Thank you, Operator. Hello, everyone, and welcome to RoboSense 2025 full year results earnings conference call.

The company's earnings results were released earlier today and are available on our IR website, www.ir.robosense.ai, the website of Hong Kong Stock Exchange, www.hkexnews.hk, as well as on newswire services. Today, you will hear from Mr.

Mark Qiu, our CEO, and Mr. Kelvin Law, our CFO, who will take you through the company's operational and financial results for the full year of 2025.

After their prepared remarks, Mark and Kelvin will be available to answer your questions. Before we continue, please note that the discussion today may contain certain forward-looking statements, which involve known and unknown risks, uncertainties, and other factors which are beyond our control, and may cause our actual results, performance, or achievements, or industry results to be materially different from any future results, performance, or achievements expressed or implied by the forward-looking statements.

In light of risks and uncertainties, the inclusion of forward-looking statements in this discussion should not be regarded as a representation by the board of directors of RoboSense that the plans and objectives will be achieved, and the shareholders and investors of the company should not place undue reliance on such statements. RoboSense does not assume any obligation to update any forward-looking statements, except as required under applicable law.

Also, please note that some of information to be discussed, including non-IFRS financial measures, are not required by or presented in accordance with IFRS. The IFRS financial measures and information reconciling these non-IFRS financial measures to RoboSense financial results prepared in accordance with IFRS are included in RoboSense annual results announcements, which had been posted on the company's IR website at www.ir.robosense.ai and the website of Hong Kong Stock Exchange, www.hkexnews.hk.

Finally, as a reminder, this conference call is being recorded. In addition, a webcast of this conference call is available on RoboSense Investor Relations website.

Now, I will turn the call over to Mr. Mark Qiu, the CEO of RoboSense.

Thank you, Grace. Hello, everyone.

Thank you all for taking the time to join RoboSense 2025 fourth quarter and full year earnings conference call. We delivered exciting results for the fourth quarter of 2025, and through our efforts over the past year, we have laid a solid foundation for our next phase of development in 2026.

Today, I would like to walk you through three key areas in a more structured way: First, in 2025, RoboSense successfully reached an operational turning point, achieving its first ever quarterly profit in the fourth quarter of 2025. Second, I will review our core strategy, deployment, and key achievements in 2025.

Third, I will share our outlook for 2026, including how we utilize the solid foundation laid in 2025 to achieve more predictable and sustainable growth and create sustained value for our shareholders and partners. Let me begin with the financial results, which I know are top of mind for everyone on the call.

I'm very pleased to announce that in the fourth quarter of 2025, we achieved our first quarterly profit since the company's establishment, making a critical operational turning point. In the fourth quarter, RoboSense recorded total revenue of RMB750.7 million, representing a year-on-year increase of 46.1% and a quarter-on-quarter increase of 84.4%.

Total LiDAR sales volume reached 459,600 units, up 183.2% year-on-year and 147.6% quarter-on-quarter. In terms of profitability, our overall gross margin improved to 28.5%, up 6.4 percentage points year-on-year and 4.6 percentage points quarter-on-quarter.

Operating profit reached RMB130.1 million, and net profit amounted to RMB103.7 million. Behind this performance were three structural drivers that came together and were fully realized in the fourth quarter.

First, our digital products entered into the stage of large-scale delivery. In the fourth quarter of 2025, the total shipments of LiDAR products for ADAS applications reached approximately 238,400 units, up 54.8% year-on-year.

More notably, sales of LiDAR products for robotics and others experienced explosive volume growth, with total shipments reached approximately 221,200 units, surging 2,565.1% year-over-year and 523.1% quarter-on-quarter, becoming the core engine of the quarterly growth. Second, our revenue contribution mix continued to optimize, with the robotics business achieving explosive growth.

In the fourth quarter, revenue from sales of LiDAR products for robotics and others rose to RMB346.7 million in the fourth quarter of 2025, up 427.5% year-on-year and 133.4% quarter-on-quarter, to become the primary driver to the quarterly revenue growth, significantly enhancing the company's overall gross profit margin and optimizing our revenue contribution mix. Third, cost reduction benefiting from our self-developed chips reflect in the income statement.

Gross margin of LiDAR products for ADAS applications rose to 22% in the fourth quarter of 2025, up 5.7 percentage points year-on-year and 3.9 percentage points quarter-on-quarter. The sustained gross margin improvement validates the effectiveness of our structural cost optimization strategy by using our proprietary chips.

Gross margin of LiDAR products for robotics and other remains stable at 37.3%, maintaining at a comparable high level amidst rapid volume expansion and ASP change. Looking at the full year, we achieved a total revenue of RMB1.941 billion, up 17.7% year-over-year.

Total annual LiDAR product sales volume reaches approximately 912,000 units, up 67.7% year-over-year. Overall gross margin improved to 26.5% in 2025, an increase of 9.3 percentage points from 2024, with gross profit reaching 81.3% year-over-year.

Net loss narrowed to RMB145 million in 2025, and adjusted net loss further narrowed to RMB53.5 million. Our growth was not driven by a single factor, but structurally supported by technological and product leadership, optimized revenue contribution mix, and fully unleashed supply chain capabilities, and etc.

Next, let me walk you through our core strategy deployment and key achievements in 2025. Frankly speaking, 2025 was an extremely challenging and high-productive year for RoboSense.

As many of you may know, starting from the first quarter of 2025, our cooperation with two major customers, the H Company and X Company, was suspended. These two customers had represented a substantial portion of our ADAS sales volume in 2024.

This made 2025 a highly unusual and critical year for the company. At the same time, we saw highly positive signals in the LiDAR industry.

On one hand, intelligent driving technologies and commercial applications gradually formed a reinforcing cycle, driving a surge in LiDAR demand. On the other hand, the rapid development of robotics and physical AI leads to the emerging of different LiDAR application scenarios, and each application scenarios is expanding at a scale comparable to the size of the automotive market.

