Hard Krypton’s first release | Fudan professor and former Intel chief scientist worked on the embodied brain, “Mooshen Intelligence” completed nearly 100 million yuan in Pre-A round of additional financing

Author | Qiu Xiaofen Editor | Yuan Silai Hard Krypton has learned that "Motion Brain" has completed nearly 100 million yuan in Pre-A round of additional financing. Investors in this round include China's leading property services company, a Hong Kong consortium, industrial investment platforms jointly created by multiple listed companies Jinyue Investment, Chuanghehui Capital, and old shareholders Xuhui Capital. This is also the second time that “Mooshen Intelligence” received investment within two months after the 300 million yuan Pre-A round in May 2026. In addition, the company’s nearly 500 million yuan in Pre-A+ round financing is also being completed. The company revealed that since the first half of this year, the valuation of “Mooshen Intelligence” has increased more than 10 times, making it one of the fastest-growing embodied brain companies in the industry. "Mooshen Intelligence" was established in January 2025. It is positioned as the world's most action-savvy self-evolving embodied brain company and is committed to "installing native general-purpose brains on 1 billion robots." The company was co-founded by Professor Chen Tao, director of the Deep Learning Laboratory of Fudan University, Dr. Zhang Yimin, former chief scientist of Intel China, and serial entrepreneur Mu Zelin. The core team has people with backgrounds from the three major chip giants: HiSilicon, Intel, and Nvidia. It is one of the few embodied large-model teams in China with operator-level adaptation capabilities. Although it was only registered in 2025, the technology accumulation of "Mooshen Intelligence" has already started in 2022. On the technical route, "Mooshen Intelligence" has chosen a path that is different from the mainstream VLA (Visual-Language-Action) - the "World Motion Model" with motion as the core. (Source/Enterprise) Starting from 2022, the "Mooshen Intelligence" team has released the MLD (hidden space action diffusion model). At that time, the mainstream of the industry regarded actions as discrete signals, resulting in the generated actions being rigid and lacking diversity. However, the "Mooshen Intelligence" team was the first in the world to propose mapping actions to latent space (Latent Space) and introduce a diffusion model (Diffusion Model). MLD is equivalent to letting AI learn to generate actions "without a teacher": no longer need to frame by frame