CRNC - NASDAQ

Good day, and thank you for standing by. Welcome to the Cerence Third Quarter 2025 Earnings Call. [Operator Instructions] Please be advised that today's conference is being recorded. I'd now like to hand the conference over to Kate Hickman, Vice President of Corporate Communications and Investor Relations. Please go ahead.

Hello, everyone, and welcome to Cerence's Third Quarter 2025 Conference Call. I'm Kate Hickman, VP of Corporate Communications and Investor Relations. Before we begin, I would like to remind you that this call may involve certain forward-looking statements. Any statements that are not statements of historical fact, including statements related to our expectations, anticipations, intentions, estimates, assumptions, beliefs, outlook, strategies, goals, objectives, targets and plans are forward-looking statements. Cerence makes no representations to update those statements after today. These statements are subject to risks and uncertainties, which may cause actual results to differ materially from such statements and expectations as described in our SEC filings, including the Form 8-K with the press release preceding today's call, our most recent Form 10-Q and our Form 10-K filed on November 25, 2024. In addition, the company may refer to certain non-GAAP measures, key performance indicators and pro forma financial information during this call. Please refer to today's press release for further details of the definitions, limitations and uses of those measures and reconciliations of non-GAAP measures to the closest GAAP equivalent. The press release is available in the Investors section of our website. Joining me on today's call are Brian Krzanich, CEO; and Tony Rodriquez, CFO. Please note that slides with further context are available in the Investors section of our website. Before handing the call over to Brian, I would like to mention that we will be participating in the Raymond James Industrial and Energy Showcase on August 13 and 14 and the Needham Virtual Semiconductor and SemiCap Conference on August 20 and 21. Now on to the call. Brian?

