so drivers, co-drivers and passengers in the car. Nevertheless, it's much more than large language models. We have also produced an AI computing platform for our partners, especially in the automotive world. Just think about multi-seat intelligence, emotional AI and multi-modal interactions here. And this platform is also independent of the market segment. You can also apply to any other opportunity in the market, for example AIoT.

And just -- it took us less than a minute to start on the topic of AI, and that's been the theme for our conference this year. You obviously announced the partnership with NVIDIA. You launched a product, CaLLM. Maybe double-click on that a little bit on how do you see this LLM working, particularly with your customers and what your product offering is about?

So NVIDIA is a very important partner for us. Just think, especially in the automotive world, you have to produce maybe large language models for roughly 30 languages. Our approach is hybrid, so solutions running in the car and in the cloud. With the support of NVIDIA we have full access to the infrastructure, software and also hardware. Here, we can produce our OEM-specific branded solutions, large language models in record time. We have also access to NVIDIA engineers, which is also important to optimize our solution. And we own also a huge data set in the automotive world over 60 billion token. This has a huge value for creating this OEM-specific models.

And Stefan, just -- because we hear a lot about AI, I want to know a little bit about just the vertical that you're in right now to your comment about data. How does that differentiate you particularly as -- everybody is talking about AI, but there is something unique about your product and the platform?

So at CES, we have recemented our innovation leadership, especially in the transportation space. Maybe you have heard also about our joint press event with VW Group where they have introduced our first AI solution, where we brought together the best of 2 worlds, namely ChatGPT and VW's so-called IDA in-car assistant. It's not just bringing in ChatGPT, it's much more to it because we are enabling also OEMs to feed the system with their own information, right? For example, you can ask in a VW car what is the best VW car, maybe the ID.7 or brand, it's not a GM or Ford or BMW, right? So the OEM has full control about it. And equally important, the business logic is owned by the OEM, by the in-car assistant. Yes. And this gave us a tremendous opportunity also with other OEMs. We had also -- already right after this announcement with SKODA and [ Fiat ], you will see more from the VW Group over the next couple of months. But they will be also followed by other OEMs across the globe or even in China, right? In China, you cannot leverage ChatGPT. You have to deal with other third-party large language models. Nevertheless, for also reducing costs, which is very important to OEMs and also for having this high response time behavior, so a very snappy solution. We are leveraging actually also our large language models, right, so that we can also have this kind of answers in real time. And the beauty of the solution is, for example, if you say I need a coffee as a driver, the system has all of your preferences, right, and then it's directly connected to the navigation GPS system, right, and brings it to your preferred coffee shop. Yes.

And Stefan, as we think about what the future may look like, and obviously, you're in the best seat here, no pun intended, with the AI capabilities you have. Help us understand like, what does that look like because you have the front seat, on what OEMs are thinking about using your technology? What does that road map look like with AI?

So first of all, of our underlying core technologies are AI-based made on [ Euronet ] and other stuff, right? I think we are a global -- I don't think we are a global player in the market. We offer more than 70 languages, and we own all technologies underneath, from audio AI, yes, just starting for hands-free communication, but also for our multi-seat opportunities, right, private calls in the car up to text to speech to automatic speech recognition and, of course, natural language understanding. On top, we have also multi-modal capabilities, right? The interaction with the system, with GPS data, with sensor data, we call it Proactive AI, and we combining everything together. In the past, Cerence was considered more or less as a component provider, but now we are an innovation leader here, and we are providing the complete solution. Therefore, we have created also a new assistant called Cerence Assistant with NLU Plus, and here, you have actually unlimited opportunities. One of our first customers, for example, is BYD. You have [ multi-intent ] opportunities. You can say in one short, "Open the window." "I want to listen to Madonna", and whatever, yes, in one utterance, right? You have also a lot of opportunities for doing information retrieval. Yes. You can just say, "Okay, what about this?" "I'll tell you about this." You will see a one line, say, "Oh, what do I have to do now." Right? These are really cool features, right. And compare it to solutions built in 5 years ago or even 10 years ago, it was more or less [ command ] control. Now you have the full flexibility for natural language understanding, really a natural interaction with users, drivers and co-drivers, right? And what I said also, we launched also system with Great Wall Motors, saying, okay, we have multi-seat experience, right? People can talk concurrently to the system and will get their response. One can say in the back, "I would like to see a movie from Netflix", blah, blah, blah. The others say, "Okay, make a phone call", right? And you have also the [ isolation ] with multi-bubble [ songs ] in the car. It's all produced by us. So it's much more than just providing a conversational AI solution.