We firmly believe that 2025 marked an inflection point for LiDAR industry and a critical year for our business development. It represents a key window for the company to expand our business to other markets and the optimal time to optimize our business structure so as to mitigate performance volatility from overreliance on a single market or one or two major customers.

In 2025, we lead the LiDAR industry into the digital era. In the first quarter of 2025, leveraging our long-accumulated digital architecture, we launched a full portfolio of digital LiDAR products, including MX, M4, E1R, and Airy, and etc., officially initiating the paradigm shift of LiDAR from analog to digital architecture.

This technological transformation is revolutionary, much like the transition from the film camera to digital camera. In terms of resolution, traditional analog architecture using discrete component struggle to mass-production LiDAR products exceed 128 beams.

By contrast, digital LiDAR can easily surpass 1,000 beams while maintaining high-cost performance. This fundamentally overturns the conventional perception that LiDAR has limited performance ceiling and prohibitively high cost.

Digital LiDAR has therefore become highly attractive for both advanced automotive autonomous driving function and a wide range of robotics applications. In the second quarter of 2025, we start our market expansion activities.

To date, our digital LiDAR products have achieved encouraging progress across robotaxi, ADAS, and general robotics segments, securing a large number of prospective orders. In the third quarter of 2025, alongside with the market expansion strategy, we also fully focus on preparing for mass production of digital LiDAR, including production capability, capacity expansion, and supply chain optimization, making final preparation for large-scale product delivery expected to become occurred in the fourth quarter of 2025.

During this quarter, a series of our in-house chips obtained the automotive-grade certification and finalized the product reliability verification tests. Production lines for digital LiDAR products launched earlier in the year began trial production operations.

As I mentioned during the third quarter earnings call, the delivery performance in the first three quarters largely reflect that the final phase of our legacy analog era products. The mass production, readiness, and order backlogs of digital LiDARs are the key factors to define and determine our future.

Then in the fourth quarter, all these efforts came to fruition. Digital LiDAR product for the first time officially start large-scale delivery to customer.

In the automotive sector, both M4 and MX products entered mass production. Notably, Voyah Courage and IM Motor become the first SOP vehicle model equipped with LiDAR featuring more than 500 beams, achieving Level 4 LiDAR perception capabilities.

In the general robotics sector, we also saw an explosive growth. E1R and Airy Light both entered into large-scale mass production stage.

We help customer developing high-performance solid-state LiDAR technology at scale in different robotics applications. In the general robotics sector, we achieved a quarterly product shipments exceeding 200,000 units in the fourth quarter of 2025.

Technological breakthrough, market expansion, and production capability readiness were all realized in the fourth quarter of 2025, resulting in our quarterly LiDAR shipment volume hit a record high. We complete the full cycle from strategic layout and commercial [harvest?]

. Now that I have reviewed the core development quarterly by quarterly, and let me summarize our key achievements in 2025: In the robotaxi segment, our M4 ultra-long-range main LiDAR and E1 blind-spot LiDAR combination, relying on their industry-leading performance and high level of product and technology sophistication, quickly become the preferred solution for robotaxi customers.

In the analog era, our market share in this sector was approximately 10%. In the digital era, we have established cooperation with more than 90% of the world's core robotaxi and robotruck players, including Baidu Apollo Go, DiDi Autonomous Driving, WeRide, Pony.ai, and leading North American Level 4 players, and etc.

We expect that our digital LiDAR will serve as the core safety sensors on the robotaxi players' next-generation mass-production and commercially operated vehicles. We also strengthen our ecosystem position by joining NVIDIA's DRIVE AGX Orin, DRIVE, and Omniverse ecosystem, building a comprehensive network covering mainstream automakers, mobility service providers, and Level 4 automotive autonomous driving companies, laying a solid foundation of LiDAR adoption for the large-scale deployment of robotaxi fleets.

In ADAS, for the Level 2 market, our digital MX with 192 beams, high-density point cloud, and a 300 meters detection range has become the mainstream industry configuration. For high-level Level 3 intelligent driving, our M4 has become the industry-only mass-production digital LiDAR with more than 500 beams.

Today, our digital platform products have secured design wins for over 100 vehicle models. In the domestic Chinese market, in addition to our long-term partners such as BYD and Geely, we have obtained more vehicle model design wins from other OEMs due to the industry-leading technology and high performance of our digital LiDAR products.

We add new names in our OEMs customer list, including a prominent emerging EV brand, Great Wall Motor, Leapmotor, and SERES. Beyond main LiDARs, multiple OEMs will, for the first time, install our digital fully solid-state blind-spot LiDAR in their SOP vehicle models this year.

Our international automotive business achieved comprehensive progress, covering all major automotive markets in Asia Pacific, Europe, and North America. Today, we have secured vehicle model design wins from 14 overseas and Sino-foreign joint venture OEMs.

According to the 2025 market share of LiDAR suppliers for joint venture automotive brands released by Shuju Bang, RoboSense ranked first with a market share of over 70%. In the Asia Pacific market, projects with Japan's top three automakers progress steadily, and we continue to obtain new vehicle model design wins from Sino-foreign joint venture of leading Japanese OEMs.

Among those European Sino-foreign joint venture OEMs, we have secured new vehicle model design wins from the joint venture of a leading European luxury automakers, as well as multiple new design wins from several Sino-foreign joint ventures established by the large European automotive group. In the North American market, we have obtained exclusive design wins for multiple vehicle models from several OEMs.

In 2025, overseas revenue increased by more than 90% year-on-year, marking a transit from being a leader in China to become a global front-runner. As of the end of 2025, in the automotive ADAS sector, we secured a total of 183 vehicle model design wins from 34 OEMs and Tier 1 suppliers.