Thank you, Kate. Good afternoon, everyone, and welcome to the Q3 2025 Cerence earnings call. I'm excited to speak with you today. We are very pleased with our strong results this quarter, exceeding the high end of our guidance with revenue of $62.2 million and adjusted EBITDA of $9 million. Importantly, we generated strong free cash flow of $16.1 million, marking our fifth consecutive quarter of positive free cash flow. And for the full fiscal year, we are raising and narrowing revenue guidance to $244 million to $249 million. This brings the low end of our guidance range above the previous midpoint, and Tony will provide further details on our results later in the call. We continue to make progress on our 3 key deliverables for 2025, advancing our AI road map, growing our business with new and existing customers and continuing our transformation and cost management. First, we continue to advance the development of Cerence xUI, our next-generation hybrid Agentic AI assistant platform. While we previously differentiated between xUI Gen 1 and Gen 2, it's important to understand that xUI is not a static product, but rather a dynamic platform that is continuously evolving with new capabilities and features like multimodality and emotion detection rolling out over a strategic multiyear road map. In addition, xUI is designed to scale alongside our OEM customers' development cycles, ensuring alignment with their evolving go- to-market strategies and technical road maps. We are firmly on schedule with the major milestones on this road map, highlighting our disciplined execution and commitment to delivering long-term scalable value through AI innovation. And with that, we reached several important milestones for xUI within the quarter, including increasing language availability and advancing the platform's contextual reasoning capabilities. In addition, we continue to expand our partnerships with chip providers like Arm, enabling us to flexibly distribute and share computational loads between CPUs and GPUs to deliver improved speed and performance on the edge. And we believe this approach makes xUI one of the most powerful and flexible agentic AI platforms in the automotive space. And we're gearing up for another exciting milestone in September when we will showcase xUI at the International Auto Show in Munich. In Q3, we also continued to define and advance our AI agentic strategy within xUI. We're focused on developing our own stand-alone agents, namely a real-time knowledge agent and an integrated navigation and EV charging agent as well as empowering OEMs to integrate their own agents. Our calm family of proprietary language models is critical to this effort, enabling streamlined interoperability by serving as the orchestrator that helps ensure a consistent behavior and seamless integration across agents to deliver a seamless user experience. And we believe that the flexibility and openness of our architecture alongside our unmatched automotive expertise continue to be differentiators for Cerence, especially as OEMs navigate the complexity and ambiguity of the current market while still looking for ways to improve the user experience in their vehicles. And we believe OEMs continue to choose Cerence and are excited by xUI for 3 main reasons: one, our deep automotive expertise, which allows us to quickly and efficiently integrate into the OEM's hardware to access vehicle data and manage vehicle functions. There are up to thousands of automotive-specific operations and command along with complex hardware and software that require intelligent integration to function. Two, our open agnostic architecture and business model. Technology is evolving rapidly, making it difficult for OEMs to predict the future. At the same time, their customers are expecting LLM-based solutions in their vehicles today. This puts automakers in a difficult position to having to make decisions today without knowing what the future will hold. This is a tailwind for Cerence as we offer the architectural and operational flexibility and domain expertise that can help them bridge this gap, enabling OEMs to bring LLM-powered capabilities to customers today and in the future while also maintaining ownership over their brand experience and data. And third, our team. We believe we are the best in the industry working at Cerence AI and the majority of our team have long relationships with our automaker customers. This makes us easy to work with as we know the ins and outs of how our customers work. And as a result, we continue to see strong customer interest in xUI. In fiscal Q3, we signed a deal within the Volkswagen Group for one of its brands to have xUI as the basis of its next-gen system. We also expanded our work with JLR, bringing additional generative AI capabilities to their current platform as we work in parallel to build their future in car experience based on xUI. And importantly, we're continuing to see an increased PPU with these deals, and we continue to have a robust pipeline of ongoing customer interest with a steady stream of proof-of-concept programs across North America, German, Japanese and Chinese OEMs. Continuing on our deliverables for 2025, the second is to further grow our business with new and existing customers. We have a strong global momentum with our customers, including new design wins with Daihatsu and Hyundai as well as program extensions and renewals with Great Wall Motors and GM. Six major customer programs started production this quarter, including a BYD program for outside of China that spans 14 languages, illustrating that Chinese OEMs continue to look at Cerence to support their global expansion. The previously announced generative AI program with smart as well as programs with Audi, Geely, Mahindra, Nissan and PSA also went live within the fiscal Q3. Our commitment to partnering and supporting our customers is also being recognized. We received formal letters in appreciation of our collaboration from 2 major customers in China, and we signed a new professional services agreement with Mercedes-Benz. Our work with Mercedes-Benz signals our ability to deliver core foundational technology in coexistence with Big Tech and OEMs, evolving agentic AI strategies as evidenced by our work on the latest generation of MBUX that first rolled out in the new electric CLA. We continue our efforts outside of automotive as well with deep focus on identifying new verticals where we think we have a solid value proposition and can win. We're being careful to understand these markets and the existing players within them, and we're working with vertical partners to help us with go-to-market technology. This process, along with executing proof-of-concept programs ensures we don't overinvest before we know we have a good product and a market. For example, we have developed a call center agent and are working to identify a go-to-market partner for this offering. At this time, the agent is focused on service-related customer interactions in the car, where we can use our vast car knowledge and already ingested vehicle data to support incoming calls and requests. But we could see this solution being applied across a myriad of industries. We also formally announced our partnership with LG, which is leveraging our cloud, neural and edge text to speech to power voice interaction across its global television lineup, spanning 65 voices and languages across tens of millions of households. Cerence TTS enables LG TVs to quickly respond to user inquiries via natural spoken words, creating a more natural human-like user experience and simplifying users' interactions as they search for content and programming. So for example, the user could say, find me movies with Tom Cruise. And the TV can respond out loud while displaying options on the screen saying something like, you must like Tom Cruise. I looked up some of his movies. Do any of them interest you? Beyond the clear user experience benefits, Cerence TTS also enables LG to meet increasing accessibility requirements worldwide by ensuring that its television can be easily operated by people of all abilities. As a reminder, we believe the impact of our work to expand beyond automotive will be seen in our revenue and profitability in late fiscal 2026 and beyond. The third key deliverable for 2025 is to continue our transformation and cost management. As you can see from our continued strong cash performance, we are seeing the real benefits from this work, and it is being delivered to the bottom line for our shareholders. We continue to identify opportunities for further cost savings. Importantly, as we stated in our last earnings call, we did not see meaningful impact from tariffs on this quarter's results. For fiscal Q4 and therefore, for fiscal 2025, we believe the impacts will remain limited. However, it's important to note that the situation remains fluid and may evolve over the remainder of the year. We are seeing third-party projections of vehicle volumes down approximately 2.5% for Q4, and we're seeing some programs continue to push out and be delayed. With this in mind we're working cooperatively with our customers to find ways to optimize our partnership to best support them while also maintaining favorable conditions for Cerence. Lastly, we have a long history of investing in our technology, and we have and will continue to defend our strong IP. We want those who infringe in our intellectual property to know that we're serious about this effort. We have recently filed actions against Sony and TCL for their television's infringement of our voice technology patents. In conclusion, we're proud of what our team has accomplished this quarter and are encouraged by our results and our ongoing opportunities. We believe that we remain well positioned to support our customers and differentiate by our combination of technology and innovation, our diverse and expansive customer base and our deep automotive expertise. Our forecast for fiscal Q4 and 2025 demonstrates the strength of our products and the great work we've done to drive financial returns for our shareholders. With that, I'll turn it over to Tony.

