X Square Robot, a Shenzhen-based embodied AI company, concluded its participation at the 2026 World Robot Conference in Beijing, showcasing its "one model, many applications" approach across multiple robotic platforms and real-world tasks. Powered by its WALL-B AI model, the company demonstrated this approach across several physical environments, extending a logistics livestream it began on August 12 into a five-hour challenge during the conference.
At the event, X Square showed off WALL-B's capabilities across a range of robotic platforms. In a simulated home setting, two wheeled humanoid robots responded to spoken commands, carrying out tasks including organizing household items, retrieving and delivering objects, cleaning, watering plants, and handling pet-related chores. Using coordinated two-arm manipulation and obstacle avoidance, the robots moved between rooms and completed the requested tasks.
This demonstration is part of a broader push by X Square to deploy robots in actual homes. In March, the company partnered with Chinese local-services platform 58.com to launch a home-cleaning service, with robots working alongside professional cleaners during household visits. In May, X Square began placing robots into participating households for extended in-home service through what it calls its "X Family Member Program."
Elsewhere at the conference, a dexterous five-fingered robotic hand with 22 degrees of freedom, meaning it can move in 22 independent ways, completed a multi-step task unboxing a fan, powered by WALL-B. With its view of the object partially blocked by the box itself, the system opened the packaging, removed the fan from a tight space, placed it on a table, turned it on, and visually confirmed it was running. A separate dual-arm robotic system responded to visitors' spoken requests by selecting flowers of a specified color and arranging them in a vase, adjusting its grip and positioning to handle flexible stems and delicate, deformable flowers, and re-planning its movements when the flower basket or vase was moved mid-task.
X Square also extended its focus on real-world deployment into logistics. The WALL-B-powered system is already running on a live parcel-sorting line and is designed to support continuous, around-the-clock operation. The demonstration at WRC served as a public test of that capability: during the challenge, the system sorted 10,000 parcels in 5 hours, 14 minutes, and 1 second, working out to an average throughput of 1,911 parcels per hour, or roughly 1.88 seconds per parcel.
Throughout the challenge, the robots handled parcels that varied in size, shape, material, and position, requiring the system to repeatedly identify, grasp, orient, and sort each item, while correcting failed attempts and adapting to changes on the conveyor belt. Maintaining consistent throughput across thousands of parcels served as a practical test of the reliability and endurance needed for commercial logistics operations.
Wang Qian, founder and CEO of X Square Robot, said embodied AI requires a foundation model built specifically for the physical world, developed in parallel with the foundation models built for the digital world. He said X Square's goal is to build a shared intelligence foundation capable of operating across different robots, tasks, and physical environments, continuing to improve through real-world deployment.
Taken together, the company frames these demonstrations as reflecting its broader view that embodied intelligence shouldn't be tied to a single robot, task, or environment. X Square said it will continue integrating its AI models with robotic hardware as it expands into additional home, logistics, and industrial applications.