In the robotics and general robotics sector, by using our core digital LiDAR products, including E1, E1R, Airy, Flair, etc., we have developed various perception solutions applying in diverse robotics scenarios. In lawn mower robots market, in addition to Mammotion, we secure an exclusive design win with Navimow, a brand owned by Segway-Ninebot.

Notably, we recently obtained an exclusive design win from a leading cleaning robot brand for its new lawn mower robotics business, with our digital LiDAR products delivery scheduled to commence within this year. In unmanned delivery vehicle sector, we now serve more than 90% of leading customers in the industry.

Our digital LiDAR products have already been deployed at scale on the next-generation unmanned delivery vehicle deployed by these global industry leaders, including Neolix, Zelos, Rhino.ai, JD, Meituan, Cainiao, Minieye, Coco Robotics, and etc. In the embodied robotics sector, the demand for LiDAR products is rising rapidly.

We have already established partnerships with nearly 50 leading customers, including AgiBot, Unitree, and EngineAI, to jointly accelerate the deep adoption of robotics across a wide range of application scenarios. We are also expanding our LiDAR supplier business into more demanding applications, such as autonomous mining trucks and low-altitude drones.

According to public data issued by multiple third-party research house, we achieved 3D LiDAR products sales volume number one in the robotics sector in China. We have also been awarded with LiDAR product sales volume number one in global lawn mower robotics sector, in commercial cleaning robotics sector, in China humanoid robot sector, in China unmanned delivery vehicle sector, and in China embodied robotics sector.

The business of LiDAR products for robotics and other has truly become the second major driver for the company's growth. After reviewing the past, let's look ahead to 2026.

We are highly confident with our business development in the year ahead. We will continue to validate our growth through the actual quarterly performance.

We will convert the competitive advantage in technologies, customer base, and product performance accumulated in 2025 into tangible results in 2026 and beyond. In terms of technology, we will continue to further strengthen our leadership in digital era.

The superiority of digital architecture has gained industry-wide recognition. We believe that chips development will become the next core focus of competition.

We will further deepen and advance our self-developed chips capabilities and continuously launch new proprietary chips to create generational product differentiation. Meanwhile, we will continue to explore the opportunities for applying our products into more different market, from automotive to general robotics, from large enterprise customer to small business customers, and eventually to consumer applications.

Our goal is to make LiDAR a ubiquitous product category, much like cameras today. In terms of market segments, our business will no longer rely on any single market segment or a small number of key customers.

We expect the revenue from ADAS business sector will maintaining a strong growth. At the same time, revenue from robotics and other business sector will enter into another year of breakout growth, and we anticipate that RoboSense will again lead the industry in sales volume across the general robotics market in 2026.

Ultimately, we target an approximately 50-50 revenue mix between ADAS and robotics business. In terms of production capability, we have completed on production layout with an annual capacity of 4 million LiDARs units, which will sufficiently match the ramp-up needs of both ADAS and robotics business and fully ensure large-scale product deliveries in 2026.

In terms of product deliveries, we expect the sales volume of our LiDAR products in both ADAS and robotics applications to increase by at least 2 to 3 times year-over-year. We are fairly confident in this level of growth that we will achieve.

We remain firmly believe that LiDAR products are our core foundation. Automotive and general robotics business will serve as our dual growth engines, while innovative new business represent our long-term growth driver.

In 2026, our strategy is clear: RoboSense is a robotics company. We will continue to expand our boundaries of physical AI, firmly securing our position in robots ecosystem by developing key components and core capabilities in mobility and manipulation functions.

Upholding our mission, safer world, smarter life, we will continue to invest in building a stronger technological moat and create long-term value for society, our partners, and shareholders. Once again, thank you all for your support.

Next, I'd like to hand it over to our CFO, Kelvin Law, who will provide a detailed breakdown of the financial data. After that, we'll open the floor to your questions.

All right. Thank you, Mark.

Now I would like to take you through the 2025 full year results financial highlights. Total revenue increased by about 17.7% to RMB1,941 million for the year ended December 31st, 2025, from RMB1,648.9 million for the year ended December 31st, 2024.

The increase was primarily due to the increase in sales of product in 2025. Our revenue from sales of product increased by about 18.4% to RMB1,815.8 million in 2025 from RMB1,533.7 million in 2024, primarily due to the increase in sales of product for robotics and others, partially offset by the decrease in sales revenue of products for ADAS applications.

The total number of our LiDAR products sold increased by about 67.6% to approximately 912,000 units in 2025 from approximately 544,200 units in 2024. In 2025, despite the increase in number of LiDAR products sold for ADAS application by about 17.2% to approximately 609,000 units in 2025 from approximately 519,800 units in 2024, our revenue from sales of LiDAR products for ADAS application decreased to RMB1,105.9 million in 2025 from RMB1,335.3 million in 2024, representing a year-on-year reduction of about 17.2%.

The reduction in revenue from LiDAR products for ADAS application was mainly due to the decrease in the average unit price of product for ADAS applications to approximately RMB1,800 per unit in 2025 from approximately RMB2,600 per unit in 2024, resulting from the increase in sales of our lower-priced MX and EM series LiDAR products in 2025. Our revenue from sales of products for robotics and others increased significantly to RMB709.8 million in 2025 from RMB198.5 million in 2024, representing a year-on-year increase of about 257.7%.

The total number of LiDAR products sold for robotics and others increased significantly by 1,141.8% to approximately 303,000 units in 2025 from approximately 24,400 units in 2024, whilst the average unit price of product decreased to approximately RMB2,300 per unit in 2025 from approximately RMB8,100 per unit in 2024. In 2025, the sales of our lower-priced new E1R and Airy LiDAR products to robotics customers, especially lawn mower manufacturer, increased significantly.

In 2024, most of our products sold in this category were those mechanical LiDARs, such as Helios and Bipoo series, which had higher average unit price. Our revenue from sales of solutions decreased by about 20.8% to RMB77.6 million in 2025 from RMB98 million in 2024.