Thank you, Brian. Good afternoon, everyone, and thank you for joining our Q3 FY '25 earnings call. We appreciate your time and interest in our company. Today, I'll be reviewing our Q3 results for fiscal 2025 and providing some guidance for our fourth quarter and resulting full fiscal year. Let's get into the Q3 operating statement. For Q3 FY '25, we reported total revenue of $62.2 million, which was above the high end of our guidance range of $52 million to $56 million. We saw strong contributions across all of our revenue lines. Variable license revenue was $34.2 million, up 48% year-over-year, reflecting solid utilization across our customer base and reflecting more in-period revenue from license shipments as compared to the prior year quarter, given the lower level of recent fixed contracts than in historical periods. We saw higher-than-expected volumes, possibly due to manufacturers producing ahead of potential tariff impacts and received benefit from favorable exchange rates on the euro. As expected, we recognized no fixed license revenue this quarter compared to $20 million in Q3 of last year, reflecting our continued strategic shift away from large upfront license deal. To reiterate, fixed license arrangement involves customers paying early on a nonrefundable basis in exchange for discounted rates. While these deals provide short-term cash benefits and meet certain customer needs, we've made a deliberate decision over the past 2 years to strengthen our pricing strategy and better align revenue recognition with product delivery. Historically, fixed license revenue reached as high as $70 million annually. Last year, it totaled $30 million. For this fiscal year, we planned approximately $20 million in fiscal Q2, primarily tied to budgeting cycles of our Asian-based customers. Going forward, we are comfortable in this $20 million range annually. As a result, while total license revenue declined 20.6% year-over-year this quarter, the underlying mix continues to shift toward more recurring and scalable usage-based models versus upfront fixed license arrangements. Connected services revenue was $12.8 million, up 17% year-over-year, driven by steady growth in our connected installed base. Professional services came in at $15.2 million, down 8% from prior year, reflecting a lower mix of custom work and greater implementation efficiency. Gross profit for the quarter was $45.9 million, yielding a gross margin of 74% compared to 72% in Q3 of fiscal 2024. The increase was largely due to a higher mix of technology revenue, offset by the absence of high-margin fixed license revenue. Non-GAAP operating expenses came in at $39.6 million, a 3% or $1.1 million reduction year-over-year. Q3 of last year benefited from a $2.4 million bad debt recovery. Otherwise, the year-over-year reduction would have been 11%. This reflects focused expense management while maintaining strategic investment in R&D. As a result, our adjusted EBITDA for the quarter was $9 million, which is well above our guidance range of $1 million to $4 million. Our GAAP net loss for Q3 was $3 million compared to a net loss of $314 million for the same quarter last year. In Q3 of last fiscal year, the company recorded a goodwill impairment charge of $357 million. This was a noncash charge that only affected our GAAP results. From a metric standpoint, we shipped 12.4 million units this quarter, an increase from 12.1 million in prior year. We also grew our number of connected cars shipped by 12%, underscoring the continued momentum in vehicle connectivity. We captured 52% of worldwide auto production remaining in line with our historical penetration. Adjusted total billings were $226 million, an increase of 3.5% year-over-year and consistent with plan. As a reminder, when we look at total licenses shipped, pro forma royalties is an operating measure we use representing the total value of variable licenses shipped in a quarter, including the shipments from fixed licenses where the revenue was previously recognized upon contract signing. We refer to the shipments where revenue was recognized in a prior period as fixed license consumption. Our pro forma royalties were $43.2 million, which were higher by $3.5 million as compared to $39.7 million in Q3 of last year. However, the consumption of our previous fixed license contracts totaled $9.1 million this quarter, lower than the same quarter last year by about 7% and in line with expectations. This drops more of the pro forma royalties into revenue in the current period as compared to a year ago. As discussed in previous calls, we anticipate a lower level of consumption given our lower level of fixed contracts than in historical periods. Our PPU metric increased to $4.91 for the trailing 12-month period, up from $4.47 for the same period last year, reflecting continued implementation of our improved pricing strategy and an increase in the adoption of connected solutions. When evaluating our liquidity position this fiscal year, we have been successful in reducing our total debt by $87.5 million using cash on hand. In Q1, we repurchased $27.4 million in principal value of our 2025 convertible notes and fully repaid the remaining $60.1 million in June. Combined with our $16.1 million of positive free cash flow during the quarter, our fifth consecutive quarter of generating positive free cash flow, we ended the quarter with $79.1 million in cash and marketable securities. We're comfortable in operating the business at these levels, supported by the expectation of continued positive cash flow. Now turning to our guidance. For Q4, we currently expect revenue to be in the range of $53 million to $58 million. As mentioned, we believe some portion of the Q3 upside was due to higher-than-expected production within the quarter, possibly due to OEMs producing ahead of tariff impacts in Q4. In addition, we expect no material fixed license revenue to be signed in Q4, and we expect to see normal seasonality where volumes in Q4 tend to be lower than Q3. With no fixed license revenue forecast in Q4, we expect gross margins to be in the range of 68% to 69; GAAP net loss to be in the range of $18 million to $22 million; and adjusted EBITDA to be in the range of $2 million to $6 million. With that, given our strong performance in Q3 and our Q4 guidance, ranges for full fiscal year expectations for revenue, adjusted EBITDA and free cash flow have improved and narrowed. Revenue guidance is moving from a range between $236 million and $247 million to a range between $244 million and $249 million. This brings the low end of our guidance range above the previous midpoint. In addition, we are raising our adjusted EBITDA guidance range to $42 million to $46 million, and our expected free cash flow range has increased to $38 million to $42 million. To summarize, we believe our Q3 performance demonstrates meaningful progress across variable revenue, recurring connected services and expense discipline. We remain confident in our long-term strategy, and we are actively managing our expenses while continuing to invest in innovation and customer value. Thank you again for your continued support. With that, I will turn it back to Brian to close our remarks.

