Shanghai Moushen Intelligent Technology

Shanghai Moushen Intelligent Technology Co., Ltd., also known as Motion Brain, is developing an end-side embodied brain for robotic systems. The company was officially registered in 2025, though its technological development dates back to 2022. It was co-founded by Professor CHEN Tao, the Director of the Deep Learning Laboratory at Fudan University; Dr. ZHANG Yimin, a former Chief Scientist at Intel China; and MU Zelin, a serial entrepreneur who serves as the CEO. The core team includes talent with backgrounds from HiSilicon, Intel, and NVIDIA.

The company's core technology is a motion-centric "World Motion Model," which differs from the more common Vision-Language-Action (VLA) models. This model focuses on generating and directing motion. In 2022, the team released the Motion Latent Diffusion Model (MLD), which allows AI to generate natural and coherent movements from random noise. Building on this, they launched MotionGPT in September 2023, which breaks down human postures into approximately 3,000 "motion primitives" to better direct robotic actions. This technology enables the rapid adaptation of their embodied large model to different robot bodies, a process that can be completed within two weeks. The system has been adapted for robots such as Fudan University's "Guanghua No.1," Unitree G1, and the Shanghai Humanoid Robot Innovation Center's "Qinglong" and "Linglong".

Shanghai Moushen has demonstrated early commercial traction, reporting audited revenue of 10 million yuan in 2025 and 30 million yuan in the first half of 2026. The company has attracted significant investment, securing a CNY 300 million Pre-A round in May 2026 from investors including Green Technology Private Equity, Leo Group, and Huafu Capital. This was followed by an additional Pre-A financing of nearly 100 million yuan from investors such as Jin Yue Investment and Chuanghehui Capital.

Keywords: Embodied AI, World Motion Model, MotionGPT, Motion Latent Diffusion, robotic motion, humanoid robots, AI brain, deep learning, autonomous systems, motion primitives, robot adaptation, AI-generated motion, robotics software, intelligent systems, Fudan University, Intel, NVIDIA, CHEN Tao, ZHANG Yimin, MU Zelin