Robot business surges 6 times, ROBOSENSE starts second growth curve

On August 21, ROBOSENSE (2498.HK) released its mid-term performance for 2025. After the financial report was released, the stock price rose over 9% today.

In the first half of 2025, the company's total revenue was RMB 783 million, a year-on-year increase of 7.7%. Among them, total revenue in the second quarter reached RMB 455 million, a year-on-year increase of 24.4% and a quarter-on-quarter increase of 38.9%.

The improvement in profitability is particularly significant. The overall gross margin in the first half of the year increased sharply from 13.6% in the same period last year to 25.9%; the gross margin in the second quarter reached 27.7%, marking the sixth consecutive quarter of steady growth. The net loss in the first half of the year narrowed significantly by 44.5% year-on-year, to RMB 148.6 million.

Behind the data are two distinct business curves.

The growth engine comes from the robotics business. In the second quarter, the sales volume of LiDAR for robotics and other fields reached 34,400 units, a year-on-year increase of 631.9%. In the first half of the year, this business segment generated revenue of RMB 220.7 million, a year-on-year increase of 184.8%, with a gross margin as high as 45.0%.

At the same time, the ADAS (Advanced Driver Assistance Systems) application segment, which is the main business, is undergoing strategic adjustments. In the first half of the year, the sales volume of LiDAR for ADAS applications was 220,500 units, a year-on-year decrease of 6.0%; sales revenue was RMB 500.3 million, a year-on-year decrease of 17.9%. However, it is worth noting that despite the decline in revenue, the gross margin of the ADAS business increased from 11.2% in the same period last year to 17.4%, which the financial report attributed to the optimization of raw material costs and the adoption of self-developed SOC chips.

Between the rise and fall, what the market sees is a company using high-growth, high-profit new businesses to provide a buffer for the technological upgrades and product iterations of its core business. This forms the key to understanding the current value of ROBOSENSE.

The Two Curves of the Financial Statements

The mid-term report of ROBOSENSE outlines a company with steady growth and continuously improving profitability.

In the second quarter of 2025, total revenue reached RMB 455 million, a year-on-year increase of 24.4% and a quarter-on-quarter increase of 38.9%. Total revenue in the first half of the year was RMB 783 million, a year-on-year increase of 7.7%. A more critical indicator is the gross margin, which reached 27.7% in the second quarter and 25.9% in the first half of the year, marking the sixth consecutive quarter of steady growth. At the same time, the net loss in the first half of the year narrowed significantly by 44.5% year-on-year, to RMB 148.6 million The improvement in profitability stems from the interplay of two curves.

The first curve is the "explosive growth" of the robotics business.

This is a story of growth. In the second quarter, the sales volume of lidar for robotics and other fields reached 34,400 units, a staggering year-on-year increase of 631.9%. In the first half of the year, the revenue of this business segment reached RMB 220.7 million, a year-on-year increase of 184.8%. This curve is not only steep but also highly profitable, with a gross margin reaching 41.5% in the second quarter and 45.0% in the first half of the year, far exceeding the company's overall level, becoming the core engine driving overall profitability.

The second curve is the "gear shift" of the ADAS business.

In the first half of the year, the sales volume of lidar for ADAS applications was 220,500 units, a year-on-year decrease of 6.0%; sales revenue was RMB 500.3 million, a year-on-year decrease of 17.9%.

The temporary decline in revenue has a clear background. The financial report explains that this is mainly due to product cycle adjustments with two major automotive clients, one of which has shifted to products developed by its strategic tier-one suppliers, while the other has canceled the lidar configuration on its vehicle models.

However, more forward-looking than the revenue figures is the gross margin of this business segment. Despite the decline in revenue, the gross margin of the ADAS business significantly increased from 11.2% in the same period last year to 17.4% in the first half of 2025. Behind this is the optimization of raw material procurement costs, and more importantly, the adoption of self-developed SOC processing chips, which have significantly lower costs compared to previously procured FPGA chips from third parties.

The divergence of the two curves demonstrates a mature strategic execution capability.

The robotics business is not a sideline; it has grown into a high-growth, high-profit independent engine at a critical moment. The cash flow and profits it generates provide the necessary funding and time window for the core ADAS business to undergo essential product upgrades, smoothing the pains of transformation. The fact that the gross margin of the ADAS business has improved despite a decline in revenue is itself a strong testament to the cost-effectiveness of the new generation of digital platforms. The company is proactively replacing old businesses with limited profit margins with new businesses of higher quality and higher profit margins.

Digital Arsenal: From E Platform to EM Platform

The technology strategy of ROBOSENSE is a fundamental transformation from analog to digital, comparable to the transition of cameras from the film era to the digital era. At the core of this revolution is the company's self-developed SOC chip.

The competition in lidar has entered deep waters, moving from innovations in system optical architecture to a deeper "chip war." ROBOSENSE's advantage lies in integrating the core receiving system and echo feature recognition functions into a single SOC chip. This not only streamlines the link and reduces power consumption but also allows the performance of lidar to continuously improve following Moore's Law, building a solid moat capable of withstanding dual challenges from competitors and OEMs' self-development.

This digital arsenal consists of two major platforms.

E Platform: The Foundation of All-Solid State The E platform is the industry's first all-solid-state digital lidar platform, built on the self-developed CSOC chip. Its first product, "Evan" (E1), is positioned as the only choice for mass production of L3 and L4 level autonomous driving wide-angle and blind-spot lidar, and also the optimal option for robotic lidar. A dedicated product for the robotics field, "Evan R" (E1R), has also been launched.

