Hey, good afternoon, everybody. It's Jeff Osborne, the Mobility Technology Analyst at TD Cowen. Thanks for joining us for our second annual Sustainability Week. Very pleased to have Stefan Ortmanns, CEO of Cerence, as well as Rich Yerganian, SVP and IR at Cerence as well. Thanks, gentlemen, for taking time out of your busy day, especially Stefan joining us from Germany there in the evening. You know, maybe for those that are tuning in that aren't familiar with Cerence, there's certainly been a lot of news flow here with the stock and company structure, industry structure, technology evolution that we're gonna dig into.
But, for those at a high level that aren't familiar with the story, do you mind just giving us a brief overview in a minute or two of what it is that you folks do? And then we'll dive into particular questions as part of our fireside chat.
Okay. So, thank you so much, Jeff, for having us here today. It's a pleasure. And hello, everyone. Yeah, just let me introduce the company. So Cerence is a spin-off from Nuance in 2019, and we are a software company providing conversational AI solutions for the automotive segment. We have a quite high penetration rate of 54%, so it means every second newly produced car is using somehow technology or the full stack from us. We are working with every major OEM across the world, including Chinese OEMs, such as BYD. And I'm pretty sure there is a good chance you have used our products somehow. For example, just using voice for dialing, right? Dialing, calling mom, or for navigation purposes, or set the temperature. These are the typical simple commands in a car.
Now, we have a lot of expertise in this specific market, and our latest products, which we have introduced at our last CES, in Vegas, are all based on GenAI and large language model, which actually boosts the performance, its broader, deeper content, but also its real-time behavior. And I mean, with our knowledge about the car, having access to sensors, offering multi-seat solutions, right? We believe that we have set a new de facto standard in the automotive world. And beyond this, I think our own produced large language models are based on billions of data we have collected in the past in the automotive world, and we're combining those data with generic information.
Excellent. Maybe, you know, the most common question that I've historically gotten about you folks is sort of how you compete with, Big Tech, so to speak, the Googles and Apples of the world. So I'm just curious, can you walk us through sort of the industry dynamics, who the major players are, both within China and outside of China? Obviously, some of the tech companies aren't as germane in China. And then, you know, what you feel is the differentiating factor, maybe, as it relates to Cerence, relative to some of the other tech companies out there.
So maybe let me start outside of China, and to name it right, our key competitors are Google and Amazon. We or I strongly believe that we have some competitive advantages. I think we have all these long-term customer relationships. We have benchmark winning core technology. When saying core technology, it's about automatic speech recognition with large language models, right? And we are benchmarked more or less from all OEMs across the globe, and we are superior when it comes to automotive needs here. We are offering also to our OEMs and to our partners a customer-branded solution, meaning the customer owns the business logic, the arbitration logic, right? And it's really optimized and customized even for specific car lines of an OEM.
This is a bit is a big advantage, right? And secondly, the other big advantage is when it comes to our cloud solution, actually, our solution is a hybrid one, meaning we have solutions running in parallel on the head unit, so in the car and in the cloud. We're sharing also the cloud data with all our OEMs, and this gives them also some new opportunities for further data monetization in the future. Our solution, I would say, goes far beyond conversational AI. Just think about emotion detection, gesture detection, right? Biometrics, right? So overall, I see our solution as a multimodal, seamless solution with access to sensors, yeah, for offering also proactive AI. And then we have also capabilities just with the interior microphones to detect who's speaking in the car.
Is it the driver, the co-driver, or passengers in the back? And then, obviously, we are offering also arbitration solution. Some OEMs want to have multi digital assistants, for example, connecting with your home device. It could be Google, it could be Siri, whatever, right? And we're providing this solution to our OEMs. When looking to the Chinese market, this is a specific market. Here we are facing competitors such as Baidu, also other players like iFLYTEK or iSpeech. But also here we are working with key OEMs. Just to give you some examples, BYD, Great Wall Motor, SAIC, Geely, NIO, and especially we are offering them overseas solution. So also we're working more or less with all OEMs across the globe, including Korea and Japan.
The basis for this is our large language portfolio. We are offering around about 70, so seven, oh, languages.
Got it. Would you say that Chinese starting to export to non-Chinese markets, you know, could be a tailwind to you?
Yes.
A safe assumption?
Absolutely right, and we have also this broad OEM portfolio. So, for example, if the Chinese OEMs are coming to Europe here, right, and you know that we have excellent relationships with more or less all European OEMs, right? We are still very well positioned here.
Got it. Well, before we dive into the most recently reported quarter, which had a lot of fireworks in it, we'll come to that in a second. We have been getting a lot of phone calls, you know, from newer investors in the story, or people kicking the tires for the first time in a while. So on the surface of your income statement, maybe we'll just peel back the onion. From a high level, you've got the three revenue lines: license, connected services, and professionals. Can you just walk us through at a high level what each of those are, what they represent, and what the drivers are for each of them?