Thanks, Tony. In closing, we're pleased with our results this quarter and proud of what our team has accomplished. Through the remainder of the fiscal 2025, we remain focused on our key deliverables, advancing our AI road map, growing our business with new and existing customers and continuing our transformation and cost management. And we look forward to sharing more on our view for fiscal 2026 in the next quarter's call. We'll now open it up for questions.

[Operator Instructions] Our first question comes from the line of Nick Doyle with Needham.

Congrats on the PPU progress. Can you expand on what drove the increases this quarter? Was that across both license and connected services? And you talked about 6 programs, these Gen AI programs going into production. But it's right to think that's too early for that to really impact the PPU line this quarter. So any more detail there would be great.

Sure. And you're right, it's too early for those to impact PPU. As we know, it's a trailing 12 months number, and those will come in as we really deliver against that. But yes, on the delivery for Q3, it was across all 3 lines. So if you think about the embedded side, we saw more volume from the embedded license this quarter. So -- and if you remember from previous discussions on revenue recognition for the embedded licenses that are shipped, those will get recognized in quarter. So we saw a higher volume there. And because of that higher volume, the way the fair value works for fair value accounting is some of it when we originally booked these contracts, some of that value goes to professional services. And so when we saw higher value on the embedded side, we also had an increase in the usage side of professional services. So you saw embedded up, you saw professional services up as well. And then from a connected side, we saw an increase as we noted, but that was really kind of in line with expectations. We've talked about this before. Remember that our customers self-report their royalty volumes. And so we have to accrue a certain amount until we get those royalty reports. And in some periods, we have catch-up kind of true-ups to what we've accrued to what this actually came in. So we also had some additional true-ups, which impacted the embedded line and the professional services line as well. And those are normal for the business, but they -- I think they were a little bit higher than normal this quarter. And then lastly, we saw an increase in the euro exchange, the dollar that benefited all of the line items as well.

