The 2026 World AI Conference in Shanghai highlighted a wave of embodied general AI products from more than 200 vendors, turning the expo floor into a fiercely competitive arena where robots performed a wide range of adaptive tasks.
Embodied general AI in action
Exhibitors demonstrated what is meant by “embodied general AI”: intelligent systems built to execute variable tasks and adapt to different application scenarios. On the show floor, robots illustrated that versatility by performing actions from picking items off a shelf to handing objects to visitors, and executing precision operations.
Concrete examples presented in Shanghai
Demonstrations included robots retrieving a mock medicine package from a shelf, serving drinks to attendees, grasping and transferring items between a person and the machine, opening a washing machine to load laundry, and sorting parcels. These vignettes underscored the machines’ ability to sequence distinct physical actions while adapting to human interaction and the constraints of an exhibition environment.
A highly competitive market
With more than 200 vendors present, the event had a distinctly competitive feel: each supplier showcased variants of so‑called “generalist” robots aimed at domestic, logistics, or service use cases. Interoperability of functions and adaptability to changing tasks emerged as the key differentiators promoted across booths.
The presentations offered a concise tour through diverse hardware and software approaches to realizing versatile intelligence, with the show format favoring live, practical demos over speculative talks.
In summary
The conference provided a tangible snapshot of embodied general AI today: prototypes and products that prioritize operational flexibility and the execution of heterogeneous tasks, displayed in a live setting where competition between suppliers was plainly visible.
