OUST - NASDAQ

Hello, and welcome to Ouster's Third Quarter 2025 Earnings Conference Call. [Operator Instructions] The call today is being recorded, and a replay of the call will be available on the Ouster Investor Relations website an hour after the completion of this call. And with that, I'd now like to turn the conference over to Chen Geng, Senior Vice President of Strategic Finance and Treasurer. Chen, please go ahead.

Thank you, operator, and good afternoon, everyone. Thank you for joining our third quarter 2025 financial results call. Today on the call, we have Chief Executive Officer Angus Pacala; and Chief Financial Officer Ken Gianella. As a reminder, after the market closed today, Ouster issued its financial news release, which was also furnished on a Form 8-K and is posted in the Investor Relations section of the Ouster website. Today's conference call will be available for webcast replay in the Investor Relations section of our website. I want to remind everyone that on this call, we will make certain forward-looking statements. These include all statements about our competitive position, anticipated industry trends, our business and strategic priorities, the development and expansion of our products and our revenue guidance for the fourth quarter of 2025. Actual results may differ materially from those contemplated by these forward-looking statements. Factors that could cause actual results and trends to differ materially from those contained in or implied by these forward-looking statements are set forth in the third quarter 2025 financial results release and in the quarterly and annual reports we file with the Securities and Exchange Commission. Any forward-looking statements that we make on this call are based on assumptions as of today, and other than as may be required by law, Ouster assumes no obligation to update any forward-looking statements, which speak only as of their respective dates. In today's conference call, we will discuss both GAAP and non-GAAP financial measures. A reconciliation of GAAP to non-GAAP measures discussed today is included in the financial results release that was issued today. I would now like to turn the call over to Angus.

