RoboSense (02498) marked its tenth consecutive year of participation at the 2026 Consumer Electronics Show (CES), a significant milestone, and showcased its groundbreaking achievements in embodied intelligence with the global debut of an AI-powered robot. At the exhibition booth, RoboSense's globally premiered robot "Delivery Guy" autonomously completed a long-range task involving gift packing, transportation, unpacking, and box recycling, smoothly executing nearly 20 complex operational steps. This demonstration relied on RoboSense's self-developed "hand-eye coordination" solution, which fully integrates core technologies including the world's first VTLA-3D model, the robot's eye Active Camera, and a multi-degree-of-freedom dexterous hand. Also on display were products such as the second-generation all-solid-state LiDAR E1 Gen2, the world's first 3D safety LiDAR Safety Airy, the ultra-compact digital LiDAR Airy Lite, the automotive digital LiDAR models EM4 and EMX, alongside core technologies like RoboSense's fully self-developed chip portfolio. At this year's CES, RoboSense not only announced a collaboration with the globally leading brand in smart robotic mowers, Segway Navimow, and jointly exhibited the new Navimow i2 LiDAR mower, but also showcased unmanned delivery vehicles with strategic partners like Neolix and Coco Robotic. Furthermore, in partnership with Geely Auto (00175), it displayed the flagship six-seat SUV, the Lynk & Co 900, highlighting the commercial application of RoboSense's technological solutions across multiple scenarios. Through the exhibition of robots, AI models, core hardware, chip technology, and application deployments, RoboSense comprehensively outlined its blueprint for the robotics ecosystem business in the AI era, spanning from underlying technology to downstream applications.
The robot's "packing, unpacking, and gift delivery" demonstration represented a new breakthrough in embodied intelligence. At the RoboSense booth, the globally premiered robot "Delivery Guy" became the center of attention. Without any human intervention, it autonomously completed a long-range operation involving gift packing, shelving, transportation, unboxing, gift delivery, and folding the box for recycling, presenting New Year gifts to the audience. The core enabler of this demonstration was RoboSense's self-developed VTLA-3D manipulation model. Pioneering the integration of force-tactile modality with 3D point cloud information in the industry, it fuses visual, force-tactile, and language data with high-precision 3D color point clouds generated by the Active Camera. This allows the robot to perceive environmental states more comprehensively, significantly enhancing the success rate of highly dexterous manipulations. Additionally, the rich spatial structure information provided by the 3D point cloud reduces reliance on massive training data, enabling the VTLA-3D model to demonstrate higher data efficiency. Building on this, the solution also trains a task-planning AI capable of decomposing complex, abstract tasks into atomic sub-tasks and scheduling their execution, forming a "fast-slow dual-system" that achieves both long-range planning and precise operational capabilities. On the hardware front, the solution establishes a complete perception and execution system. The Active Camera series sensors act as the "robot's eyes," where the AC2 provides close-range, high-precision stereo and color information to support fine manipulation perception, while the AC1 handles long-range navigation and obstacle avoidance. The end-effector is equipped with a dexterous hand featuring multiple force-tactile arrays; the feedback from tactile signals effectively addresses visual blind spots, ensuring compliant and precise operations. Products like the AC1, AC2, and the multi-degree-of-freedom dexterous hand were exhibited individually at the booth. Benefiting from numerous pioneering innovations, the robot successfully completed the continuous operation demonstration in the novel environment of the CES venue, demonstrating the powerful generalization capability of RoboSense's embodied intelligence solution. The demonstration involved numerous high-precision, compliant actions, showcasing the advanced level of current robotic technology in handling long-range, high-precision tasks. Currently, RoboSense's embodied intelligence solution forms an end-to-end technological closed loop, applicable across various scenarios such as logistics, industrial, and service sectors.
Global partners shared the stage, showcasing commercial deployment across multiple scenarios. At this CES, RoboSense collaborated with several globally leading robotics and intelligent driving companies to exhibit commercial application results in vertical scenarios across various fields. In the consumer robotics sector, RoboSense partnered with Segway Navimow to jointly exhibit the Navimow i2 LiDAR robotic mower, integrated with a digital all-solid-state LiDAR. This collaboration deeply merges RoboSense's technological advantages with Navimow's globally leading consumer-grade platform, aiming to set a new benchmark for yard automation development. In the unmanned delivery field, the L4-level unmanned logistics vehicle, jointly exhibited with strategic partner Neolix, is equipped with RoboSense's digital high-precision LiDAR Fairy and has already entered mass production and delivery. Simultaneously, the unmanned delivery vehicle jointly exhibited with global urban robotics delivery platform Coco Robotic, equipped with RoboSense's all-solid-state digital LiDAR E1R, has been widely deployed across North American streets. In the intelligent driving sector, RoboSense and Geely Auto jointly exhibited the flagship six-seat SUV Lynk & Co 900, together showcasing advanced AI-driven intelligent driving products and their immense market potential to the world. These collaborations across diverse fields fully attest to the scalability and commercial viability of RoboSense's technological solutions.