Despite the increase in number of delivered solution projects to 98 projects in 2025 from 331 projects in 2024, the average selling price per project increased to approximately RMB792,200 in 2025 from approximately RMB296,000 in 2024, primarily attributable to the increase in demand from customers for more customized perception-related solutions. Our revenue from provision of services and others increased by about 177% to RMB47.6 million in 2025 from RMB17.2 million in 2024, primarily due to the number of completed technology service projects has been increased in 2025.

Our cost of sales increased by about 4.5% to RMB1,426.9 million in 2025 from RMB1,365.3 million in 2024, primarily driven by the increase in sales of product in 2025. Our gross profit increased by about 81.3% to RMB514.2 million in 2025 from RMB283.6 million in 2024.

Our gross profit margin improved by about 9.3 percentage points to 26.5% in 2025 from 17.2% in 2024. Our overall gross profit margin was largely affected by the changes in the sales contribution from different product categories.

The increase in overall gross profit margin was mainly attributable to the gross profit margin improvement of both our LiDAR products for ADAS applications and LiDAR products for robotics and others. For LiDAR products for ADAS applications, the gross profit increased by about 17.7% to RMB210.9 million in 2025 from RMB179.2 million in 2024.

The gross profit margin for this product category improved to 19.1% in 2025 from 13.4% in 2024, increased by about 5.7 percentage points. The gross profit margin improvement was primarily attributable to the decrease in raw material procurement cost and the adoption of our in-house developed SoC processing chips, which have lower cost as compared to the FPGA chips acquired from third-party suppliers.

For sales of LiDAR products for robotics and others, the gross profit increased by about 311.4% to RMB281.6 million in 2025 from RMB68.5% in 2024. The gross profit margin for this product category increased to 39.7% in 2025 from 34.5% in 2024, increased by about 5.2 percentage points.

This was primarily attributable to the reduction of raw material procurement cost and production overheads resulting from the increase in scale of production. For provision of LiDAR perception solutions, the gross profit decreased by about 24.9% to RMB39.5 million in 2025 from RMB52.7 million in 2024.

The gross profit margin for this product category decreased to 50.9% in 2025 from 53.8% in 2024, dropped by about 2.8 percentage points. The decrease in gross profit and gross profit margin was mainly attributable to the increase in raw material procurement cost for customized solution projects.

For our provision of services, we recorded a gross loss of RMB16.8 million and RMB17.9 million in 2024 and 2025 respectively. The gross loss margin for this product category improved to 37.7% in 2025 from 97.6% in 2024.

Our R&D expenses increased by about 5.1% to RMB646.7 million in 2025 from RMB615.4 million in 2024. The increase was mainly due to: number one, the increase in employee benefit expenses by about RMB21.6 million, which were mainly attributable to the increase in employee remuneration package and share-based compensation; and number two, the increase in raw material consumable, design and development expenses by about RMB9.5 million incurred in developing new and more advanced products.

Our R&D expenses excluding share-based compensation as a percentage of revenue reduced to 29.9% in 2025 from 33.6% in 2024. Our sales and marketing expenses increased by about 16.1% to RMB128.1 million in 2025 from RMB110.3 million in 2024.

The increase was primarily due to the increase in employee benefit expenses by RMB18.5 million, which was mainly attributable to the increase in employee remuneration package and share-based compensation. Our sales and marketing expenses excluding share-based compensation as a percentage of revenue reduced to 6% in 2025 from 6.1% in 2024.

Our G&A expenses increased by about 2.5% to RMB168 million in 2025 from RMB164 million in 2024. The increase was primarily due to: number one, the increase in employee benefit expenses by RMB18.4 million, which was mainly attributable to the increase in employee remuneration package and severance payment, partially offset by; number two, the decrease in professional service fee.

Our G&A expenses excluding share-based compensation and listing expenses as a percentage of revenue reduced to 8% in 2025 from 9.1% in 2024. Net impairment losses on financial assets decreased by about 90% to RMB1.2 million in 2025 from RMB11.8 million in 2024.

The decrease was mainly due to the decrease in provision for impairment on trade receivables. Other income increased by about 146.8% to RMB129.6 million in 2025 from RMB52.5 million in 2024.

The increase was primarily due to the increase in government grants, interest income, and a lump-sum monetary compensation received from one of our customers in 2025. Other gains or losses improved from a loss of RMB18.8 million in 2024 to a gain of RMB115.9 million in 2025, representing an improvement of RMB134.7 million.

The increase was primarily due to the increase in fair value gains on financial assets at fair value through P&L in 2025. Net finance income decreased by about 11.1% to RMB88.6 million in 2025 from RMB99.7 million in 2024.

The decrease was primarily due to the decrease in interest income from cash and cash equivalents. Share of net loss or profit of associates accounted for using equity method turned from a net profit of RMB10.5 million in 2024 to a net loss of RMB13.4 million in 2025, representing a deterioration of RMB23.9 million.

The share of net loss was primarily due to the operating loss incurred by one of our associates in 2025. Impairment loss of an associate accounted for using equity method increased to RMB16.5 million in 2025 from nil in 2024.

The loss was primarily due to the impairment of goodwill incurred by one of our associates in 2025. Our net loss decreased by about 69.9% to RMB145 million in 2025 from RMB481.8 million in 2024.

All right, this finished my report about the 2025 full year results. Please allow me to continue to present to you briefly the fourth quarter of 2024 financial results highlights.

In Q4 2025, our total revenue increased by about 46.1% to RMB750.7 million in 2025 Q4 from RMB513.9 million in 2024 Q4. Our revenue from the sales of product increased by about 53% to RMB707.6 million in 2025 Q4 from RMB600 from RMB462.4 million in 2024 Q4.

The total number of our LiDAR products sold increased by about 183.2% to approximately 459,600 units in 2025 Q4 from approximately 162,300 units in 2024 Q4. The number of LiDAR products sold for ADAS application increased by about 54.8% to approximately 238,400 units in 2025 Q4 from approximately 154,000 units in 2024 Q4.