Yes. And I think for us, what -- this new kind of what we call the immersive cabin experience with all of this technology and then being able to do that with our technologies and our capabilities, but branded by the OEM. And the other differentiator that we have is the OEM owns that data. And now through our solutions can also design how they want that to work. And that's all integrated into one platform. And not only I think have we moved from being kind of a component supplier to be this integrated solutions provider around this immersive cabin, I think what the team saw in CES was it flipping a little bit to where the customers are pulling on us. You can get at this technology where -- I think for a long time, I've been involved with this business for 15 years, when it was part of Nuance. We had to go push this. We had to go sell it. We had to get people to implement this. And what we saw at CES, the press release with Volkswagen, which was in the car and working, right? So it wasn't a test. We had a bunch of people come in and try and trick it. They thought it was just a demo and it just worked perfectly. And then what happened in the rest of the show over the next 3 or 4 days is OEM started coming to the booth and saying, "I want this." "How do I get this", right, which has led to like this really huge expanded pipeline that our sales force is off trying to drive the drive. The other thing from a financial -- just so you understand the financials. So Stefan talked about embedded, that's a onetime license for the car. So that revenue is taken upfront. And that's where we have the 54% penetration. So some piece of our technology is in 54% on a trailing 12 months of every car produced. On top of that 54%, we add -- about 21% of those get our hybrid solution, which includes the cloud. That's a service agreement with the OEM. We got the money upfront, which is great, and then we amortize the revenue over the service period. And then we have a small professional services group which allows us to create that differentiated experience for the OEM, which is very different than what the big tech companies will do. So I think we're at a really cool inflection point here.

And when looking also at the gross margins for embedded, it's in the mid-90s, right, for Connected Services, mid-70s, and even for Professional Services, mid-20s.

And so maybe, Stefan and Tom, as I think about this, maybe one way to put this is vertical software platform that you have. And to your comment about the data and some of the big tech competing, the fact that you're playing independent is a huge advantage, which is why all the OEMs want to come to you. And as we think about AI in the future, it seems like you're very well positioned for that. Is that fair to say?

That's a fair statement for it, and that was also actually the outcome of CES, right? And we have a huge post-CES momentum here. Coming back to the VW where we deployed the solution from concept to product maturity with -- in less than 3 months, that's also kind of new innovation, speed, right, AI innovation speed, right? VW to the [ CEO ], they're very pleased with the solution, right, and we see also a huge appetite from other OEMs, right? And development times in the past was roughly 12 to 24 months. China is always a bit faster. But now it's in the range of 3 to 6 months, right? And this gives us new opportunities. Also when considering those cloud solutions, right, so we have now also the opportunity for upgrading cars on their own, so a platform that can already [SOP'ed].

Just kind of building up on that a little bit, your momentum. In the recent call -- earnings call, you talked about the growth in China with Chinese OEMs. Maybe talk to us a little bit about what you're seeing in different geographies and where you're seeing the most traction?

So what I see, and I'm spending a lot of time in China, companies like BYD, Great Wall Motors, Nio, Li Auto, Geely or traditional OEMs like SAIC, putting a lot of attention to the in-cabin experience and aforementioned so-called immersive in-cabin experience, right? And here, we are well positioned. And we can also learn quite a lot from those OEMs because their speed in terms of from concept to deployment is tremendous. Yes. I was also 2 weeks ago in Japan, meetings with Toyota, Subaru, Mazda, Honda, right? Also here, Chinese OEMs were normally always very traditional, cultural driven. But also here, they have a new speed for bringing in this new in-cabin experience. And it's independent, but it's a typical combustion engine or EV, right? So we're not affected by those elements here, which is great for us. So I think there's a huge appetite for adapting to our solutions, right, especially when it comes to this multi-seat intelligence where also passengers can interact with the system.

