Hello. Welcome to the 30th Deutsche Bank Depositary Receipts Virtual Investor Conference, dbVIC. My name is Zafar Aziz from the DR Investor Relations Advisory Team at Deutsche Bank. I am pleased to announce that our next presentation will be from WeRide. Before handing over to our presenter, some points to note. At this point, I am very pleased to welcome our speaker from WeRide.
Thank you for the opportunity to introduce WeRide, a global leader in autonomous driving and technology. At WeRide, our mission is simple yet ambitious: to transform urban living with autonomous driving. We're building a universal autonomous driving platform that scales from spans from L2+ passenger vehicle solution all the way to full driverless L4 mobility services, enabling safer, more efficient, and more intelligent transportation system around the world. Founded in 2017, WeRide has quickly become one of the few global leader in autonomous driving space that has demonstrate the technology depth and the real-world operation and financial scaling simultaneously. Today, we operate in 12 countries and in more than 40 cities with trial and commercial operations staying in China, Middle East, Europe, Southeast Asia, and other part of, like North Asia.
The breadth matters because autonomous driving is not just a software challenge. It's regulatory, operational, and localization challenge. Our ability to operate across different transportation system, legal framework, and rider behaviors is a core strategic mode. Operationally, now WeRide manage one of the world's largest autonomous fleet, that including more than 2,100 L4 autonomous vehicles and more than 1,100 robotaxis. The scale provides not only the revenue opportunity, but also one of the richest real-world data loops in the industry. On the financial side, we just published our 2025 full year result, our revenue increased 90% year on year and get to RMB 685 million for the yearly revenue. Annual revenue. Recently, we started to disclose more numbers on the robotaxi operation.
Here you can see are the some of the latest number we have. In terms of the total cost of ownership, it has decreased 38%. It's mainly driven by the improved operational efficiency and the lower, like, vehicle BOM cost. On the second part, you can see we have a core ratio in our industry called remote assistance ratio. In 2024, that number is 1 to 10, meaning, like, one remote safety officer can overlook 10 vehicles at the same time. Right now that ratio has improved to 1 to 40, which it's a dramatic, like, decrease in terms of, like, the safety officer cost. Mm-hmm. And also, we are, we have, like, a significant improvement on the vehicle platform cost in one year time, and that BOM cost has dropped by 15%.
Mainly it's enabled by the new HPC 3.0 computing platform as well as, like, we are building in larger scale BOM factory L4 pre-installed robotaxis, so that also streamlining some of the expenses. In terms of our robotaxi fleet side in China, we have more than 800 that's up and running every day, mainly in Beijing and in Guangzhou, and soon in Shenzhen you will see a lot of our robotaxis running on the street. In terms of the coverage, like, ODD, right now the official operation zone covers more than 1,000 sq km.
For some of the daily order numbers, for the past six months average, every day we have 15 daily orders per robotaxi at some of the peak time, that number can go to 26. Even though in China we're not breakeven yet, we think we are in the next year or so, there's a high likelihood since we are also expanding ODD, we can get to the breakeven for China robotaxi soon. Given WeRide is one of the key player globally, our outside China fleet has already way past the breakeven point. Since we're talking about the China operation, here are some of the other numbers we can disclose. Year-on-year, we have a 900% growth on the registered users.
That shows, like, since we are expanding ODD and releasing more vehicles on the street, now we have see a higher, like, user engagement, like, rider engagement rate from the app. We have launched the Free PUDO. Free PUDO, meaning it's a free pickup and dropoff points. It's this unlimited pickup and dropoff points selection within the ODD. That's exactly the same, like, pickup and dropoff mechanism that Uber or DiDi, those, like, major platforms are using. We have already adopt the same, like, selection, point selection in robotaxi.
Technically, that's a lot harder for robotaxi to do this, but we managed to achieve this last year. Also in order to reach to more, more riders, we also work with Alibaba's Amap, as well as WeChat on the WeChat mobility program, as well as their mini program so that so we have broadened the domestic reach with the other major mobility platform. WeRide's global footprint is one of the strongest proof points of scalability. We have autonomous driving permit in eight countries, and we have deployed autonomous vehicles in 12 countries. So if you're looking at our AV fleet outside China, it's over 250 robotaxis that spans from like, we have fleet in Japan and in Singapore, outside China and Asia.
In the Middle East, we have a quite a large presence in U.A.E. as well as in the Kingdom of Saudi Arabia. If we're looking at Europe here, we have the fleet in Belgium, in France, in Spain, and in Switzerland, as well as in Slovakia. We do have a very global reach for our fleet. Middle East is the standout success story. Like two years ago, WeRide became the first company that secured like a nationwide autonomous driving license from the U.A.E. federal government. Last year, we became the first company to secure a city-level, fully driverless robotaxi commercial permit outside U.S.