Revenue from sales of our LiDAR products for ADAS application decreased to RMB360.9 million in 2025 Q4 from RMB396.7 million in 2024 Q4, representing a year-on-year reduction of about 9%. The decrease in the average unit price of product for ADAS application to approximately RMB1,500 per unit in 2025 Q4 from approximately RMB2,600 per unit in 2024 Q4.

Revenue from sales of product for robotics and others increased significantly to RMB346.7 million in 2025 Q4 from RMB65.7 million in 2024 Q4, representing a year-on-year increase of about 427.5%. The total number of LiDAR products sold for robotics and others increased significantly by about 2,565.1% to approximately 221,200 units in 2025 Q4 from approximately 8,300 units in 2024 Q4.

The average unit price of products decreased to approximately RMB1,600 per unit in 2025 Q4 from approximately RMB7,900 per unit in 2024 Q4. Revenue from the sales of solution decreased by about 63.9% to RMB14.5 million in 2025 Q4 from RMB40.2% in 2024 Q4.

The number of delivered solution projects decreased to 36 projects in 2025 Q4 from 145 projects in 2024 Q4. The average selling price per project increased to approximately RMB403,100 in 2025 Q4 from approximately RMB277,500 in 2024 Q4.

Revenue from the provision of services and others increased by about 153.6% to RMB28.6 million in 2025 Q4 from RMB11.3 million in 2024 Q4. Cost of sales increased by about 34.2% to RMB537.1 million 537.1 million in 2025 Q4 from RMB400.1 million in 2024 Q4.

Gross profit increased by about 87.7% to RMB213.6 million in 2025 Q4 from RMB113.8 million in 2024 Q4. The overall gross profit margin improved to 28.5% in 2025 Q4 from 22.1% in 2024 Q4, increased by 6.4 percentage points.

For our LiDAR products for ADAS application, the gross profit increased by about 23% to RMB79.3 million in 2025 Q4 from RMB64.5 million in 2024 Q4. The gross profit margin for this product category improved to 22% in 2025 Q4 from 16.3% in 2024 Q4, increased by about 5.7 percentage points.

For our sales of LiDAR products for robotics and others, the gross profit increased by about 345.1% to RMB129.4 million in 2025 Q4 from RMB29.1 million in 2024 Q4. The gross profit margin for this product category decreased to 37.3% in 2025 Q4 from 44.2% in 2024 Q4, dropped by about 6.9 percentage points.

For our provision of LiDAR perception solutions, the gross profit decreased by about 79.1% to RMB4.4 million in 2025 Q4 from RMB21.2 million in 2024 Q4. The gross profit margin for this product category decreased to 30.3% in 2025 Q4 from 52.7% for 2024 Q4.

Our R&D expenses increased by about 5.2% to RMB to RMB158.2 million in 2025 Q4 from RMB150.3 million in 2024 Q4. The increase was primarily due to increase in employee benefit expenses by RMB10 million and the increase in raw material consumable expenses by 12.9 million incurred in developing new and more advanced products.

Our sales and marketing expenses increased by about 35.3% to RMB36.3 million in 2025 Q4 from RMB26.8 million in 2024 Q4. The increase was primarily due to increase in employee benefit expenses by RMB8.1 million.

Our G&A expenses decreased by about 7.8% to RMB41.8 million in 2025 Q4 from RMB45.3 million in 2024 Q4. The decrease was primarily due to the decrease in professional service fee.

Net impairment losses on financial assets decreased by about 44% to RMB6 million in 2025 Q4 from RMB10.8 million in 2024 Q4. Other income increased by 484% to RMB68.4 million in 2025 Q4 from RMB11.7 million in 2024 Q4.

Other gain or loss improved from a loss of RMB44.3 million in 2024 Q4 to a gain of RMB90.5 million in 2025 Q4, representing an improvement of RMB134.8 million. Net finance income decreased by about 21.3% to RMB17.2 million in 2025 Q4 from RMB21.9 million in 2024 Q4.

All right, finally, our result improved from a net loss of RMB131.1 million in 2024 Q4 to a net profit of RMB103.7 million in 2025 Q4, representing an improvement of RMB234.8 million. All right, this concludes my financial highlights.

Operator, we are ready for questions.

Thanks for taking my question. This is Xiaoyi from Jefferies.

I've got one question for management. Could you walk us through the customer mix of your ADAS business in '26, and how should we think about the shipment expectations for this year, particularly contributions from BYD and Geely?

Also, maybe a little bit more color about the new customers such as Leapmotor, Great Wall Motor, Xiaomi, and others. Thank you.

Thank you for your question, Xiaoyi. I'm glad to share with you further about our business in ADAS.

As mentioned in the earnings call just now, the launch of digital LiDAR marks an extremely significant milestone, and it has spurred industry adoption, boosted penetration rate, and further accelerated the arrival of Level 3. The fourth quarter of 2025 marks our first mass production quarter, and we delivered strong results.

2026 will be our first full year of mass production for digital LiDAR, and we expect this momentum will continue. In ADAS sector, on the one hand, we have secured more customers, which will further reshape our customer mix in 2026, making our revenue more resilient and better able to withstand risks.

In the other hand, we are doing our very best to work together with some top OEMs, target to further drive the widespread adoption of LiDAR across the industry. Since evolving to rapidly, the growth logic, project timelines, and opportunity points are different across different customers.

In order to give a more responsible answer, I believe I have to break down our key customers individually. First, the customer G and B.

Unfortunately, I'm not able to disclose their name directly, Xiaoyi. About G and B, both of them have a very strong sales volume.

As industry leader, they are setting the trend. We're going all out with them to get LiDAR equipped across all models and price range, and we are seeing very positive signals.