Maybe switching gears a little bit. Talk to us about Cerence's FOU agreement that's coming up for expiration soon. Just for the benefit for the audience, the spin-out of Nuance had you, I wouldn't call it restricted but focused on automotive. What does that mean? What are some of the opportunities in AIoT that you could potentially explore?

Maybe let me start first, and then I will hand over to Tom because he did all the transition when he was one of the transition Officer at Nuance here. Yes, we had some FOU constraints, I think, for the next 2 quarters. So we are also very excited about new opportunities, generating new revenue avenues for the company here. We're already also quite successful here. Last quarter, we launched 4 products with companies like Motorola, Zebra, also in China and in Korea. We see actually 4 areas where we can also bring a lot of value to customers, industrials, wearables, smart home and also in the health, well-being segment here. Currently, our focus is on typical voice control on audio processing, voice biometrics, but also information retrieval. Just think about a car manual. You also easily apply it to a device manual, right, for example, for medical applications, so on, right? And we're seeing also here a strong demand for some of the key players here across the globe, especially here in North America.

Yes. I mean, just clarity, the field of use restriction worked both ways, right? So Nuance, Microsoft couldn't compete in the transportation space. But just so you know, we didn't violate the FOU and stuff like that. We launched some products in the last couple of quarters. There were certain technologies that were developed in Cerence that weren't subject to the field of use. So we've got to launch this a little bit in some defined areas to get this going. So we put a small technology team that really takes our core technology, but that [indiscernible] "miniaturize" it or get it to a point where it can go into a smaller device. We've had a couple of business development people. So we got to kind of launch this early. And then at the end of September when the field of use expires, then kind of the worlds are [ oyster ], and we can expand with all of our technologies into some of the areas that Stefan talked about. We're not really worried about the reverse coming, that Microsoft's all of a sudden going to come in and swoop in and take the automotive business the same way as we're not going to swoop in and take their core healthcare business, because both of these businesses require really deep understanding and lots of data and language data. They have that on healthcare. We have that in transportation, which is why we can leverage it in these large language model solutions for the OEMs. They don't have access to that post the FOU. They didn't have to go collect all of that. So I think it's a great opportunity for us after the field of use to grow that segment of the business.

Maybe talking a little bit about how you've changed the management team? Recently, you've added new members to the senior management team, including new CTO, new CRO. Talk to us about the additions and what's the recent takeaway from how they're thinking about the company going forward?

Yes. So actually, I feel very good about the management team now, what we did over the last 1.5 years. We brought in Nils Schanz as Product Officer from Mercedes. He was responsible for their MBUX solution. So for the digital experience at Mercedes, then we brought in Iqbal Arshad. He was prior to Cerence at Google, Lenovo, Motorola, [ so as much as ] automotive and he has a great [ reckon ] when it comes to AI. He is driving the CTO organization, and he is driving all core technologies, including the new voice [indiscernible] solution, which is based on a large multi-modal actually, which is very action-oriented, which is actually needed if you're talking to device or working with devices here. And then we have also Christian Mentz. He started in October. He is from Amazon and that Amazon Automotive or the smart vehicle. He was running sales, business development and professional services, and now he is running sales for us and marketing, and he is doing a terrific job for Cerence. And he is also well connected across various OEMs, which is actually needed here. And we see already the results from this new business team in record time, right? So a proof point for us was CES, and we got really extremely positive feedback, not just from OEMs but also from investor, analysts and also from journalists here. And then we had also another -- unfortunately, I mean, Tom wants finally to retire. He looks like 50 years.

I'm not.