Hello, everyone, and thank you for joining us today. I'll start with a brief recap of the quarter and review our strategic priorities. Ken will cover our financial results in more detail before I close with some final thoughts. Our third quarter results reflect the continued growth we are seeing across our business with revenue of $39.5 million, representing our 11th straight quarter of revenue growth. We set a new quarterly record with over 7,200 sensors shipped, bringing physical AI to life across multiple applications, including yard logistics and traffic intersections. Gross margin remained strong at 42%, and we ended the third quarter with $247 million of cash and equivalents and no debt. This performance further demonstrates our ability to convert pilot programs into large volume orders as we deepen our relationships across our diverse customer base. In our Smart Infrastructure vertical, we expanded deployments of Ouster Gemini and REV7 at logistics yards around the country, helping our customer improve throughput, efficiency and safety. We continue to progress with testing new Gemini AI algorithms at key customer sites during the quarter with the potential to expand use cases and more than quadruple the number of sensors per logistics yard. We also won new deals to bring Ouster BlueCity to additional intersections across Utah to enhance traffic flow, safety and operational efficiency. In our industrial vertical, we shipped a significant number of REV7 sensors to a leading global technology company as it continued to expand the use of autonomous mobile robots across its warehouse floors. Ouster's sensors are mounted on a variety of warehouse equipment, including AMRs, forklifts and tuggers, enabling our customers to detect and avoid nearby objects and helping heavy machinery to work safely in unstructured environments. We also secured a substantial order to supply REV7 to a large European industrial equipment manufacturer, which is upgrading the sensor stack on its next-generation electric mining trucks. These trucks are part of an autonomous haulage solution that increases safety, productivity and efficiency while producing 0 greenhouse gas emissions during operation. During the quarter, we delivered REV7 sensors to support the continued expansion of Serve Robotics last-mile delivery fleet across the United States. Last month, Serve announced its 1,000th robot deployment compared to an average of 57 active daily robots in the fourth quarter of last year and expects to reach 2,000 robots in service by the end of this year. Serve is a prime example of Ouster's engagement with companies that are rapidly accelerating from initial testing to commercial deployment. Turning to our 2025 strategic priorities. We progressed across all 3 key focus areas: one, scaling the software attached business; two, transforming the product portfolio; and three, executing towards profitability. Our software attached business gained traction during the quarter. Yard logistics was a key driver of demand, and we also won a deal to deploy Gemini for crowd management solutions at major tourist sites and large events in South Korea. More customers are recognizing the benefits of our Lidar solutions, and we won pilot deployments for intelligent perimeter security, spanning energy and industrial sites, Tarmac, data centers and defense facilities across the world. In September, we announced a strategic partnership with Constellis, which now offers a unified security solution enabled by Ouster Gemini and Ouster Digital Lidar. By investing in AI perception, Ouster has built a core platform to enable our customers to develop targeted market-specific applications. With Gemini, Constellis can provide real-time analytics, threat classification and automated response protocols to bring physical AI to advanced security operations. Constellis' operational expertise and global network positions us to rapidly advance Gemini for critical and large-scale security operations. In the ITS market, I'm excited by the continued growth of our Ouster BlueCity solution, where we won large deals in the U.S. and Canada. We continue to expand our distribution network and signed 7 new exclusive partnerships to bring BlueCity to additional states, including Illinois and Missouri. We are proving the value of AI-powered Lidar to state and local governments across the nation, and our Blue City partnership network now covers the majority of a nationwide market of over 300,000 signalized intersections. We also brought on a transportation integrator in Europe following a successful pilot deployment in Brussels. These partnerships in conjunction with expanding our BlueCity bundles to provide customers with more setup options are key actions to support the continued growth of our software solutions. Moving to the product portfolio. We continue to make major investments in retraining our AI algorithms on an ever-expanding corpus of field data. In the third quarter, these efforts delivered better detection accuracy at longer ranges and higher vehicle speeds to support use cases like tolling and highway monitoring. We also released real-time localization or RTLS, in our Ouster SDK. RTLS empowers our customers to understand the position of their assets with centimeter level accuracy, enabling features like geofencing, automatic speed limit enforcement and custom go/no-go zones. In addition, Lidar-powered RTLS significantly reduces the investment and infrastructure required by legacy sensors. Our team continued to progress with testing and validation of our next-generation L4 and Cronos custom silicon. These investments will unlock major performance, security and reliability improvements for our OS sensors and become the backbone of our solid-state digital flash line. The innovations from this next phase of our product roadmap are expected to more than double our current addressable market and are the most profound investment in our product roadmap to date. Finally, I am proud of our team's consistent execution towards profitability as we deliver record results in the third quarter as we remain on track to meet our long-term financial framework. I'll now turn it over to Ken to cover our financial results in more detail.

Thank you, Angus, and good afternoon, everyone. I want to open my comments by noting that since joining the company in May, I have witnessed firsthand the incredible dedication and laser focus on execution towards our 2025 company goals that Angus discussed, and I'm excited about the opportunities in front of us. Now turning to third quarter financial performance. As Angus noted, our results reflect the underlying strength in our business. Revenue of $39.5 million was a record, representing growth of 41% year-over-year and 13% sequentially. We delivered more than 7,200 sensors, which also represented an all-time high. As a reminder, we do expect a level of fluctuation of volumes on a quarterly basis as it is largely dependent on meeting our customer delivery and timing needs. Smart Infrastructure was the largest contributor to the third quarter revenue, followed by roughly equal contributions from our robotics and industrial verticals. GAAP gross margin of 42% increased by 4 points compared with the third quarter last year and reflects the benefits of steadily increasing revenues and product mix, offset by continuing tariff headwinds. While our gross margin performance has been strong this year, we maintain that 35% to 40% is an appropriate long-term annual gross margin target for the business. Next, GAAP operating expenses were $41 million in the third quarter, up 7% over the prior year. The increase was primarily driven by investments in R&D to support the new product development cycle. As I mentioned last quarter, we remain focused on managing our operating expenses, but anticipate there will be variability on a quarterly basis, largely due to the timing of investments in our innovation and go-to-market expansion. Adjusted EBITDA was a loss of approximately $10 million, flat year-over-year and a decline of $4 million sequentially. The sequential decline is primarily due to a favorable employment tax refund we received in the prior quarter. We are pleased with our ability to drive growth and have the operational capacity to meet our customers' needs. Creating strategic and operational flexibility for the company to innovate and grow as we continue to execute towards profitability remains a top priority. Our balance sheet is one of the strongest in the industry, which is important for our customers as they depend not only on the long-term support of our products, but also our long-term financial security as a key supplier. We ended the quarter in a stronger position with cash, cash equivalents, restricted cash and short-term investments of $247 million. This includes approximately $35 million of net proceeds from our ATM. At September 30, we had approximately $4 million of authorization remaining on our ATM. Moving to guidance. For the fourth quarter, we expect to achieve revenue between $39.5 million and $42.5 million. Thank you for your continued interest in Ouster. I'll now turn the call back to Angus for his closing remarks.

