MoSense Raises RMB100M for Robot Electronic Skin
Shanghai robotics startup MoSense reportedly closes an RMB100 million-level Angel+ round to develop tactile hardware, algorithms and robotic motion models.

Robots are getting better at recognising objects through cameras, but understanding physical contact remains a difficult part of manipulation. MoSense is building technology intended to make robotic bodies more sensitive to touch.
What happened
Shanghai-based MoSense reportedly completed an Angel+ round at approximately the RMB 100 million level, co-led by CDH Investments and Oasis Capital, with participation from robotics and technology investors.
The funding was first reported in late October, with an initial report on 26 October and additional coverage on 27 October. It should not be interpreted as a separate new round on each date. The round amount is approximate rather than a precisely disclosed cash total.
Founded in May 2026 by a team associated with Hong Kong University of Science and Technology, MoSense develops full-body tactile sensing solutions for robots. Its MoSkin hardware integrates sensing into structural components, while MoTact processes contact data. It is also developing MoWorld, a proposed model intended to incorporate tactile feedback into robot control.
Why it matters
A vision model may identify a glass, but a robot still needs to know when it touches the glass and whether its grip is secure. Tactile feedback can help close that gap, particularly for delicate or unpredictable physical tasks.
Combining sensors, processing algorithms and robot-control models could create a more integrated system than selling sensing hardware alone. The technical questions are signal quality, robustness, calibration and cost across different robotic designs.
The bigger picture
As investment in embodied AI accelerates, component suppliers that provide useful physical feedback could become important infrastructure for robotics manufacturers. MoSense is an early example, but its planned full-stack approach still needs validation across commercial robots and real working environments.