Yes. Nick, I think part of the rest of your question was 6 new programs, are they affecting PPU this quarter? You're right. No, they'll go into implementation over the next quarter or 2, depending on the program. We are seeing, as Tony said, more connected devices or vehicles. We're seeing some upgrades as OEMs choose to push out some of the newer programs, but then we can offer them upgrades through the cloud, things like that, that are driving that PPU. So more connected, more features. The cloud gives us the option to really have that flexibility to offer them things that we can download overnight literally.

And then a clarification. I mean, sometimes you call out what that true-up number was. Was it not material enough to call it out this quarter?

Yes. And I think what we've said in the past that we have a material number with any single customer, and I think there was nothing material within a single customer. But we had true-ups across the board that were probably a little higher than in previous quarters. And again, those -- as we do those, it's a normal course of how we record revenue, right, as we get actuals versus what we've accrued for in the past. So nothing material for any single, but a fair amount compared to previous quarters overall.

Understood. And a bit of a macro question here. Just what are your customers seeing in terms of auto production in the next 6 months? I mean you mentioned this in your script, Tony. I think investors are concerned about weakening second half and some of that's contemplated in the guide. But are you thinking this will impact the December quarter as well? Like I guess, how much do you think was pulled in company-wide, industry-wide, just from your point of view?

Let me start and then Tony can go into more detail. But -- so each quarter, we've kind of told you we don't see much of an impact in the next quarter in tariffs and things like that. And so far, that's held. We do think there was a certain amount of build ahead in Q3 that occurred. It's a little bit hard for us to pin it down exactly how much, Nick, was, but that's a little bit why Q3 was so high and Q4 is kind of more of a standard quarter where that's just some of the build ahead we're seeing. And then -- but in the overall year, we're actually up and we've raised our guidance and raised the bottom of our guidance above the midpoint prior. That all shows that we believe the year will end out strong, right? As we go on to next year, I'm just going to be honest with you, we're not going to try and forecast next year yet. We have some initial IHS data that suggests there's some decrease in volume. What we don't know is there's a bunch of -- we talked about the number of POCs and new programs. We talked about the 6 that we won this quarter, those all have to be factored into that. So we have offsetting. We believe the connected vehicles will grow. So all of those things will balance each other, and we'll sit down and figure out how 2026 and Q1 of 2026 fiscal looks as we go through this quarter, and we'll give you those numbers at the end of the quarter. But right now, we believe Q4 will hold to our numbers. It means a strong year, raising guidance, raising midpoint, raising above the midpoint of the prior forecast. So closing out the year strong.

Yes, I think that's well said. And I think, if you think of our second half, when you combine Q3 and our Q4 guidance, we're ahead of kind of where our second half was. So yes, maybe a little bit of pull forward. We factored that into a little bit of volume decrease in Q4. But the other side of it is, as I mentioned, we had some true-ups this kind of across the board this quarter. Well, that part of it is a little bit of catch-up, but the other side is kind of new run rates for some of these programs that we have to bake into -- and since they're a little bit higher, bake those into our overall volume run rate. So a lot of puts and calls, but that's reflected in our full year guidance, which now is higher than as we said from last quarter. So we've pulled that guidance upward. And we'll take a look at how that impacts '26 as we finish out the year.

Our next question comes from Colin Langan with Wells Fargo.

I guess just following up on some of the recent comments. The guide does imply, I think, about a 10% decline at the midpoint quarter- over-quarter in sales. I think you mentioned about a 2.5% production decline. I think it might be a little worse. But why the big move down? Is that related to some of the true-ups in the quarter? Or is there other factors we should consider?