Thanks, Ken. Ouster has a strong financial foundation, a robust distribution and partner network and a diverse customer base of emerging and world-class companies. Ouster is at the forefront of technology that is reshaping how the world engages with the physical environment. Our physical AI solutions are helping deliver improvements in safety and efficiency across a wide range of industries. All of this, coupled with our cutting-edge product roadmap positions us well to further accelerate the adoption of physical AI. With that, I'd like to now open up the call for Q&A.

[Operator Instructions] Our first question today comes from the line of Colin Rusch with Oppenheimer.

Can you talk about where you're at in the testing process with the Rev8 and the Cronos offerings? Would love to get a sense of kind of how that testing is going. Any sort of concerns or kind of accelerations that you're thinking about with the platform given the potential growth and addressable market here?

Colin, thanks for the question. So we really try not to talk ahead of the release of our next-generation products other than to make sure that it's clear that we remain incredibly committed to the investments we're making in the digital Lidar portfolio. So our L4 chip, the Chronos going into the DF platform, these are things we talk about each and every earnings call because they're still the biggest source of investment that we have at Ouster and because of all the promise, the importance of these products to the future expansion of Ouster's business. So the points that we've made in the past and we continue to make on this earnings call, a doubling of the overall TAM the most significant set of products -- hardware products in Ouster's roadmap in Ouster's history, all remain true. We're incredibly committed and focused to getting these products out as soon as humanly possible. But beyond that, I'm going to just leave it at that.

Okay. Fair enough. And then as you work through the design cycle with your customers, and we know that there's an awful lot of customers you guys are working with. There's a lot of innovation happening in industrial hardware design. Can you talk a little bit about the cadence of those programs moving forward? We know that you have a number of wins and moving from kind of smaller volumes into more series production, particularly with some of the off-road vehicles. But what you're seeing in terms of just the cadence of design cycles, the adoption rates, any sort of win rate data that you can share? I would love to get a sense of how that's evolving here.

Yes. I mean so we have over 1,000 end customers. And one of the points that we've made and one of the kind of bright spots about Ouster's promise of the future is that there's a small minority of all of the high-quality customers that we have that have actually reached full-scale production and commercial release of their products that are built with an Ouster Lidar inside. And so that means there's immense opportunity in our existing latent customer base for tranches of these customers to go from development all the way to commercial lease. And we mentioned on the call, Serve Robotics, a great example of a customer that shows how the volumes shift from a kind of development, small-scale pilot style production where they had 57 robots deployed with our technology, if you look back a year or so ago. Now they're on track to have 2,000 robots deployed with our technology in the next couple of months. So that kind of fundamental order of magnitude shift is a big part of our growth strategy for the foreseeable future. And a small fraction of our overall customer mix under 10% is actually in that full-scale production. So -- and -- but one of Ouster's core kind of muscles that we've built on the commercial side is our ability to support our customers developing these challenging new technologies and close gaps that maybe we have better expertise closing than our customers do, either on the hardware or the software that processes our Lidar technology so that we're getting customers to market faster and they're seeing that we're a valued partner in that process versus just a hardware supplier. So I think there's a lot of -- yes, there's a lot of kind of goodwill and deep partnerships that we've built along the way. And we're continuing to do that. It's something that our customers value at this point.

