The WALL-B robot from X Square, a $2.8 billion Shenzhen-based embedded AI company, sorted 1,816 packets per hour during a one-hour livestream, achieving an accuracy of more than 98 percent.
The benchmark mark that X Square set out to beat — 1,248 packages per hour — matches Figure AI's sustained average on its 200-hour run in May. WALL-B surpassed it by about 45%, setting the challenge ahead of time. @aipost
You may also likeMicrosoft services almost restored after 24 hours of outageUnlike humanoid systems, WALL-B employs two 6-axis robotic arms with specialized grippers. The system continuously adjusts its strategy as packages move, overlap, or become obstructed. After each pickup, it reanalyzes the stack, adapting to warped bags or misaligned labels, minimizing human interventions.
The underlying AI model — the WALL (world unified model) family — also operates in less predictable environments, from housework (housekeeping with 58.com in Shenzhen) to tabletop handling, requiring advanced insight and reasoning. The company, backed by Xiaomi, Meituan and Alibaba, is betting on the generality of the model rather than the hardware.