Our solution is hybrid, yeah, but working seamlessly together, and this has advantages when it comes to privacy, right? So the wake word is in the car or other aspects here. And the license revenue reflects royalties paid for embedding our software inside a head unit, yeah, for example, for controlling functions inside the car. This is a one-time license fee for the lifetime of the car, and therefore we take the full revenue when the license is applied at the time of production. Secondly, for our connected services, cloud services represents a revenue stream, which is associated with our cloud hosting platform. To give you also here a practical example, the driver is requesting real-time, for example, travel information, or real-time information such as weather or whatever, right? Traffic, yeah.
This is covered with our connected services, and obviously it goes far beyond this, restaurant reservation and so on and so forth, yeah? And here, when a subscription begins, we bill the full term of the subscription, collect the cash, yeah, and then the amount is added to our deferred revenue, which is then amortized over the service period. And you see also that we have, for this fiscal year, a positive cash flow, but slightly a negative EBITDA. This is changing now. Then pro services is actually also a key differentiator, especially, when comparing us with the Big Tech players here. We are offering engineering services to our customers for doing the complete integration and customization. Nowadays, with the emergence of Gen AI, solutions are getting more and more complex.
In the past, we acted more like a technology provider, but now we prefer to integrate and build the complete conversational solution for the OEMs, yeah? As the revenue and payment are typically built on milestones. Rich, anything you want to add?
No, I think that about covers it. So the three revenue streams are licensed business, which is the one-time license for the life of the car. That's the software that goes in the car. Connected services, which is cloud hosted, which is a subscription-based area, and then the professional services. You know, the nice thing about the connected services is, again, as Stefan said, we bill and collect the cash for the full amount of the period up front. So, you know, as time moves on here, as each quarter goes by, we're actually, you know, generating a greater than one CFFO to EBITDA.
Mm-hmm
... ratio.
Makes sense. Maybe just the last sort of, I don't wanna say remedial question, but just sort of higher level, is you've had a lot of moving parts within the license revenue line, in particular, between pro forma royalties, the consumption of royalties, fixed licenses. Sort of maybe just walk through what those are, what the strategy's been around licenses. I know each customer maybe is a bit unique in what they want or how they pay for it, but, how does that work?
So let me start also here with the license revenue, and we have two ways to report our licenses: so fixed or variable. So variable represents the number of car produced in the quarter in which a license is applied. Yeah. So now when looking at the fixed contracts, yeah, that's actually when a customer, and normally that's a first-tier supplier, mainly, mainly based in Asia, approaches us to buy a quantity of licenses upfront, yeah? And the consumption rate is typically 4-6 quarters, yeah? And then they pay us for this period, yeah, and for this, they will get a discount.
Got it. So you get the cash upfront, maybe they get a 10% or thereabouts discount, so-
Exactly. So in the past, looking back, yeah, so I mean, the fixed licenses was quite high. And Tom, the former CFO, and I decided two years ago to reduce the number of fixed-
Right
license deals, which is very essential, because I don't want to give them a lot of discounts here, right? So, and now, for this fiscal year, we're in a range of $30 million, and, for next fiscal year, the plan is, in the range of $20 million. This requires some discipline, obviously, right? Because, I mean, that's a typical approach of Japanese, Korean first-tier suppliers doing this kind of purchasing here. But we're sticking now to this, reduced, fixed license deals.
Makes sense.
One of the benefits, Jeff, of the reduced fixed licenses is that, you know, what has to happen, right, on a quarterly basis, is whatever fixed contracts had been executed in the past, now that inventory has to be consumed. And so when that fixed contract is expired, meaning all of those licenses are used, then it goes back into the variable model, right? And so next year or this year, you know, we're expecting to consume a certain amount of that inventory. Next fiscal year, we're expecting it to be about $20 million or so lower, you know, $15 million-$20 million or so lower than this year. And that should translate into the variable license revenue that we report on a quarterly basis.
The pro forma royalty line that you referenced is kinda like an operational metric. What that represents is, had we never done a fixed contract, that's the amount of license revenue we would've reported in any particular quarter.
Got it. No, definitely there's a lot of metrics to try to peel back the onion on the call, which is helpful. Maybe just getting into the quarter itself, the headline numbers were actually pretty good. The problem was the forward-looking guidance and the multi-year plan, you know, was withdrawn, so to speak. And so I'm just curious, can you give us a sense of what's going on, either within Cerence or the broader auto industry? But how did we sorta get to the point that we are, that you've decided to withdraw the guidance, completely?