No. I mean you got to stop and -- I mean that's what we've been attempting to communicate is some of that was pull-ins from Q4 into Q3. So if you look at Q3, we really exceeded even the top end of our guidance for Q3. And we believe some of that was pull in from Q4 as people were trying to -- as OEMs were trying to get ahead of the tariffs that were supposed to kick in, in August. So some of that was simply a pull-in from Q4, let's call it, into Q3 to get ahead of that curve. But when you look at the full year and you take a look at Q4 and Q3 combined, because remember, quarter-to-quarter, we're going to bounce around as production varies a bit. And with tariffs coming in, going out, adjusting every month, week, sometimes, that's going to cause some of this fluctuation. You got to take a look at that long-term. That's why we are very confident in we're raising the year -- we're raising the bottom end of the year even above the midpoint of the prior guidance. That shows you the strong combined back half of the year. So it's really just pull ahead.

Yes, it's that. And then the volume number, we do look at overall IHS projections, which is that's the 2% to 2.5% that we discussed. But we also have to think about our specific customers within that population. And I believe that they were -- as you can see, our volume was up a little bit higher than even IHS. If you look at our volumes for Q3, you can see that we were higher than year-over- year and quarter-over-quarter than the IHS numbers. So correspondingly, I think we would pull that down in similar fashion, so a little bit higher than the 2% in anticipation of how much was pulled forward. Couple that with, I'm not sure, a lot of our revenue gets recorded in the tail end of the quarter as we get royalty reports and so forth. So I don't know what the euro will look like in the last month of our fourth quarter. So we've got some pullback there. And then some of the true-ups that we've seen, similarly, we don't want to anticipate that we'd have the same amount of kind of true-up volume as well. So we factored all 3 of those things in, and that's why you see the revenue kind of coming down a little bit in Q4. But still, if you look at -- it's a solid Q4, if you think about it apples-to-apples to where we were, would have been a year ago, without any fixed in either of the quarters.

Got it. And you flagged that you paid down the debt in the quarter with cash. How should we think about your cash balance now? Where do you think minimum cash needs are? And are you considering sort of adding some debt to boost the cash cushion in the future? Or are you fine where you are now?

Well, I'll answer that last piece. Absolutely not. I think about adding debt, no. Now would there -- could there be a scenario of potentially looking at the 2028 converts and refinancing those? Well, they're trading below par. So there could be some math there that you could refinance those. But if you think about our cash balance, we ended the quarter with $79-plus million in cash and marketable securities. The way we look at it is we can absolutely run the business with that level of cash on hand as a cushion and considering that we still project that we will be cash flow positive in the future. So our goal would be, of course, just like we did in this year, where we paid off all of our requirements, the $87 million from cash on hand that as we build that war chest going into 2028 that we could conceivably do the same and not have to either certainly not add to the debt number, but even not refinance and be in a position where we don't impact liquidity. So I think we're at a very solid base right now. Now we will always look at the capital markets and see if there's a more optimal recipe in that mix. But right now, it's very solid, and we don't have any anticipation of adding to the debt.

Our next question comes from Jeff Van Rhee with Craig-Hallum.

On the free cash flow. It looks great. A couple of questions. First on -- obviously, with the connected products, you've got some visibility into the features that people want. I'm curious if there's any discernible trends in there that are surprising you in terms of what's being used or not used. And along that same line, you talked a bit about AI. You talked about some of the capabilities. But as you move more into AI-centric offerings, what's the main opportunity to differentiate from tech bellwethers?

Sure. So there's a lot that I can answer with this. But let's start actually with the second part of your question about how do we differentiate because that also talks to you a bit about what are people using and what this connected really get you. The way we differentiate, there are -- and I tried to talk about it in the call was, one, we believe our technology is leading edge, right? We're moving into a world where we're multimodal, we can use sensors and cameras in the vehicle and really add a lot of feature capability to the vehicle. Second is we're agnostic. So when you talk about Big Tech, they tend to lock you in. They're going to lock you in to -- you've got to use a certain navigation tool. You've got to use certain LLMs. And every time somebody does a call out to an LLM, what's the newest Mexican restaurant, who won the football game last night, tell me about the weather. That costs money. And as those prices change, as more competitive market comes, OEMs want to be able to move across the LLM marketplace to optimize the features and the costs. And our option of agnostic and able to work with any of them really allows them to do that. And that's a big deal for these guys. It also futureproofs them, right? We see a lot of development going in LLMs right now. And you don't know who's going to be necessarily the leading player in that space 2 years from now. If you lock yourself into a Google-only answer or an Amazon-only answer, then you're locked in. And whereas you choose Cerence and we can go anywhere, whether it be Claude or you want to use Google LLMs or you want to -- whoever it is, we're able to do that. What are people using? They're starting to use this as they see the value in this. And really, what we're going to show at the International Auto Show in Munich next month is really the true agentic nature, where your car becomes a partner and you can literally just talk to it. And we're seeing people more and more enjoy that feature. Hey, my feet are cold. The system knows then to turn on the floor heat. Hey, I want to look at the -- what's the moon like tonight? Well, let me open this skylight and let you look out, right? Being able to have that natural discussion is where people are enjoying these features. This is where the real agentic value comes. And just like what we explained with the TV with LG, where you can just say, hey, what are all the Tom Cruise movies that are available right now. It allows you to just ask questions in a natural way and get answers back that are meaningful and helpful. That's what people are using.