He jumped in when we had a need for having a really strong CFO. It's a pleasure working with Tom, right, and we have also some good links to Dan Tempesta, who was the former Nuance CEO. He just finished his work at Microsoft Word, and we saw that was a great opportunity if he can step in and support us. I was working with Dan when I was running the Automotive division at Nuance, right. [ Rather let ] Tom [ speak as he ] has much longer relationship with Dan.

Yes. I mean I've been involved with the auto business and Cerence since 2008. As Dan has -- Dan joined Nuance 3 or 4 weeks before I did. And I led the spin in 2019 and joined the Board and then just about 2 years ago stepped in to support the team as the CFO. And I will reiterate what Stefan said. This is the strongest operational leadership team we've had, I think, ever in the company. And not only that, they're working extremely well together, which is very important. And then I had said all along when I jumped in a couple of years ago that I wasn't going to leave until I knew we had a really solid transition plan. And as Stefan said, very lucky to attract Dan. He had lots of opportunities. But I've been talking to him for a while, trying to convince him that this was the best opportunity for him. And so he'll join us on the 18th of the month. And I'll stay for a very short transition just to help him a little, and then I'll continue on the Board. So I don't think we could have had a better transition, at least in my specs on the finance side.

No, that's great. And thank you, Tom, for doing what you've done. And obviously, you'll continue to stay engaged with Cerence as part of the Board. So we're at an inflection point. You have a great team. Curious to hear your thoughts on, as the teams around AI have picked up, as you look across the landscape, what are you seeing in terms of new competitors coming? We talked a little bit about the large tech. How do you feel you're positioned when you look across the landscape in terms of competition?

Yes. I think the competition is still the same. When looking backward, it's a Big Tech with Google and Amazon. In China, obviously, it's a bit more Baidu and other players. Overall, I'm really proud what the team has delivered over the last couple of months. And I think with our change also in our long-term roadmap strategy to focus now on conversational AI is paying off. That was also the feedback from CES from large OEMs. They want to work with us. We know also that OEMs selected Big Tech a couple of years ago. I'm not very satisfied in terms of branding, for example, in terms also of bringing the right solution into the car, depending also on the brand experience here. And this gives us now huge new opportunities and big opportunities, strategic design wins. And I'm really happy to the next couple of quarters where we see also first results in winning back those programs.

Stefan, maybe just talking a little bit about competition, going a little bit deeper. So obviously, your numbers speak for themselves and the penetration that you've had. Tell us a little bit about when you go and replace one of these large techs for the OEMs, what are like the top 2 or 3 reasons why people are choosing Cerence over, call it, Microsoft or...?

It's about quality. It's about flexibility. It's about who owns the business logic and also equally important who owns the data.

And data being the main piece in terms of neutrality?

Exactly. We also think about future opportunities for the OEMs, right? They're struggling a bit also in leveraging the data -- connected cloud data, right? How can they also create a better data monetization for themselves, right? We just announced also with Renault with a new 5 EV car at Geneva Auto Show, a new solution where we are supporting the avatar, right, and it's a groundbreaking technology, right? I mean, we're talking to another client here, it's real. And this is based on our solution also -- and Mercedes also announced at CES a new MUX solution also with emotional intelligence and this all is coming from us here, right? So we are really grounded here and the OEMs. We are doing also roadmap alignments with key OEMs across the globe, right, and this drives new innovation. And we're spending quite a lot in innovation.

I think a lot of it is to do with execution, right? I mean we've made tremendous strides, I think, under Nils who runs our Professional Services organization along with the Product teams. And then what we hear is that some of the big techs have kind of overpromised and underdelivered. And then, of course, all the other elements that Stefan talked about. So we have had some win-backs and I think we will have a bigger opportunity for win-backs kind of going forward.

Maybe Tom, to you a little bit on the financials. Obviously, there have been a few changes in terms of how you're helping investors understand the business model. Recently, you talked about the Legacy contract with Toyota. Maybe give us a few highlights of like how the business model has changed in terms of like how you're helping investors understand how you track your targets? Just some visibility on that will be helpful.

