At the World Internet Conference in Wuzhen, Yushu Technology founder Wang
Xingxing said robotics data utilization remains low, with large volumes of
sensor and operational data unused. He expects physical/embodied AI models that
can self-evolve to raise data-use efficiency and speed technical iteration. Wang
gave an optimistic timing, saying 2027 is very likely to be a critical point for
the robotics industry or robot models. He defined an embodied-intelligence “GPT
moment” as robots completing roughly 80% of tasks in about 80% of unfamiliar
scenes via voice commands, and said leading global models are close to that goal
but still face key technical barriers.