And our next question comes from the line of Andres Sheppard with Cantor Fitzgerald.

This is Anand on for Andres. Congrats on the quarter. With the rapid acceleration of self-driving vehicles, both passenger and commercial vehicles, do you guys expect to pursue this vertical a little bit more aggressively going forward? I know it wasn't as much of a focus this quarter. But are you looking for any major OEM agreements? And what would be an ideal candidate? Because it seems like most companies with the exception of Tesla are really reliant on Lidar for this.

Yes. Thanks for the question, Anand. So I mean, it's -- first, it's great to see the renewed kind of resurgence and interest around self-driving vehicles. A lot of this is because of the advancements that Waymo has made in really providing commercial service to customers out here on the West Coast and in Texas and Arizona and then also some of the advancements from Tesla. So -- it's great to see this. Ouster already has some really strong partners in this area. We're talking about robotaxi specifically, Motional and May Mobility, both are strong Ouster partners. We've seen a lot of great partnerships that May Mobility has been inking with -- to build their customer base and actually expand their commercial robotaxi deployments. So Ouster already has some of these great customer relationships. When it comes to the OEMs and kind of direct OEM integration of this technology into a car you and I can buy, that's where I have tempered expectations in the past, just basically because of the long time horizon for OEMs directly integrating self-driving technology into the cars that you and I can buy. That's largely because of technology difficulties on the OEM side versus like the readiness of compute and sensor technologies that Ouster is responsible for. So -- but on that latter point, we're absolutely interested in this space. What I've always said is it's important to have the right products with the right -- at the right point in time for that adoption to happen. I think a lot of things are converging. We have put a lot of investment into the DF and the future products on that internally at Ouster so that, again, we have the right product at the right time for this massive opportunity in direct OEM integration. So definitely interesting to us in the future, something I've tempered in the past, but I think the stars are aligning in the next couple of years here.

Got you. That's helpful. Just switching gears a little bit. I guess, for the past few months, the elephant in the room has been the Blue UAS certification. So I was wondering what are maybe some of the most recent updates related to that? And if you could potentially give us any granularity on sensor shipments? Or if not, do you continue to believe that you have a moat in this segment? Or are you seeing more competitors pursuing this now?

Yes, specifically asking -- so for those listening on the call, the Blue UAS certification was really -- it was a certification for using Ouster's Lidars on defense DoD use cases and payloads, specifically for drones. So the common use case here is -- or the traditional use case for Lidar on drones is a surveying payload, surveying and surveillance payload. Ouster is a robust business as a surveying payload on drones already and the UAS announcement made us the first DoD Blue UAS certified company in the mix. And it's definitely a boost for our business. We're not splitting out specific sensor volumes, but we do see inbounds from customers that are interested in making sure they're operating certified payloads. And whether or not, sometimes it's because the end customer is a DoD customer. And sometimes the end customer isn't, but values the fact that we're using a certified American-made technology. So definitely a bunch of benefits there. I think we do have a moat. We're the first -- we're certainly the first mover in this space. And we have a great set of products that apply really well, small form factor, super high resolution, robust Lidars that don't weigh a lot. And all those things make sense if you want to put these on a small form factor drone like the Blue UAS certification is positioned for. So yes, not splitting out any specific numbers, but definitely a benefit to our business.

[Operator Instructions] And our next question comes from the line of Madison de Paola with Rosenblatt Securities.

This is Madison calling on behalf of Kevin Cassidy. I was just wondering with so many customers moving from prototype to production, what steps are you taking to mitigate potential supply chain constraints that could impact growth? And just as a follow-up, what's the long-term target for BlueCity's attach rate?

