Zero Square Robotics

About Zero Square Robotics

Zerith is a Chinese robotics startup that develops humanoid robots for the service, education, entertainment, and exhibition sectors. Founded in January 2025 by a group of Tsinghua University graduates, the company was jointly incubated by the university and the Jianghuai Advanced Technology Center. The founding team consists of students from the university's artificial intelligence and robotics laboratory who have published academic papers in top robotics conferences and journals. Zerith has attracted talent from major technology companies like Baidu, ByteDance, and iFLYTEK.

The company's business model revolves around the research, development, and mass production of intelligent humanoid robots. In February 2025, Zerith secured tens of millions of yuan in an angel financing round and announced it had also received a multi-million-yuan order. It followed this with Angel+ and Angel++ funding rounds in July 2025, raising the equivalent of tens of millions of US dollars from investors including Baohe Venture Capital, Cowin Capital Group, and Waterwood Asset Management. The funds are intended for R&D of foundational algorithms, accelerating commercial implementation, and constructing mass production facilities.

Zerith has developed a comprehensive embodied intelligence solution, encompassing both hardware and software, and launched the Z1 and H1 robots. The company utilizes AI-assisted structural design to enhance dynamic performance and reduce energy consumption in its robots. A key technological feature is a physical interaction world model algorithm for motion control, which allows the robots to navigate and adapt to complex environments. For operational tasks, Zerith has developed specialized scene-generalized solutions inspired by autonomous driving technology, using vision-language models to analyze scenes and generate task descriptions. The company has also created the GP-6D Universal Grasping and Placement Model for mobile operations like delivery and a modular framework for complex sequential tasks.