Since this is really a big deal, timing and progress may still change. Therefore, our guidance has to be relatively conservative, but we keep the possibility of upward adjustment.

Our current forecast for each of them, these two company, is about 450,000 to 550,000 units in 2026. Customer W is a milestone new client for us.

Their vehicle come with a relatively high price tag, they have very high requirements for LiDAR performance, and they also have a strong intention to adopt LiDAR on a large scale. Since we launch our digital LiDAR, we basically been selected for all their new models.

Several of their high-profile models with supplied now has successfully entered the ramp-up phase. We expect to start mass production and delivery in the second quarter, with more models in the pipeline.

Some of their main model haven't reached their facelift cycle yet. Other companies' LiDAR are still selling, we don't expect to have a dominant share with this customer in 2026.

Therefore, the estimate volume for this year is around 150,000 to 200,000 units. Likewise, we won both customer L and customer X as new clients last year with our outstanding digital LiDAR.

It is really exciting because they are both phenomenal Chinese EV new players. For customer X, the platform we won so far are expected to enter mass production around Q4.

Since it's only one quarter, we won't give very high guidance. We expect around 50,000 to 100,000 units.

I guess we can see higher expectation for 2027. Customer L's, their project was supposed to start mass production in March, but it's be delayed for reasons that beyond our control.

We're only forecasting 50,000 to 100,000 units for this project, too. Here's the important point: joint venture and overseas business are becoming a really strong growth area for us.

Their market share has dropped a bit, but joint venture still make up nearly half of Chinese annual car sales, and they're going all in for the next stage for autonomous driving. Last year, ZhuoSi Auto Research 2025 report showed RoboSense was number one in LiDAR supplier for joint venture brands with over 70% market share.

This year, that advantage is continue, and we build up a considerable order volume. Based on our current ADAS outlook, four of our top 10 volume customer are joint ventures.

They will gradually become major contributor starting this year, and we expect around 200,000 units to 250,000 units this year. That's all, Xiaoyi.

Thank you for the detailed explanation. No further questions from me.

Thank you.

Hi, management. This is Shiwen Li from CICC.

Firstly, congratulations on the break-even and solid development in the past quarter. I have a further question related to the robotics business, which has became a key growth driver in 2025, especially quarter four.

Could you share the shipment guidance for robotics LiDAR in 2026? If possible, could you break it down by subsegments, such as lawn mower, robovan, industrial safety applications, and others?

Thanks.

Happy to. Thank you, Shiwen.

I'm glad to notice a great deal to notice that, too. As Jensen Huang put it, the AI revolution of the past 10 years has mostly happened on screen, but in the next decade, AI will truly move into the physical world.

Elon Musk has also said that humanoid robots market would one day be 10 to 100 times the size of the auto industry. We place very strong emphasis on business development in the robotics field.

We were probably the LiDAR company that invested the most in this sector. Actually, LiDAR is already widely used in many robot applications, and is now ramping up quickly.

Benefited from our digital LiDAR and our heavy investment in robotics over the past few years, we have achieved unparalleled results. In the fourth quarter of last year alone, our robotics shipment hit 230,000 units.

This is a historical moment, and for our company and for the entire entire industry, and this is truly an extraordinary achievement. Let me emphasize, just like in automotive application, Q4 marks only the first quarter of mass production for our digital LiDAR.

In 2026, we're set to carry forward and scale it to new heights. This year, we expect our robotics LiDAR sales will triple last year's volume, bringing us to near nearly 1 million units.

To be precise, we're targeting a sales volume of 800,000 units to 1 million units. As you mentioned, lawn mowing robots are currently one of our largest, most matured, and highest volume segment.

It was also the core scenario that saw the most significant volume growth in 2025. Thanks to the leading edge of our digital LiDAR products, we have secured a majority share among the top-tier customers.

Like the Ninebot, Mammotion, and I also mentioned today that we have recently secured an exclusive design win for another leading cleaning robot brand for their lawn mower segment. Deliveries will also start within this year.

For 2026, we expect shipments in the lawn mowing segment to reach 450,000 to 600,000 units. Autonomous delivery is characterized by rapid growth and broad customer coverage.

We already work with over 90% of the industry's leading players, including, like I mentioned in the earnings call, the Neolix and Zelos, Rhino.ai, JD, Meituan, Cainiao, etc., and all of which are using our solution. As robovan moves into mass production and deployment, the industry gearing up to boost operational speed.

Products requirements are increasingly aligned with those for robotaxi. Our EM platform and E platform fully meet this demand.

As it achieved a very successful result in the robotaxi era. Each vehicle is likely to be equipped with two to four units.

We expect shipments in this industry can be reached 100,000 to to reach 100,000 to 150,000 units. Other segment like the industrial safety, the cleaning, sanitation, and other scattered.

This field has a wide range of scenario and varied customer type. Altogether, we expect shipments for around 150,000 units this year.

While the current total volume of embodied intelligence is still modest, but it holds very important strategic position, much like the automotive or other robotics sectors several years ago. We are investing earlier to prepare for the upcoming market explosion.

Our LiDAR has been adopted by nearly 50 key global customers. Our AC series and hand-eye coordination solution are now undergoing scenario validation with leading customers, making this a key strategic focus for our future.

In short, beyond our automotive business, our robotics segment our robotics segments has clear clearly become an important growth driver. In 2025, our strategy of multi-application growth and diversified customer and customer and business structure has paid off.

Compared with focusing on a single market and a few major customers, I believe this will make our development more healthier and more sustainable. Shuwen.

Thank you so much, Mark. That's very helpful, and that's all my question.

Thank you. Thanks for taking my question.

This is Cherry from JPMorgan. Just wanted to say congrats for achieving break-even in Q4 last year, and also all the positive progress mentioned earlier.

I actually just have one question that we kind of noticed that both RoboSense and peers have recently announced orders from Baidu's Apollo Go and also WeRide. Could you kindly share more details on these orders?