Sure. Under the prior CEO and during COVID and the pandemic, the company kind of did a much higher level of what we call fixed contracts, and there's 2 types of fixed contracts. There's pre-pays in which case -- these are only Tier 1s, and they're mostly in Asia. On a shipping platform will come to us and say, "Hey, I want to buy an inventory of licenses usually 4, 6, 7 quarters. We want a discount, but we'll pay you upfront." And this has better contracting going back all the way to the early days of auto. It's kind of relationship. It's driven really by the procurement in those Tier 1s. And then there were minimum commitment deals. So under some of the new accounting rules, if you have a commitment, you take the revenue upfront, but then you have to bleed through all of that. They're a different structure and that -- the revenue gets taken upfront, but the cash is collected as the cars are shipped. So therefore, it's a smaller discount. Stefan and I stopped doing those when I joined as the CFO. I just didn't think there was the right economic value for those. So we haven't done one almost 2 years now, and we won't do anymore. But we are still bleeding through those. On pre-pays, we made a commitment last year to only do $40 million of pre-pays. We did $36 million, so we hit our objectives. We said this year we would do in the $20 million for this year, so half, because we think our position is stronger, because most of these decisions these days get made by the OEMs themselves, but they'll still contract through the Tier 1s. So we hope to bring that down. So why is that important? Well, it makes us more of a running royalty or our Connected Services coming in through the ratable services programs that you [ won ]. And it doesn't create all this crazy lumpiness. So I think that was a big thing. And then the other thing that happened is, and we talked about, there was a Legacy contract that goes back all the way to 2017 and an acquisition that Nuance made in the auto space, that this company had sold a solution to Toyota that had a very, very high price and have had a 7-year service period. At the time of the spin, Toyota was shipping very few cars with that technology on them, but had built up this huge deferred revenue base on all the cars that had shipped prior to the spin. And we have been amortizing that deferred revenue off since the time of the spin. And it's very predictable because there's no new cars going into it. And it had a fixed termination on the service, which was 12/31/25, our fiscal Q1. And we were going -- we've been amortizing about $8.5 million a quarter off of that, very predictable, very transparent. We've told everybody about it. In Q1, Toyota informed us that they were shutting that service off because they had very little usage. Nobody was really using it. They still had to keep it going. So there was a cost element to them. Because they terminated that and weren't going to use that service anymore, we had to accelerate all the remaining deferred revenue, which is about $76 million. I'm happy though, because, again, it cleans up the financials on a go-forward basis because now you don't have all this revenue, it doesn't have any cash associated with it. I don't have to explain it every day. And then the other thing we've done is we've kind of increased transparency around billings. We've instituted a 5-year backlog model to help people really understand, are we driving towards our multiyear plan and objectives? So not only do I think we're in this like awesome position from a technology and a customer standpoint, but kind of on a go-forward basis, the financials are much easier to understand, and we'll really start to show the real strong trending of the business. And this really awesome model that we have, right, that you get built on a platform, [ we ] don't really get ripped off, right? And so it's not a pure [ ARR ] model, but in some cases, it's even better, right, because it's just -- they're shipping cars on these platforms for years, right? And now with the ability to update the Connected Services and then keep adding services on top of that. And then if this ability to attack and put this on, on the road vehicles, it's a whole new opportunity for us. And we've talked about it for a few years, but the OEMs control that relationship. So we're 100% dependent upon them being value in that. And I think it's another proof point that these products are so good that they actually want to -- they know how to make an investment, right, because they'll have to pay us and then they'll put them on those cars. And then hopefully, those customers will see value in it. And that can potentially drive what they're really trying to get to the OEMs is the subscription monetization opportunity that they all want.

And Tom, just maybe expanding on this a little bit more. As you think about the opportunities going forward, particularly around pricing, some of your AI offerings, how do you envision that getting rolled out in terms of is it pricing per module? How would you go about pricing those?

