Massive data is one of the key differences separating embodied intelligence from traditional industrial robots, alongside advanced AI "brains" and multimodal physical designs that adapt to specific industrial workflows, said Wang Ziqi, Chief of Staff at Agile Robots.

He made the remarks at a closed-door forum on embodied intelligence and industrial chain innovation during the 2025 T-EDGE on Tuesday.

First, the "brain" – including large models and VLA – endows robotic end-effectors with generalization and robustness. Second, the use of massive data; Agile Robots has built data factories across different scenarios to drive a Real2Sim2Real closed-loop system. Third, the robots' multimodal nature, where the physical embodiment of an intelligent system is tailored to fit specific workflows, ensuring optimal adaptation to industrial tasks.

Wang emphasized that these features collectively elevate embodied intelligence beyond conventional automation, enabling more flexible, data-driven, and context-aware industrial operations.