This means a lot to autonomous driving companies because by having city-level permit to run robotaxi, which means like we have the flexibility to add on as many vehicle as possible to get a higher penetration, as well as to make sure that all of the robotaxi fleet can achieve a relatively high utilization rate. It's one of the most like significant and important permit we have so far. Speaking of Abu Dhabi, now our service covers nearly 70% of the city core area, and the market is already approaching its. If we were talking about WeRide and Uber, which is also one of the main platform we're operating as a whole, we are reaching to the break-even point under the driver-out operation.
As for the WeRide side, as if we're looking at the Middle East subsidiary at a standalone basis, it already is a profitable entity for the past two years. Yeah. We have also launched the public robotaxi rides in Dubai and we are right now is the only public robotaxi service provider in Riyadh. Within the region, we have more than 200 robotaxis, making it one of the largest scaled robotaxi deployment outside China. Yeah. In Europe, we secured the first driverless robotaxi permit for passenger service beginning in Switzerland, while we are also expanding to Slovakia and other market. This year, we'll announce, we will enter into another two European key cities, major cities for the robotaxi deployment as well.
This is very strategically important because Europe's regulator bar is exceptionally high. The early permit here is also a strong signal of like the safety credibility in Europe for WeRide. And in Singapore, we commenced robotaxi and robobus public operation, creating a strategic like beachhead in Southeast Asia. The ability for us to scale across Asia, Europe, and the Middle East demonstrate that WeRide is building not just a tech stack, but it's a global autonomous mobility operating system. On the other side, on the L2+ side, we have launched the WeRide WePilot 3.0. This is a high-end L2++ like solution.
We work with tier one, the global tier one, like Bosch, and we have OEM partnership clients like Chery and GAC. We have demonstrated industry-leading performance, recently we have won three consecutive wins in the national urban NOA competitions. Our system delivers, if we're talking about the performance level, we can do a few hours driving in the city without intervention most of the time. In terms of the driving efficiency and the lower intervention as well as the regulatory compliant behavior in a very complex urban environment, WeRide has demonstrated to be the best in the country so far.
Later on in the slides, we'll talk about the competition result and how the competition it's actually scored. Yeah. On the technical technology side, why we're different. The heart of WeRide's differentiation is our WeRide One, which is our universal autonomous driving technology platform. This platform integrate the software, hardware, and the cloud infrastructure into a reusable system that powers every product category from the L2+ assistant driving to fully driverless L4 vehicles. We have a few layers. You can see we use the same like the software core, and that's including like the one-stage end-to-end and that also include all the perception, prediction, planning, simulation layer. That can also be reused across different robotaxi, robobus, robotruck, and the ADAS solution.
This dramatically shorten the iteration cycles and reduce the marginal R&D cost per new product. One of our most like underappreciated strengths is ability to create a dual data flywheel between our L2+ and L4 system. Through the WePilot 3.0, our one-stage end-to-end L2+ solutions, we collect large scale human driving and the L2+ like driving like data that enhanced all the robustness like corner case understanding and policy learning across China and global. Those learning feed directly into the L4 system, which accelerating the model convergence and those long tail scenario performance. This is very strategically powerful because it combines the scale of the consumer ADAS data as well with the precision of the robotaxi standard type of like data quality.
Also one of the core standout capability is WeRide Genesis. This is our generative engineer neural environment for simulation intelligence in self-driving. What it means is like it's a huge like world model, like the simulation system. It can generate virtual driving environment in seconds, enable a rapid safe safety validation and also scenario expansion and in a very cost efficient way that can we can also do a lot of the edge case testing in the system. Rather than waiting for the real scenario to occur like in real world on public road, the system can like synthesize those scenarios at scale and to make sure to make the algorithm more robust.
It significantly reduced our on-road testing cost while improving statistically and systematically the safety validation side. We have a few layers of the safety on the statistical safety, on the behavior safety, and systematic safety for that. With all of the three layers, it creates a safety architecture not only for demos, but for very large scale like public deployment.
[Presentation]
Last year we upgraded our GXR to complete like a pre-installed version. From factory, from the OEM factory, it's already everything, all the sensors are mounted and the computing unit are 100% automotive-grade. Now the entire vehicle is 100% automotive-grade. On the computing side, it's 2,000 tops, which is powerful enough for the L4 algorithm. We have a thousand-line LiDAR to increase the point cloud, and also we extended the detection range to like a higher range of a few hundred meters.
Since now we are doing everything from the assembly line, the vehicle assembly time is shortened to under 10 minutes. This is the national ADAS competition we were just talking about. You can see WeRide is WeRide's ADAS is embedded into the Chery ET, this is the vehicle we take to the competition. We have won three times consecutively for this national ADAS competition. Even though the second or third place always changes in different stops, the Xinhua stop that happens in March, this is the Wenzhou stop in January, Taizhou stop in December. There's no competition in February, given this is Chinese New Year.