Well, let me start with the latter first. Thank you, Madison. The bigger thing is we're not breaking that out and giving the long-term target. It is part of our overall robotics and industrial outlook that we have already. So if you just stick with those growth rates that we talk about, that's the majority of what would be covered there. Turning to your first part about capacity. One of the things we've done very well. And if you look at the growth just these last 2 quarters, we set 2 record quarters of shipments pretty much quarter-over-quarter, our sensors grew year-over-year for this quarter alone was 85% and quarter-over-quarter from last quarter to this quarter is a 31% growth. So having that capacity is very important to us as we continue our growth journey. So part of the capital investments we make aren't just strategic. It's also financial flexibility. What's important for us is meeting our customers' scheduled demands because while we pride ourselves on the continued growth, our customers are growing just as fast. And so, we have to have the capacity to deliver their needs so they can meet their customers' needs. So we will always be investing in capacity to ensure that we can meet the customer demands.

And our next question comes from the line of Richard Shannon with Craig-Hallum Capital Group.

This is Tyler Anderson on for Richard. So Amazon has been talking about adding a lot of robots in the future and including the humanoid robots, do you think there's going to be a benefit to you from this? I have seen some pictures of robots that look similar to yours. And how would they show up in bookings when that starts moving forward? Is this something that takes a long time that needs to be built out? Just any way to think about that would be helpful.

Yes. It's a good question, Tyler, because this is a fast-evolving space. I'm amazed how many humanoid robotics companies have been announced in the last year. Overall, definitely a great thing for us. Humanoids need sensing technology like any other robot and Lidar is the best possible sensor you could put in the mix. And we already have some customers that are using our Lidars in their humanoid robotics platform. So definitely good news there. You would see -- the way that's going to impact our business is it would boost our robotics vertical, right? That's where this would fall into the financials or the financial performance of the company. I'd say it's still early days, like there aren't thousands of humanoid robotics -- humanoid robots that are deployed at end customer sites right now. It's a prototyping environment. So I don't expect it to be a big impact, positive impact on Ouster's business for the next year or so, foreseeable future. But this is all about laying the groundwork for a future tranche of customers to reach commercial deployment just like what we've seen with some of our other verticals happening all the time. So we love investing in new customer sets. I think the humanoid thing is interesting, but I think it's going to take a couple of years to play out.

Great. And then you mentioned something about a majority of intersections, about 300,000 in the U.S. Is this the total addressable market that you're speaking to? Or is this something that you already have plans and that's moving forward with business in hand? Just want to get a look at that. And then also, is there any way that you could categorize the attach rate for your traffic data?

Yes, absolutely. So what I said was that we had signed exclusive partnerships and distribution partnerships that covered regions for the majority of signalized intersections in North America. There are about 300,000 signalized intersections in North America. That's the total addressable market. But I think it's a market that we can largely go after aggressively today. BlueCity is a best-in-class intersection -- intelligent intersection product. It can cover a wide swath of the use cases today. We haven't quantified exactly how many of those 300,000 intersections exactly that BlueCity can go and capture, but it's a significant fraction. And a major impediment to going and addressing that market is just having regional partners that we can sell through that can support the end customers, not just in installing the technology upfront, but also supporting them for the long term. It's important that a municipality has support on their traffic infrastructure for many years to come. So we have a lot of inertia there. We announced 7 new exclusive partnerships. And so overall, we are -- we have partnerships that cover the majority of the North American market. When we're talking about attach rates, BlueCity is by default, a software-attached product. You cannot just buy sensors and you cannot just buy software. You have to buy the whole complete solution that goes turnkey onto your intersection. So I wouldn't -- every BlueCity cell has an attach rate of 100% for Lidars, has an attach rate of 100% for a software component. So it's more -- what we're seeing is that we're growing pretty fast in this market. Smart Infrastructure was our biggest vertical this quarter. And so my goal is instead of looking at attach rates per se for BlueCity, it's looking at the growth rate for BlueCity versus the rest of our business. I think there's some early signs that there's some really positive fast growth happening there.