The E1 officially entered mass production in the first half of 2025, becoming the industry's first large-scale commercial all-solid-state lidar. The company expects that the shipment volume of the E platform series products will reach six figures in 2025.

EM Platform: Long-Distance Lidar

The EM platform is a digital platform specifically designed for long-distance detection, based on the self-developed SPEDSOC chip, aimed at solving point cloud quality issues in adverse weather conditions such as rain and fog. This platform has two core products targeting different battlefields:

• EM4: Currently the only digital lidar in the industry with over 500 lines that can be mass-produced, it is a powerful tool for L3 and L4 level main perception radar.
• EMX: A true 192-line lidar that provides optimal performance for the vast L2+ advanced driver assistance market. Compared to competitors' previous generation 64-line products based on analog architecture, the EMX achieves a significant improvement in resolution and cost advantage with the support of digital architecture and chips, representing a generational leap in architecture and chip technology.

Both products are set to enter large-scale mass production and delivery in the third quarter of 2025.

Robosense's product pipeline is not merely a simple accumulation of models, but a meticulously designed, hierarchical weapon system. The EMX serves as the "battering ram" to capture the mainstream L2+ market, the EM4 is the "precision-guided weapon" to seize the high ground of L4 technology, while the E1 is the "multifunctional vehicle" to consolidate the L4 ecosystem (blind-spot detection) and simultaneously ignite the growth curve of robotics. This comprehensive product combination not only allows for precise strikes on various niche markets but also creates significant cross-selling opportunities, deeply locking customers within its ecosystem.

In the Robotaxi field, which has the most stringent technical requirements, digital lidar has completely replaced old mechanical products. Robosense claims to have captured 90% of the market share in this field.

The combination of "EM4 + E1" has become the preferred configuration for the new generation of Robotaxi and has completed product mass production verification with eight leading global customers. Its client list includes almost all industry leaders: Didi Autonomous Driving, Baidu, Pony.ai, WeRide, and several leading Robotaxi and Robotruck companies in North America.

The newly launched EM platform is making waves in the ADAS market. Within six months of its release, it has secured 45 model designations from eight major OEMs.

Among them, there is an exclusive supply designation from "a globally leading new energy vehicle company" for as many as 32 models. Its designated range covers the globe, including Japan (becoming the exclusive supplier for two of the Top 2 automakers), North America, and Europe (including a leading luxury automotive joint venture brand) core markets In June 2025, the company completed the delivery of its 1 millionth vehicle-mounted lidar, marking an important milestone.

The absolute dominance in the Robotaxi field is transforming into a strong driving force in the ADAS market. For passenger car manufacturers, choosing a supplier validated in the most stringent L4 scenarios undoubtedly provides dual guarantees in technology and brand. When a car manufacturer can claim to consumers, "The lidar we use is the same as that in fully autonomous vehicles," it transforms from a technical parameter into a powerful endorsement of safety and technology branding. The success of Robotaxi provides the strongest "certificate of trust" for the expansion of the ADAS business.

The customer base for the robotics business has surpassed 3,200. Its growth is driven by breakthroughs in multiple vertical fields:

• Lawn Mowers: Exclusive partnerships have been established with several of the top 5 manufacturers globally, with order sizes reaching seven figures.
• Unmanned Delivery: Collaborations have been formed with domestic companies such as Meituan, New Stone Age, and White Rhino, as well as leading delivery platforms in North America and COCORobotics.
• Embodied Intelligence: Deep collaborations are ongoing with over 20 globally leading embodied robotics companies, including Yushu Technology.

The company maintains its guidance for robot lidar shipments to reach six figures by 2025 and expresses confidence in "slightly exceeding expectations."

Next Frontier: The Eye of the Machine

The launch of the AC (ActiveCamera) platform marks SuTong JuChuang's ambition to go beyond being a lidar hardware supplier, aiming to become a foundational technology platform company in the robotics era.

AC is defined as a brand new category—"the true eye of the robot." It is an "all-in-one" sensor that can directly output RGBD data, efficiently synchronizing and integrating the geometric accuracy of lidar, the semantic information of cameras, and the posture data of IMUs at the hardware level.

Its goal is to address the pain points of existing depth sensors (such as Intel's RealSense) that perform poorly and have unstable accuracy in strong outdoor light. Once resolved, it could unlock a huge market for industrial automation and consumer applications (such as real-time 3D environment modeling).

The AC platform is equipped with an "AI-Ready" ecosystem, including open-source algorithms, SDKs, and toolchains, aimed at providing development convenience for the entire robotics industry, forming a closed loop from data collection to model training, thereby establishing a highly sticky technology platform.

The launch of the AC platform can be seen as a strategic response to the "pure vision" route. By creating a sensor that natively integrates lidar and cameras, SuTong JuChuang is laying the hardware foundation for a "fusion-first" perception route for all robots (from cars to humanoid robots). The future it bets on is that the ultimate form of artificial intelligence is not the victory of a single sensor, but the intelligent fusion of multimodal sensors. SuTong JuChuang is committed to becoming the company that provides core hardware and software tools for a fused future.

The numbers in the financial report depict a company that is consolidating its current leadership position while systematically building for the future