Two-month-old dexterous-hand startup ManuPrax raises nearly $10 million

Founded in September by two young researchers, ManuPrax Robotics closed two rounds led by Alphaist Partners and Yaotu Capital.

ManuPrax Robotics, a dexterous-manipulation startup founded in September, has closed two consecutive funding rounds totaling nearly $10 million. The rounds were led by Alphaist Partners and Yaotu Capital, with follow-on backing from Jieli Angel and the controlling shareholder of a components manufacturer.

The company was co-founded by two young researchers. Chief executive Li Puhao, a student in Tsinghua University's Department of Automation, has worked on dexterous hands and manipulation since 2021. Chief technology officer Chen Zeyuan, who holds a master's degree from Peking University, researches interaction algorithms for dexterous manipulation and previously worked at Amazon FAR, Agibot and JD.com. He contributed to the ClutterDexGrasp and CADGrasp research projects.

ManuPrax is building an integrated software-and-hardware system for dexterous hands, using reinforcement learning so models learn manipulation through interaction, trial and error. It combines simulation, human demonstrations and generated interaction trajectories to build training data, and is testing whether models trained in simulation transfer to real hardware. The aim is an operations-model platform that works across different dexterous hands, turning high-level goals from multimodal models into physical manipulation skills robots can call on.

The team is also weighing building its own hardware or partnering with manufacturers to speed up technical validation and gather real-world data. The round lands in a crowded corner of the market: IT Juzi data puts financing for dexterous-hand startups above 25 billion yuan in the first half of 2026 alone.