Got it. Helpful. And on the sales front, can you talk about just new bookings in terms of do you feel you're taking share, holding share, losing share? Curious what you're seeing there. And then just last for me, modeling related. For '26, I realize you're not giving a guide, but can you give even early thoughts on maybe the consumption you think is a reasonable gross consumption of prepaid for the year that you think is a reasonable range?

So prepaid is easy. We've said that we are very comfortable in that same $20 million.

I was asking consumption, not actual prepaid, just the consumption of licenses.

Yes, yes. So you think of consumption, you saw really the first and second quarter, our consumption was higher, and you've now seen a kind of a run rate of consumption for the last 2 quarters in like $9 million range, right? And so which again meant that in the second half, and this is why our variable license was up so much. It was one of the components of why it was up 48%, right, is that more of it fell to the bottom line. So when sometimes people hear, well, it's easy. You didn't have prepaid this quarter. Is that a bad thing? Well, no, we've consciously reduced that number annually to this $20 million -- roughly $20 million number. What that means is as we do that, that more of that variable shipments that go out falls to in-period revenue. So we're about $9 million of consumption right now. We haven't put guidance out for this, but our consumption certainly year-over-year will not be up. It will be down just by the level that you've seen historically of fixed. The consumption has to go down for the entire year.

Yes, I'd say right now, our share is kind of holding where it's been. We're not losing anything that we're currently in. There's a few RFQs out right now that we'll see over the next quarter or 2 that are out for bid. And it's -- we're very competitive. We've typically made it past round 1, and we're in the final round of selection. So right now, we're holding. I'd say it's -- share-wise, we're flat.

Our next question comes from Mark Delaney with Goldman Sachs.

You have Aman Gupta on for Mark Delaney. Maybe just on this nonautomotive expansion that you guys are doing, and congrats on the LG news. Can you maybe help size kind of what that looks like in terms of revenue as it ramps in fiscal '26 and onwards? And how we should think about maybe something like a PPU in a large application for nonauto market relative to auto and how that pipeline of just general nonauto deals is continuing to evolve?

Sure. So I can't give forecast for '26, but I can give you a general -- how we view those kinds of markets. And TVs is a good one to just let's focus on. Typically, the price per unit is going to be less, partly because the complexity of the application is not quite there. You're not having all the car knowledge and having to do navigation along with real-time knowledge like who won the football game and things like that, you're typically just doing searches within a data set that's tied to your provider and things like that. So the PPU tends to be a lot less, but the units are quite a bit more, right? If you just take a look at the number of vehicles sold versus the number of TVs sold worldwide, you can far and away outpace the number of vehicles. If you add a couple of the major TV providers, you can get to a fairly large number. So we look at it from that perspective in that it's going to be a lower PPU, but it's lower technology and a bit simpler of an implementation. It's mostly cloud- based and all. And we're not going to do forecasts for '26 though yet. Sorry.

Yes. And thinking about that, as you know, the PPU number that we've started last quarter, given and continue this quarter is an automotive PPU that's part of, obviously, our primary market. So when we think about outside of automotive, I'm not concerned about a decrease in overall PPU because effectively, as Brian mentioned, it will be a lower PPU, higher volume, but more importantly, it will be additive revenue once those markets start to progress. So that's the way you should probably think about it.