Well, again, as I said earlier, I think it's nice that we're starting to get more of a pull from the OEMs. So there's -- obviously, they're seeing real value here. That should give us some pricing opportunity and -- on both the embedded and the connected side and then the ability to take that 20-ish percent attach on top of embedded, that's a big opportunity because then you get the embedded, you get the additional cloud-based services on top of that. And then to the point you can do upgrades to that, then the whole value chain goes up. And where -- people ask us all the time, clearly, auto production matters to us. But what really drives this business is this innovative, exciting technology that the OEMs really now see how they can leverage, that drives up the value per car that we're receiving.

Tom, one more for you. As you think about product development -- and so 2 questions there. Help us understand how do you think about the R&D spend going forward? And then the second part was, as you lay out more and more software offerings, any impact we should anticipate on gross margins?

Yes. So we continue to invest in R&D at very high percentages of revenue as we went through this revenue transition the last couple of years. Stefan and I thought it was critically important because it's the future of the company. So, we did take a hit to earnings and to cash flows. But clearly, we wouldn't be in the position that we're in and what we've announced if we hadn't done that. Iqbal's come in. He's added some new processes and structure to the whole R&D and product development cycle. We did some -- he's done a reshuffling of resources around. We've done some minor kind of restructuring activities. But we have a really strong global team. Clearly, there's an appetite always in R&D for more and more spending, but we're aligned with Iqbal. And as the revenue grows, there should be leverage there. And then, of course, we now have the advantage of leverage with NVIDIA, right, because we have to leverage their R&D resources, their tech stack, their capabilities, right? This is a partnership that benefits both. So there's a lot of money going back and forth between the 2 companies. We're going to make the money collectively in each of our companies. And then Stefan alluded earlier that our license business is in the mid-90s. I don't think that changes much. Connected, you'll see it's going to dip a little because of all the -- it's going to go way up this year because of the accelerated Toyota, or it's going to dip down a little bit in '25. But we believe steady state going forward, we're back in the mid-70s, which is where we've always been. The model is that the small amount of that has to go to ChatGPT. We'll get paid for by the OEMs and we'll go through our P&L and then as Stefan said, PS is about 25%. So that's how we believe it's going to play out here, and that's what we're driving.

Maybe I'll pause there to see if any questions before we move on. No. Okay. Stefan, maybe talk to us a little bit about EVs and if any impact at all on how you think about your end market?

So 2 years ago, right, there was this huge hype about EVs, right, also about creating new in-cabin experiences, right? What we see here, most of the traditional OEMs have adapted. I mean they're driving now the new in-cabin experience, what you have seen, for example, with VW traditional large volume manufacturer, which is actually great for us. We are working also with -- as I said at the beginning, with the largest EV maker with BYD. To give you also some ideas about our new speed here, the new Cerence Assistant for [ Mandarin ] could be deployed within 2 weeks. That's in record time, right? And this is really needed for this new market here. Overall, I think also with our new AI roadmap right, we are on a very good path. I mean, auto production is still a watch area for us but nevertheless with this new OTA, over-the-air update opportunities, right, we are well positioned for future growth, especially in the connected fields, right? And we're discussing also with OEMs new business models, right, subscription model, right. You have now endless opportunities. And I have to say also the quality of the new in-car system, right, advanced with ChatGPT or other third-party language models, right, gives a lot of new opportunities also for the OEMs compared to the older in-car system.

So is it safe to say no impact on you -- how you go to the market?

Yes.

And if anything, it's...

The only impact I see if you have a time, when had COVID, when we had actually 90 million newly produced cars on annual basis, right, trending towards 100 million, and in COVID it went down to 70 million. And now it's in the low [ 80s ], right? If you have another supply chain issue or another war in Ukraine or so on, this could hurt us. But nevertheless, I think we are well positioned. We have also a very diversified OEM network. You know there are already some battle ongoing in Europe, right, the German, French, Italian OEMs versus the Chinese OEMs, right? But we're working with all of them together, which is good for us.

That's amazing. And once again, thank you both for taking the time. I know we're up on time, but we appreciate you both joining us.

Well, thank all of you at the very end of this conference to come and join us instead of running home. Thank you.

Thank you. Thanks, again. Thank you.