At the 2026 World Physical AI Conference and Unitree Partner Ecosystem Conference, Wang Xingxing, the visionary founder of Unitree Technology, offered his profound perspectives on the pivotal "GPT moment" for embodied AI. In his view, this transformative milestone is reached when robots can adeptly accomplish approximately 80% of their primary tasks in roughly 80% of unfamiliar environments, all through simple voice or text commands. Presently, large-scale AI models have showcased remarkable prowess in robot planning, navigation, and various other domains. However, to truly usher in the GPT moment for embodied AI, the paramount challenge remains the seamless integration of AI models' input and output with the physical capabilities of real-world robots. Wang Xingxing envisions that, as early as the end of this year or next, robots will be capable of executing tasks like room tidying based on clear instructions.