Mm-hmm. Yeah, when receiving the Q1 royalty reports, we saw some concerning trends here, right? Then we conducted deep account-by-account reviews, which commenced during Q2 and finished in April. We saw a couple of areas to watch, which actually led to a reforecasting of our FY 2024 guidance. I think there was an internal review of forecast projection based on latest data. We are typically collecting from the OEM and from IHS, and taking into consideration historical trends. Yeah. Our penetration was still high, it remains at 54%, yeah? Also in our discussion, we saw actually a downward trend, especially for European customers. You know, there are actually three key trends in the automotive world. EV is also slowing down.
Then we saw also, I mean, software-defined car, which is actually increasing the complexity of software in a car to a new dimension here. And this led also to delays in SOP, where we have actually a much, much higher PPU, so price per unit or value per car. So and then we saw also some slower ramp for newly SOP'd solutions here. And the other aspect is that we saw also a slower ramp in adjacent markets, so two-wheelers and trucks. So overall, we signed nine deals with two-wheeler manufacturers across the globe. Six went live, but still the revenue was below our expectations, yeah? And similarly, we saw a similar trend in non-transportation, which we call AIoT.
So everybody needs to run out and buy their Harley-Davidson, I guess, but, anyways.
Yeah, I think, Jeff, just to summarize, I think unfortunately it was a perfect storm of various contributing factors that led to that.
Right.
Unfortunately, that overshadowed the progress that we're making, and rightfully so, but the progress that we're making with the newer products that we introduced at CES and, you know, the momentum that we have with customers in terms of that technology.
Yeah, I definitely wanna touch on the six wins from CES, but maybe just to touch, keep on the financials theme. You referenced sort of a run rate revenue, you know, outlook for 2025 without sort of giving guidance, but just pointing the analysts in a direction of travel, so to speak. So I'm just curious, you know, how did you come up with the run rate, or what do you mean by that, the term run rate, as we think about a framework for 2025 more broadly?
So, Rich, do you want to take this first?
Yeah. Yeah, I'll take it. So, you know, what we're referring to there is, you know, as you, as you may recall, or for those that aren't familiar, in our Q1, you know, we had about $130 million in revenue for the quarter. The reason it was so high was because a customer of ours, that dates back to 2000, a program that dates back to 2013, was terminated by the end customer, in this case, it's pretty well known it was Toyota, in the December quarter. So we had to accelerate any remaining deferred revenue from that particular program. And, you know, that was a non-cash revenue. Most of that cash was collected by Nuance prior to the spin-out. But anyways, we had to accelerate that revenue. So what we...
rate is, if you take the, you know, the midpoint for where we think the full year revenue is gonna end up, and that includes that accelerated revenue, and if you strip that out just to get an apples to apples moving forward, meaning representing the business as we move forward in time, then you get to a run rate for the total revenues for the company, in that $240-ish million range. And that's what we wanted to, you know, make sure people understood, because that legacy revenue was no longer gonna be part of the mix. So we wanted to make sure people understood what the baseline of the business is moving forward.
And then what we did was we said, "Yes, we think, you know, middle single digit framework for potential growth next year." And a big part of that growth-- And that was assuming everything else in the business is flat, so we're not assuming any growth in auto production or anything like that, that you know, we should see that growth, if for no other reason, then the consumption level of those fixed contracts will be, you know, lower by a fair amount next year, and that translates into the variable revenue for our licensed business.
Got it. That's helpful, Rich. Maybe the last financials question for now is, you, you mentioned or alluded to on the call about taking, you know, cost out of the business, you know, given the macro of what's going on. And then, you know, you didn't really quantify anything. I don't know if that was just something you sort of came up with right before the earnings call, or if that was something that just requires a lot more study. You know, certainly there's been a lot of moving parts post the spin, and then the quick move to AI, that you highlighted, which I'll... I, I wanna come to next in the line of questioning.
I'm just trying to understand, you know, when will investors have a better appreciation of how much potential costs can be taken out of the business, and sort of what the go-forward EBITDA and cash flow dynamics may or may not be?
So, first of all, cash flow and positive EBITDA are important for us, and we are working now on a cost reduction initiative. We brought on an external firm supporting us here. They bring in a lot of experience. But we are in a very early stage here and are not yet prepared to share any finalized plans for cost measures or final numbers for cost savings. Nevertheless, I think it's important to right-size the business, yeah, while still delivering on our Gen AI roadmap and also on our customer commitments. That's very important. And secondly, as already Rich referred to, CES was a big success for the company. We're getting great feedback.
We're having lots of opportunities, and it seems also that OEMs now see the benefit of large language model optimized for the automotive segment here, and we need to convert these opportunities. Yeah.
