Gewu Tongzhi Raises Nearly RMB100M Angel Round for Robot Physical-Interaction Brain

Gewu Tongzhi, a Hangzhou-based startup developing dexterous manipulation capabilities for robots, has completed an angel round of nearly RMB100 million, according to reports. Investors include Hongfu Asset, Peak Investment, Hangzhou Capital, Hangzhou Talent Group, and several industrial investment institutions and listed companies.

Founded in Hangzhou in 2026, Gewu Tongzhi is researching how robots can understand objects and the physical changes that occur during manipulation, and improve their actions through interaction with real environments. Its core technical route combines world models with physical-interaction learning.

Rather than relying only on static datasets, the company aims to build a robot brain that learns from contact and dynamic feedback in real-world tasks. This approach is intended to support dexterous manipulation in industrial scenarios, where robots must handle varied objects and adapt to changing physical conditions. For industrial users, such capabilities could reduce the need for highly customized programming and make robot deployment more flexible.

Founder and CEO Cai Jia previously worked at Huawei’s 2012 Laboratory and the Vehicle BU 2030 Laboratory, participating in R&D and mass-production delivery of Huawei’s ADS assisted-driving system. That background spans large-scale model research and automotive-grade deployment, according to the company.

Gewu Tongzhi says the new funds will be used for model pretraining, hardware and robot-body R&D, talent recruitment, and real-world scenario data collection. The funding signals continued investor interest in embodied AI and robot “brains” that can connect perception, physical reasoning, and action in industrial settings. It is among a growing group of embodied-AI startups seeking industrial deployment.