That's very helpful. And then maybe for the second one, going back to the auto market. I think last quarter, Brian, you spoke about some potential pricing pressure from customers. And given some of the tariff changes, how has that evolved? Are you still seeing that pricing pressure? And then similarly related, are you seeing any continued delays with new awards or design activity? I know you mentioned some program pushouts in your prepared remarks as well.

Yes. We -- I'd say there's still pricing discussions. We haven't done any pricing reductions per se. What we've done is come back and we've talked about this in the past is how do we make this a win-win for both sides. And what we've tried to do is say, look, we can lower your overall costs by giving us more of the business, maybe more of the features. And we'll give you a lower cost for that bigger picture. That way, it's more revenue for us and a lower price point for them. So we have several of those offers out. And Tony talked about the PPU going up this quarter just in general. So that shows you that we didn't have to do any dramatic price reductions from that standpoint. And then there was a second part of your question, pushouts. The -- we are still seeing a few programs push out. There's several that are -- as I said, the RFQs are out. We're in the final stages. They should several of them close this quarter, we'll see or do those push out. We haven't -- as I mentioned in one of the earlier questions, we haven't had any major losses where we were competitive, and we were trying to do win backs or we had the business before and we've lost it. There haven't been any of those. But there are still several that have pushed, and we have several that are coming due, and we're hoping those will close and of course, hoping that will close in our favor, but we'll have to wait through the quarter.

[Operator Instructions] Our next question comes from Itay Michaeli with TD Cowen.

I wanted to dig in into some of the xUI wins that you talked about before. Could you share some more detail around your number of models that you would launch with a particular OEM, just the penetration levels there and the kind of mix between more premium luxury models versus mass market. I think you mentioned VW as one of the wins, maybe as an example.

Sure. So I can't give you like other than what we gave in the prepared remarks, I can say it's with somebody within the VW Group. But I'm not able to tell you exactly which group and which models, which brand within that group. We did mention JLR as an example. In general, what we're seeing with those programs is that we are seeing an increase in PPU. They are typically going across the brand. So it's not just the luxury end. It's at least down through the mid -- I'll call it, the mid-level, mid-section of the brand. There's some very low-end brands that maybe they go into other geos and things like that, that can be offering lower levels of the subscription. But in general, it's pretty broad across the brand as far as a model standpoint. But I'm just not at liberty to tell you there's marketing plans and announcements between the 2 companies that I just -- VW Group and all is like about as far as I can go with that as an example.

Yes. No, absolutely. Totally understood. That was actually helpful. Just as a follow-up, as we think about nonauto ramping in late fiscal '26 and beyond, should we think about any incremental OpEx to support that growth as well? Or do you think you can achieve that growth kind of with your current OpEx base?

No, I do not expect OpEx growth. I mean, for example, Tony and I did a review today on how we're becoming more efficient by using artificial intelligence internally to help us write our code to do some of our Q&A work, to build more tools that allow us to do analytics against our software at a much faster rate. So we think there's just in there, we can get 15% to 20% productivity improvement within our engineering team just by doing that work over the next year and expanding that to a broader. I'll tell you right now, we're kind of early in those stages, but we believe we can continue to grow that capability. And so our goal is to push that across the company and to fund our growth through that efficiency. And so we are going through that. We have regular like forums that we're reviewing at my level, the group's progress in this space. And we're talking about how to really make it measurable for next year so we can really see the productivity improvement. And the team is really excited about that because they look at this as a way to get to go and do new work and new things. It's actually fun for them to use these tools and they don't see it as a risk to their job or anything. They see it as an enhancement in a way that they can be more productive and get more done and do things that they actually didn't have time to get done in the past but make the product and the job much better. So no, I'm not going to fund it with increased OpEx. I'm going to continue to make improvements in OpEx next year and instead fund this through productivity.

That concludes today's question-and-answer session. I'd like to turn the call back to Brian Krzanich for closing remarks.

So I just wanted to close with, thank you, everybody, for taking the time out of your busy days to listen to our earnings call. We're really excited about the results we had in Q3. We're closing out the year, as we've talked about, strong by raising our guidance for the full year from a revenue and from an EBITDA and free cash flow. And we look forward to talking to you at the end of our fiscal Q4 and giving you the results for Q4, but also our forecast for 2026, which I know based on your questions, is what you guys are all looking forward to and need in order to continue to help us look at this company. So thank you very much for your time.