A recent research report from CICC highlights that general humanoid robots (not fully anthropomorphic) must possess multimodal perception and end-to-end large model capabilities to efficiently handle logistics sorting operations. Currently, these robots have reached the hardware commercialization threshold for logistics applications, yet domestic embodied model capabilities require further enhancement. From an economic standpoint, assuming a payback period of two years, the input-output ratio of general humanoid robots is already on par with that of human sorting workers. It is anticipated that in the second half of this year, these robots will transition from demonstration environments to customer trials, with explosive growth forecasted for next year. Given the high labor costs, overseas markets are poised to become the primary destination for the implementation of logistics scenarios involving these robots.