Embodied AI Moves Into Factories and Warehouses as Efficiency Takes Focus

At the CeMAT Asia logistics exhibition, which opened in Shanghai on Nov. 3, embodied AI was presented less as a speculative concept than as equipment already handling limited work in factories and warehouses. The tasks include moving goods, storing and retrieving items, and sorting.

Exhibitors focused on mechanical and operational details: robots that climb racks, retrieve boxes in four directions, avoid obstacles in narrow aisles and coordinate in multi-robot fleets. The improvements were measured in concrete terms, such as the time needed for a single turn and the space occupied by one storage slot.

The emphasis reflects a shift in the debate. Efficiency, cost and commercialization remain open questions, so vendors are trying to show that gains come from repeated, small engineering refinements rather than only from a smarter “brain.” For warehouse operators, such metrics matter because they directly affect throughput, layout density and operating expense.

The exhibition suggests embodied AI is entering real logistics settings through incremental capabilities, with competition now extending to reliability, space efficiency and coordination at scale.