Also, I think for robotaxi, it seems like RoboSense has captured most of customers and also orders. I'm just curious, when do you expect to see these orders to ramp up?

Thank you.

All right. Thank you, Cherry.

Thank you very much for recognizing RoboSense achievement in robotaxi last year, and I'm really proud of it. As during the analog architecture area, RoboSense only held around 10% market share because we focus on developing our next-generation platform, and it even led to doubts about our technical capabilities.

Everything changed with the launch of our digital LiDAR. We have built deep partnership with over 90% of the major players in the market: DiDi, Baidu, WeRide, and Pony.ai.

The industry's leading companies you mentioned are all among them. This means a lot to us.

Now I'd like to share more details about our cooperation. For Baidu Apollo Go and DiDi, we have secured exclusive design win for their next-generation vehicle models, providing the full solution of main LiDAR plus the blind-spot LiDAR.

With WeRide, we supply both the top main LiDAR and blind-spot LiDAR, while a competitor will provide their mid-range solution. In other words, we have secured orders from most leading customer for products at the most critical sensing position with the highest performance requirements.

This reflect our generational advantage in technical capabilities. I believe these top players will all put their next-generation solution into mass production this year, and our LiDAR will soon be featured in this next-generation vehicles.

A single robotaxi is typical equipped with 6 to 10 LiDAR units this year. Leading companies are all advancing their iteration of their new models, making 2026 to 2027 a critical window.

As these vehicles are gradually put into operation, our orders will be fulfilled accordingly. Cherry.

Understood. That's very clear.

Thank you, Mark, and expect to see more achievements this year. That's all from me.

Thank you.

Thank you.

Hi. Management, thank you for taking my questions.

This is Joey Yung from BofA Securities, and again, congratulations on the strong results. My question is on the high-resolution LiDAR.

We see that Huawei recently released an 896-beam high-resolution LiDAR. We also see RoboSense M4 can be the maximum number of beams to reach around 700.

I'm wondering if the maximum number can customize to 1,000 or even a higher level? How do you see the trend of LiDAR evolving towards higher beams?

Thank you.

Yes, Joey, and it's an extraordinary insight. High-performance LiDAR is a highly certain trend going forward, and here are two key logics behind this: The first one is autonomous driving has real demands for performance upgrades.

From urban NOA to full-scenario unmanned driving, the perception system needs denser point clouds to detect more distant obstacles and low users at complex intersections. This is a strict requirement for safety redundancy.

The second is our digital architecture can deliver high-channel LiDAR at a reasonable cost. Analog architecture can achieve high-channel, too, but they come with huge drawbacks in size, power consumption, and cost.

In other words, the higher the channel count, the more and even exponentially more advantage digital architecture has. The H Company launch of the 896-channel LiDAR is a perfect testament to this trend.

We made our bet on digital architecture much earlier, and have built far deeper expertise. In fact, we introduced the M4 in 2025, the world mass-production-ready 1,000-channels LiDAR, and it can be customized from 520 beams to 2,160 channels.

I believe that in the high-performance race, we will go all out to be the definer, not a follower. Starting from Q4 last year, we have Zeekr and IM Motor become the industry's first volume production customers to launch vehicle equipped with LiDAR with more than 500 channels.

To date, our M4 platform has secured design wins for more than 10 vehicle models across over six automakers. As the market increasingly recognized the advantage of high-channel LiDAR, and as more models migrate from Level 2 to Level 3 functionality, this number continue to grow.

Joey.

Thank you very much. That's very clear, and we look forward to another stronger year.

Thank you.

Thank you.

Hi, management. This is Zhang from Huatai Securities.

Thanks for the opportunity. I have two questions.

First is about cost. In the trend towards high-performance LiDAR, do the numbers of beams and cost still have an almost proportional relationship?

For example, would the cost difference between a 100-beam LiDAR and a 500-beam or even a 1,000- or 2,000-level beam LiDAR make a big difference? The second question is about new products.

We've seen that RoboSense launched a series of new products last year. Will you continue to launch new products this year, and what's the new direction?

Thank you.

Thank you for your question, Zhang. Generally speaking, higher channel counts do require more advanced technology and usually come with higher cost.

This is exactly where the real advantage of digital architecture begin to show. Under the traditional analog architecture, increasing channel count mainly means pile up more discrete components.

Digital LiDAR is different. The channel counts upgrade is driven much more by chip design, and it follow the Moore's Law.

In the other words, higher channel count does not mean cost raise in a one-to-one linear way. In fact, it's much more closer to how the evolution of cameras.

Cameras move from thousands, hundreds of thousands of pixels to millions of pixels. Performance keep improving over time, while cost keep coming down.

Of course, at any given point in time, camera with different performance level are priced differently. That exactly how we see the future of digital LiDAR.

For our next-generation chips, we are pushing forward higher integration, strong performance, and better cost effectiveness. More importantly, we have already made very solid progress.

This year, we can expect to see a series of new chips and new products launched from us. It is foreseeable that as chips continue to iterate, LiDAR will witness a chip-level generational gap.

Zhang.

Thanks. That's really clear.

Appreciate it.

Hello, thank you for taking my question. This is Shiqi from Soochow International, and my question is from the industry evolution perspective.

I was wondering if we are seeing a shift in competition between L2 and L4, as L2 used to be more about automotive grade, mass production, and cost, while L4 focused more on system performance, scenario adaptability, and commercialization. How do you see this change, and what kind of capabilities are you prioritizing right now?

Hi, Shiqi. This is a a good question, also a tough one, huh?

Let me try to answer. First, from my perspective, both Level 2 and Level 4 are pushing hard on performance, just in different way.

People used to think the Level 2 segment was only about cutting cost, but that's really a supply-side problem in my perspective. The industry just couldn't offer good enough products at a reasonable price.

