Chery physical AI ambitions rest on a simple wager: that car-industry validation and scale are harder to copy than model size
Chery physical AI projects in robotics and low-altitude aviation will draw on the perception, control and manufacturing standards the automaker developed for cars, it said at its 2026 International User Summit. At their centre is CheryMind, a foundation model built to learn how the real world behaves from road and harsh-climate test data.
CheryMind and a companion system, the Chery All-World Engine, underpin Aimoga Robotics’ MoAgent agent, a factory-integrated level 4 robotaxi and the Lingxiao passenger electric vertical take-off and landing (eVTOL) aircraft. Chery says a shared intelligence layer can lower development costs across sectors and let each product learn from the others.
The robotics arm draws wholly on Chery’s car supply chain while reusing 70% of the automaker’s core vehicle research and development technology. Aimoga has shipped over 3,000 robots for upwards of 100 applications across 60-plus countries and regions; the robotaxi and eVTOL also pool perception, computing and backup safety systems.
Cost and immature supply chains have so far held back commercial use of such machines. Chery is betting that car-industry scale, quality control and a six-dimension testing regime spanning extreme heat, cold, salinity and rough roads can move them beyond prototypes. The automaker argues physical AI should be judged on safe, reliable operation at scale, an approach it is applying from cockpits to home services.
Source: GlobeNewswire
Source: www.automotiveworld.com
