RoboScience

Based in Beijing, China and founded in 2024, RoboScience Co. , Ltd, trading as RoboScience, primarily engages in the development of embodied intelligence cerebellar and neocerebellar models and products. The company's founder and CEO, Tian Ye, graduated from the Department of Physics of the University of Science and Technology of China. He worked at Apple and built Apple's machine learning platform.

The company launched the Manipulation Foundation Model in 2024. In May 2026, RoboScience raised CNY 1 billion in Series A funding. The company’s main business is the research and development of cross-embodiment general-purpose embodied intelligence (Cross Embodiment AI), enabling robots to achieve universal manipulation capabilities across diverse hardware, objects, and tasks. The company’s self-developed innovations include a fast-slow system robotic learning framework, self-supervised training technology, and the Manipulation Foundation Model, which bridges vision-language models with the physical world.

Its products are used in the field of robotic operations. The company generates revenue by providing embodied intelligence solutions and licensing its proprietary AI models to robotics manufacturers, automation enterprises, and AI-driven industries, capitalizing on its advanced frameworks and scalable data systems.. The Series A funding raised in May 2026 will be used to further develop its core VLOA large-scale model technology, as well as to advance the engineering and mass production of its self-developed robot bodies, accelerating the large-scale deployment of general-purpose embodied intelligence solutions.