You can see the second and third place always changes, WeRide always remain to be number one, because our algorithm continuously to perform at its very best. I can show you some of the scores and how they get the scores and ranking. On there's like a compliance scores and efficiency scores, if you can looking at the number of intervention, usually for like a few hours of the competition, even though it's under the very complex like city scenario, WeRide usually have like very little intervention or no intervention, while others have a lot more intervention. The completion time for WeRide vehicle is usually like short, means like our efficiency of driving is high. This is some of the core highlight for our L2+.
Recently we just launched another vehicle with GAC, which will be shortened to be it's already released in the market, and we have received very good feedback for the Auto Plus like performance. Finally, let's talk about some of the financial validation. In financial year 2025, we delivered a record high total revenue of over RMB 685 million, representing a 90% year-over-year growth. More importantly, if you're looking at our robotaxi revenue, it's grown like 210% year-over-year to RMB 148 million. Increasing its contribution from 13% all the way to 22% on the total revenue. This is a major inflection point.
Robotaxi is becoming a larger and more meaningful revenue driver for the company. On the margin side, our margin always remains to be quite stable to get to an industry-leading 30%, while our net loss narrowed significantly. Operationally, I think we already talked about the, our expansion plan and our like current fleet size and expansion plan. We believe our net loss will continue to shrink for the next few years until we get to the break-even point. To conclude, we combine some of the rare quality in autonomous driving. We have the world-class full stack technology, and it's a universal platform that spans from L2+ to L4.
We also have like the broadest global deployment footprint, and rapidly we're increasing the unit economic for robotaxi. Okay, the first question is, quite a few new OEM L2+ peers are also working on robotaxi. What's your view? We see this as a positive validation of the long-term strategic importance of L4. We definitely welcome more players into the field. L4 robotaxi is a trillion dollar, like, total addressable market, so we understand the popularity for this market. That said, moving from, like, L2 plus assistant driving to a commercially scalable L4 robotaxi, it's not a linear extension.
The core difference for L4 is, it requires the highest precision in terms of the safety and also it's a lot more in terms of, like, the regulator approval, operational efficiency, and also to building the ecosystem on the local ground. It is really a system-level operational business. It requires the safety redundancy, requires, like, you have a great, like, remote operation, the fleet dispatch and regulatory approval, rider experience. You do have to reach to a city-level unit economics in order to really to expand the fleet. It's a completely different model from, business model from L2 +.
On the tech stack side, WeRide is the leader in L2+ technology, and we're also the leader in L4. If we're talking about the core metrics for L2+ and L4, now even we compare our own stack, there's a 250x-300 x difference in terms of the safety for our own L2+ stack and our L4. If we're talking about other, like, competitors, like L2+ to our L4 system, we think that the core, for the core, like, safety metrics, there's 1,000 x difference. While we still keep improving our L4, like, safety metrics, we still think there is room to improve by, like, 10x-100 x.
It's very hard for other L2+ players to catch up in what we have built for so many years, and we remain to be a leader in the L4 space. Yeah. While we are there are many player enter into the space, we believe the winner will be defined by the leading technology and leading safety record as well as the global operational execution and the unit economic scalability. For all of that, those points, WeRide remains strongly positioned. Yeah. The next question, could you walk us through the key cost deduction and tech innovation especially behind the upgraded GXR? Okay. The upgraded GXR, it's a robotaxi GXR.
It's a very important step in improving both the performance and the unit economic. From a cost perspective, we achieved approximately 15% BOM reduction. It's mainly driven by first of all, now it's a HPC upgrade. The compute architecture upgrade moved to 100% automotive-grade, and this significantly improved the compute density and integration efficiency. Also the higher performance per module lowers the overall system complexity and improves the maintainability and sustainability. We also have, like, a sensor optimization. We upgraded the sensor to the latest model provided by both Hesai and RoboSense.
That, the higher, like, LiDAR, like the thousand-line LiDAR architecture, increased our, the point cloud resolution by 17 x and extended the detection range to 600 m. This enables a long-range perception, and they can allow us to, like, to optimize overall sensor suite more efficiently on both, like, the performance and unit economic side. And also last but not least is on the manufacturing and assembly efficiency, since we are now moving everything to pre-installed robotaxi from factory, so we can now reduce the single vehicle assembly time to below 10 minutes. It's cost-wise it's much more efficient for us.
The beyond BOM is, like, only 15%, which is already a lot, but then the total TCO reduction, it is mainly due to the utilization and efficiency we just talked about. Like, we increased the remote assistance ratio from 1 to 10 to 1 to 40, now we have, like, a better fleet density and improved utilization of the fleet. Those are the key drivers behind the commercial commercial upgrade. Yeah. Okay. Thanks everyone for listening to the presentation. If you have any question, please reach out to ir@weride.ai. We are happy to answer the investors' question. Thank you.