Got it. Perfect. Maybe let's then switch to shift gears to the tech side, which I know, Stefan, it's more of a passion of yours, I imagine, than talking about cost reduction. You know, the CES show what was, I felt, compelling. I think you were at a disadvantage, just given that internet connectivity was a challenge, so some investors certainly had some concerns around latency of some of the demos. But, you know, it's great to now hear what, 5, 6 months later, that you've had progress. I think it was 6 OEMs, ha-
Mm-hmm
... have agreed to buy the product. And what was it? 14 or so, 11? I forget the number, but it was a compelling number that you were working with. So maybe just walk through sort of at a high level, what are the incremental advancements that you've made and, you know, what's truly differentiated around that product offering?
Mm-hmm. Mm-hmm. So, you're absolutely right, Jeff. At CES, or the day before, we had this big event with the VW Group, where we, for the first time, launched with a big OEM, a new solution called Chat Pro. We have integrated ChatGPT in a very efficient way and fully adjusted to the automotive needs. So the combination of general knowledge with navigation knowledge or in-car information, right? And all the logic, right? The user interface or the conversation user interface is controlled by the VW-branded digital assistant. And at CES, we have presented also various new models and solutions. Actually, we are driving to a full stack AI computing platform, where we bring in all the different large language models, right?
But also embedded and cloud, which gives us another advantage over the Big Tech giants here, right, for being deeply integrated in a car. And the success afterwards, after 5 months, we have signed 6 OEMs, and 3 will go live by end of this month, beginning of June. That's a big success here also. And you see also the speed here, right? Also various aspects here. Normally, for contracting or for doing a deal, a larger deal with an OEM, it takes you 12, sometimes 24 months. You already start with this pre-development, but it takes time. And now we have closed those deals within the first 5 months, and on the other hand, you see also the speed, because we're now in the new AI era.
Things can be easily integrated in the platforms within just a couple of weeks. To give you also some concrete data, the deployment development time for VW took us less than three months. On top, of course, you know, for in the automotive field, there's a lot of QA, cycle. That was another three months. So but overall, you see that's, that's a different speed, and also solutions are really impressive here, right? And this gives us a lot of advantages also for winning back opportunities which we have lost a couple of years ago. I think currently the focus, and we have two tracks actually. One is related to the integration of ChatGPT or other third-party large language models, controlled by our own large language model.
The next one is also, you will see also in the near term, a new Cerence Assistant, which is absolutely flexible. You can say whatever you like, and you will get the right answer, also with a fast real-time behavior. And of course, the third one is also adding more application to it, and one application obviously is a kind of car health or car manual. Now, in parallel, we're driving also our new next Gen AI platform, that's novel here. I cannot share too much about this, but we have also three OEMs engaged, where we're doing this pre-development, and in total, we have 14 pre-development programs in play with OEMs across the globe, where we need to actually close the deal, what I said, beforehand, yeah?
In total, we have now around about 37 proof of concepts in play. That's huge, right? You see now for us, I mean, on one hand, we are saying, "Okay, we need to reduce costs. We need to rightsize the business," which we are going to do. On the other hand, we need to deliver what we have committed to our OEMs, you know? We are very optimistic that we can achieve this ambitious goal.
Were any of those six wins win-backs at all or from Big Tech?
So there was, we had a win back earlier. It was actually end of last fiscal year, and it's also proven, right, that they want to expand our relationship and that we are working now on these expansions with these new large language model solutions here. But overall, I think we are in a very competitive situation now, better than ever, in my view. We are working also with OEMs, where we have lost a deal, a program, not the complete relationship with an OEM a couple of years ago.
Well, perfect. We're running out of time here, but maybe just as a concluding thought, obviously, been a lot of turbulence in the stock here. I'm just curious, as you look forward with all these new exciting things going on, is it really the connected line that investors should be focused on for these initiatives to show up? And, you know, when do you think it would start manifesting itself or be visible to the outside world?
So all of the new solutions are cloud-centric. So it's all about the cloud, but they still have certain embedded aspects. And what are actually the key revenue driver for us, I think there are three pillars. First of all, it's penetration. We have still a strong penetration rate, right? But there is still some upside for connected. Then secondly, growth in PPU, price per unit, and with the new AI computing platform, with new features, and we are offering a complete solution, right? We see also an uptick in price per unit, and third, obviously, is growth in connected services.
Yeah, and Jeff, the nice thing about these new, new products as well, as they hit the market, is they are subscription-based, which means we bill and collect the cash upfront for the subscription period. So that goes into deferred revenue, and the revenue actually comes last. So what you should see is hopefully growth in the billings and the deferred revenue line, and that will be an indicator on how this is progressing.
Since we own also the platform, which is currently on the road for most of the OEMs, right, we can also address already SOP cars. That's also new, yeah.
Well, perfect. I'd love to continue the conversation, but unfortunately, we're out of time. So, thanks so much for joining us from Germany, Stefan, and Rich, from Boston, thanks so much.
Thanks, Jeff. Thanks for having us.