We can see a clear shift this year as high-resolution LiDAR become available at roughly the original cost. Level 2 customer are also upgrading fast.

That means that the Level 2 market is shifting its focus from yes or no, that means if I equipped with LiDAR or not, to how good it is. Back in 2025, when 192-channels LiDAR hit the market, customer who use 64 or 128-channels LiDAR also started started evaluating carefully and making rational choice.

Like I said earlier, once LiDAR goes digital, channel count and cost no longer have a straight linear relationship. They can even expect Level 3-grade LiDAR, currently around 500 500 channels, to trickle down to Level 2 applications.

As for Level 4, it always choose the best. The point is, it focus more on system-level configuration: how many LiDARs per vehicle, what combination to be used to cover more scenarios.

For Level 4, safety redundancy isn't an option, it's a must. The real fundamental shift is is this: the minimum bar for Level 2 is raising, and the threshold for Level 4 also go rising.

As the Level 3 mature and trickle down, the boundary between Level 2 and Level 3 will be redefined. Competing in Level 2 just by low cost or in Level 4 just by high performance alone will not be enough to build long-term barriers.

It's very clear for us, we use the same core technology to serve both type of customers. It's all built on our own digital chip architecture.

We can boost performance and keep cost down at the same time. For Level 2, we let customer get much better performance at a competitive price.

For Level 4, we offer full system perception and multi-sensor coordination, so that they can get scenario coverage and safety redundancy they need. This ability to fight on two fronts essentially come from our consistent focus on platform and chips itself, rather than simple split the market into two separate parts.

Shiqi.

Thank you, Mark, and congratulations on our outcome result. That's all for me.

Thank you.

Thank you for taking my question, Mark. This is Nora Min from UBS.

After the MX and M4 products were mass-produced and delivered last year, ASP has shown a declining trend. Looking into this year, do you expect the downward pricing trend to continue?

With the benefits of scale and in-house chips, how much further room do you see for price reductions? Thank you.

Thank you, Nora. I don't deny that the ASP go down.

Over the past 2 years to 3 years, the biggest change in industry about the ASP mostly come from optimizing and shifting platform architecture. Now the architecture has largely converged, and most of the component that were meant to be chips integrated have already gone through all that process.

In the short term, I'm not going to see those sharp ASP drops driven by tech transition, like we did in the past 2 to 3 years from RMB3,000 down to RMB2,000 and then down to RMB1,000. To be honest, our gross margin, especially ADAS, still under pressure in the near term.

On one hand, the automotive industry is extremely competitive, and on the other hand, we're proactively pushing LiDAR from being an option on more models to achieve higher penetration and eventually to becoming standard. Along the way, both pricing and gross margin will face pressure.

However, I don't think it will be a long-term situation. As new product ramp-in continue to stabilize, yield keeps improving, sales volume scale up further, and our product and chips advantage continue to play out, we expect the ADAS gross margin to gradually improve and stabilize.

Nora.

Thank you so much for your detailed reply. Look forward to seeing your continued momentum this year.

Thank you.

Thank you.

Hi, Mark. Thanks for taking my question, and congrats on the strong results.

Here is Sarah Chen from China Galaxy International. My question is around our innovation business.

During CES 2026, the company showcased its embodied AI manipulation solutions. Could you provide an update on the current progress of your hand-eye coordination solution, including the active camera and the dexterous hand?

For the innovation business, particularly the AC series, what's the production ramp-up schedule for this year, any recent orders or revenue expectations you can share? More broadly, what's your outlook for other robotics products and applications this year?

Also, what's the level of investment should we expect for the innovation business going forward as a percentage of total R&D spending? Thanks.

Okay. Thank you, Sarah.

That's a lot of questions, and let's break it down one by one. It is actually a topic I've been especially eager to discuss with everyone.

I want to say that RoboSense, our target is to become the Bosch of the robotics industry, to provide incremental components and solutions for robots. For us, LiDAR just one category.

We're working hard to launch different category to keep growing our business and to further embrace the incredible robotics age. For humanoid robot, the mobility of the lower body is already quite mature, while upper body, the manipulation capability, is driving a whole new industry revolution.

That is all about the hand, the eye, and a full closed-loop system that from models to data. I'm glad you noticed the hand-eye coordination solution we showcased at CES, and what's special about it is that it demos a long-horizon task with nearly 20 low-level operations steps, including a large number of flexible manipulations.

This is something very rare in on-site generalized demos so far. We're still refining this technology, and we don't plan to commercialize it in the short term.

As for components, the active camera went through two iterations last year, and is continuously being refined together with many leading labs. This kind of early customers have given us very valuable feedback, and we plan to launch the third version this year, which will be a production-ready model.

This is expected to enter mass production at the end of this year, starting to generate revenue for us. We anticipate that over the next 3 to 5 years, this product will grow into a category even larger than LiDAR.

Regarding the dexterous hands, we also released two versions last year. We're continually developing, and hope to achieve satisfactory results this year.

It's obviously a huge category, and we're not pushing too hard to commercialize it, but you can look forward to our technological progress by the end of this year, too. Robotics is our second growth curve, and we will continue to invest heavily.

Currently, innovation business, including robotics, account for about one-third of RoboSense total R&D spending. We will adjust our investment if it needs, and may even increase it further in the future.

Unlike ADAS, which is already in a scale-up phase, the robotics business is still in an investment phase. Our goal right now is to build our product and build our solution and reference customers, laying a solid foundation for future growth at scale.

Sarah.

Thanks, Mark. Very clear.

Look forward to our 2026 performance.

Thank you. Now,

Thank you, Operator. If there is no further questions at present, we would like to conclude by thanking everyone for joining our conference call today.

We welcome you to reach out to our Investor Relations department directly by emailing at [email protected]. Should you have any questions or requests for additional information, we encourage you to visit our Investor Relations website at www.ir.robosense.ai.

Thank you for everyone.

Goodbye. Thank you.