Multiple new digital LiDAR products were exhibited. Regarding new LiDAR products, RoboSense centrally displayed several industry-leading offerings: the second-generation all-solid-state LiDAR E1 Gen2, which is smaller in size and offers stronger performance, is widely suitable for various robotics and autonomous driving scenarios; the world's first 3D safety LiDAR, Safety Airy, achieves an upgrade from 2D to near-hemispherical立体 protection, setting a new standard for industrial safety; the ultra-compact LiDAR Airy Lite, with its compact form factor and omnidirectional high-precision perception, serves as an ideal perception upgrade solution for multi-scenario robots. Concurrently, the mature digital LiDAR products EM4, EMX, and E1 debuted in the form of a complete automotive perception solution suite. For L2-level advanced driver-assistance systems (ADAS), the EMX achieves 192-line high performance through a highly integrated design; for L3-level autonomous driving, the EM4, with its exceptional performance exceeding 500 lines, promotes the safe implementation of high-level automated driving; for L4-level unmanned driving, the combination of the highly mature and high-performance EM4 and E1 provides long-range forward detection and ultra-wide-angle short-range blind spot coverage, enabling 360° blind-spot-free perception and accelerating the mass commercial production of L4 systems.
Fully self-developed chips propel the continuous evolution of robotics. At this CES, RoboSense systematically demonstrated the development path of its fully self-developed chip portfolio. Adhering to the philosophy of "AI-defining hardware, hardware-defining chips" since initiating its chip strategy in 2017, RoboSense has achieved full-stack self-research and mass production of four core chips: scanning, transmission, reception, and processing. Notably, its SPAD-SoC receiver chip and 2D VCSEL transmitter chip passed the AEC-Q102 automotive-grade certification in 2025, signifying that RoboSense has become the only company globally to have self-developed laser radar transmitter, receiver, and processing chips all meeting automotive standards across the entire chain. In 2021, it launched the world's first automotive-grade 2D MEMS scanning chip, solving industry challenges for front-load mass production and becoming the first supplier of automotive-grade LiDAR to the automotive industry. In 2022, it successfully developed the world's first integrated LiDAR receiver and processor SPAD-SoC chip, which for the first time integrated a SPAD array and processing SoC onto a single chip via 3D stacking, along with the industry's only mass-producible two-dimensional addressable VCSEL transmitter chip, effectively solving single-photon crosstalk issues. This enabled the launch of the world's first mass-produced all-solid-state LiDAR, the E1, providing reliable 3D perception for L3/L4 autonomous driving and various robots. In 2024, it mass-produced the industry's first LiDAR-dedicated SoC chip based on the RISC-V architecture, the M-Core, which received AEC-Q100 certification. This chip features picosecond-level time resolution, effectively improving ranging accuracy, and its high integration provides key support for significantly reducing LiDAR cost and size, promoting widespread adoption. Amid the rapid development of the AI robotics industry, RoboSense, with its complete suite of automotive-grade chip sets and mass production experience, continues to provide a solid foundational support for the large-scale industrialization of AI robotics technology.
Towards a new decade: From technology leader to ecosystem builder. From its first participation in 2016 to the present, the CES platform has fully witnessed RoboSense's leap from core technology breakthroughs to leading niche markets, and further to building a mature AI software and hardware technology system. At this year's CES, RoboSense's comprehensive display, spanning from underlying chips and perception hardware to upper-layer AI models and robotics platforms, established its industrial chain position as a core technology ecosystem builder in the field of AI robotics. Looking forward, RoboSense's path is clear and firm: to continue building on the foundation of self-developed innovation, scaling the heights of technology while commercializing its accumulated expertise in specific scenarios, driving the evolution of the global automotive and robotics industries, accelerating the intelligent transformation across countless sectors, and making intelligent robots a ubiquitous social productivity force.
Comments