[Operator Instructions] Our next question comes from the line of Casey Ryan with WestPark Capital.

Great quarter. We've talked a little bit about defense. I just wanted to get your perspective on that as a vertical because I think there's a lot of focus on drones. But as a company, do you guys define it as maybe being service-wide, meaning potentially all vehicles could sort of use automation? And as part of that, do you see sort of a retrofit opportunity as being potentially significant in addition to new weapons platforms and vehicle platforms?

Thanks, Casey. It's a thoughtful question. So the defense market is incredibly diverse. I think that's the first thing to acknowledge. And there's legacy vehicles already deployed in the DoD. There are traditional defense contractors. And then there's this new tranche of kind of faster-moving start-ups in the space. Ouster is really focused, I would say, on the non-retrofit opportunities, working with the traditional defense industry or the new players. And yes, I think that the retrofit opportunity may be not something that we're tracking. But overall, like there's a big opportunity here, but with, I would say, an unclear timeline to the scale where this is deployed universally on these vehicle platforms. I do think that that's where it goes. Automation is good in this -- no matter what in this industry. But it's going to take quite a while, I think, to field this technology in a big way. But there's some promising first -- places where this is useful even today. So Blue UAS surveying platforms, great example. It's not automation, it's surveying. That use case ready today, being widely deployed and used, great for Ouster's business. Automated systems operating in the field in life or death situations, there's a very high bar for fielding that. And I think it's going to be a couple of years before that's a major impact on Ouster's business. But Ouster is as well positioned for this industry as anyone in the world.

Okay. Terrific. That's a very thoughtful and helpful answer. Very quickly, there's the potential for DJI to be blocked, I guess, for U.S. shipments. Could you see that having an impact to your commercial opportunities because I think DJI obviously dominates market share in the U.S. for people who are using commercial drones for businesses and things like that. But I wonder if we should associate sort of a blocking as being positive somehow for Ouster in terms of serving domestic manufacturers.

Yes. I think that maybe there could be a positive impact on just the general awareness on where customers are sourcing critical technology in their supply chain. So the scrutiny of DJI, it's an adjacent market to us. But I think more would be around the general kind of perspective on strategic supply chains and knowing who you're buying from and maybe there's some bleed over that benefits our business. So net-net, a little bit of benefit for Ouster.

And our next question comes from the line of Tim Savageaux with Northland Capital Markets.

Congrats on the results. I might have dropped off there for a second. So sorry about that. My first question is you called out Serve Robotics as a kind of an example of the deployment of technology and volume. I wonder, we've seen some work. You obviously just did a deal with DoorDash for Los Angeles. There are some estimates that L.A. by itself could take 10,000 robots. As you look at the scale of this opportunity, is that something that's significant in terms of potential growth drivers, whether it's last mile delivery in general or Serve in particular that you're focused on?

Yes. Serve Robotics is -- I think they're having their Waymo moment, right? It's been many years of wondering, is this market -- is last mile delivery viable as a business, is it ready technologically for the prime time. And they've stuck to their game plan of making the technology and the commercial strategy work. And here we are with them now moving to orders of magnitude greater deployments. So kudos to them. I think that they're -- that's the best evidence that this is a real market with -- well, not just a real market with real demand, but that it's a viable market now. So it's easy to be skeptical and pessimistic and people were of Waymo and then they kind of scaled by orders of magnitude and now everyone is a true believer. I think that's what's happening with Serve, and I'm happy for it, both because they're a customer and because it also is like a harbinger of good things to come in the rest of the last-mile delivery market.

I think I just want to add on there, too, Tim. It's a great proof point for our strategy of go-to-market, right? We across multiple verticals. It's just not a one vertical play. And Serve is just one of our thousands of customers who were in these early stages who scaled to success with their business plan and prevailed and gave us the opportunity to grow with them. So not just betting on one sector and riding that but having the foresight to go across multiple sectors and really work with these companies through their success, it's paying off for us now.

