Since adopting an AI+robotics strategy at the start of 2025, RoboSense has built a full-stack AI loop from hardware through algorithms. At this year's CES, RoboSense staged a global debut for its AI robot, spotlighting the company's breakthroughs in embodied intelligence.
Among the highlights was the first public showing of a "delivery assistant" robot. With no human intervention, it autonomously handled long-range tasks spanning gift packing, shelving, transport, box opening, handover and folding the box for recycling — smoothly chaining nearly 20 complex steps.

Image source: RoboSense
Powering the demo is RoboSense's in-house "hand–eye coordination" solution, a full-stack integration of its VTLA-3D manipulation foundation model, the Active Camera robot eye, and a multi–degree-of-freedom dexterous hand, among other core technologies.
RoboSense's VTLA-3D manipulation model is an industry first to combine force–tactile sensing with 3D point clouds. It fuses vision, force/tactile and language inputs with high-precision 3D color point clouds generated by Active Camera, giving the robot a fuller read on its surroundings and sharply lifting success rates in fine manipulation.
On top of that, the solution trains a task-planning AI to break abstract jobs into atomic subtasks and schedule execution — creating a two-speed system that balances long-horizon planning with precise control.

Image source: RoboSense
On hardware, the robot carries a complete stack from perception to actuation — all developed in-house by RoboSense. AC-series sensors serve as the robot's eyes: AC2 units mounted above the eyes and at the hand endpoints deliver close-range, high-precision stereo and color data for fine operations; AC1 units on the body and back handle long-range navigation and obstacle avoidance. The end effector is an 8-degree-of-freedom dexterous hand with tactile sensing, addressing visual blind spots and ensuring smooth, accurate manipulation.
RoboSense also unveiled several new lidars: the second-generation solid-state E1 Gen2, suited to a wide range of robotics and autonomous-driving scenarios; the Safety Airy 3D safety lidar, upgrading protection from 2D to near-hemispherical coverage; and the ultra-mini Airy Lite.
In addition, its mature digital lidars — EM4, EMX and E1 — made their debut as a complete in-vehicle perception solution.
Beyond showcasing its end-to-end stack for embodied intelligence, RoboSense teamed up with several global leaders in robotics and autonomous driving to present commercial deployments across multiple verticals.
In consumer robotics, RoboSense partnered with Weilan Dalu to showcase the Navimow i2 lidar lawn-mowing robot integrating the E1R digital solid-state lidar. Weilan Dalu's Navimow is a leading brand in smart mowers and is now in more than 30 countries — including Germany, France, Denmark, the U.S., the U.K. and Australia — serving over 400,000 users.
In autonomous delivery, RoboSense and Neolix jointly showed a Level 4 logistics vehicle fitted with RoboSense's Fairy digital high-precision lidar, already in mass production and deliveries. It also teamed with Coco Robotics on a delivery robot equipped with the E1R solid-state digital lidar, which has seen broad deployment across North American streets.
In intelligent driving, RoboSense worked with Geely to exhibit the flagship six-seat SUV Lynk & Co 900.