Okay. And that's a good segue to my follow-up, which is, Angus, I think you mentioned sub 10% of your customer base having scaled into full production and I don't know how far sub-10%. But I guess if we look a little bit forward, I don't know whether it's a year or 2 and that number is 25% or 50%. What are the implications there for your overall revenue opportunity?

Yes. I mean I'm pointing straight to our model of the 30% to 50% growth, right? When you look at how we're progressing with that and with that going into production, I mean, we've had some good tailwinds with margin lately, but I think that would keep our margins in that 35% to 40% on a GAAP basis we looked at. But we fully are looking at that ramp-up. That's where you can get towards the higher end of that 30% to 50% range.

Yes. And I would say this is exactly the sub 10% in production is why we talk about Ouster being in the early innings or I think last quarter, we said we're still in the dugout. We're not even playing the game yet. So there's a long way to go and a lot of growth for Ouster to grow into our TAM numbers. We've put out TAM numbers. I think those are real in the long term. And -- but it just speaks to the confidence we have in hitting our 30% to 50% revenue growth for the foreseeable future.

And I believe we have Tyler from Craig-Hallum with a follow-up question.

Just a quick one. So thinking about your software business, are you -- or are customers able to use other sensors with your software? And essentially, could you be just overlaid in different use cases with your software for what people have already purchased?

So short answer, no, you cannot use a different sensor with our software, and that's why it's a software attached business. We really always focus on the fact that we're selling systems BlueCity and Gemini combine our sensors with our software and in most cases, our compute as well. So software attached business. There are cases, though, where you can buy -- we have some customers, distributors that used to sell just Ouster Lidars, and now they're selling Gemini on top of those Lidars maybe after the fact to a certain set of customers. Maybe a customer thinks that they can write their own software for our sensors and realizes that after trying for a little while that Ouster has something more mature and they can sidestep a bunch of difficult technical issues by purchasing the Gemini software themselves. So we do have some cases there. But again, the end result is that the customer is running a software attached product solutions product from Ouster.

Yes. And Tyler, I think the goal here for us is that software attach is buying the full system and perception and sensing fusion from us. The goal is that you would use an Ouster product set for your sensor and perception. And then once that software is integrated into your software stack, they can grow with us for generations because our -- we write this, so it's forward and backward compatible on the hardware elements that we sell. So regardless of the generation we're on, it gives us the flexibility. The other thing I would just add to that is if you start thinking about into the future, having that software attach rate, that makes us extremely sticky to our customer bases. So once you get in there, you want to provide not just quality service and quality products, but if you can provide a whole system and a platform that they can grow generation over generation, that's the ultimate goal of this play.

Awesome. Just one more for me. So just thinking about the software and the model training, this is all traditional machine learning, correct? And is there any way -- or are there any customers that they're pulling for some kind of LLM or visual model capability that isn't traditional machine learning?

I don't know if we've ever used the term traditional machine learning. I would definitely say that for this type of Lidar perception, we are using cutting-edge models, but they're not text models. So LLM is almost a misnomer for this industry. But we are using cutting-edge deep learning models in our products, trained on giant corpuses of annotated data that we've collected from the field. So true kind of cutting-edge fleet learning, true cutting-edge deep learning models used best-in-class in the perception space for Lidar. So yes, I think that there are ways we could augment our products with things like LLMs or maybe that the perception space will transition to transformers and LLMs. But the cutting edge is actually what we're using, and that's deep learning.

Got it. And I meant traditional and non-transformer models coming from a data background. I was just trying to differentiate from that.

Yes.

And with that, I would now like to say that Q&A is concluded. So I will hand it back to Angus for closing remarks. Angus?

All right. Well, with that little discussion on LLMs, thank you all for joining the call, and have a great rest of your day. Cheers.