On February 11, 2026, Xinghai Tu announced the completion of a 10 billion yuan Series B financing round. Investors in this round included industrial capital partners such as Jinding Capital, BAIC Capital, and Bihong Investment, alongside private equity firms Zhen Fund, Qianhai Fangzhou, and international fund Yifeng Capital. Five major existing shareholders—Cathay Capital, Meituan Longzhu, Today Capital, Xianghe Capital, and Gaorong Ventures—collectively participated with oversubscribed or full additional investments.
Xinghai Tu adheres to an end-to-end Vision-Language-Action (VLA) foundational model technology roadmap, continuously advancing model performance and generalization capabilities. The company released its G0 model in August 2025, followed by an upgraded version, G0 Plus, in January 2026, positioning it as the "world's first out-of-the-box VLA model." Additionally, the company's open-source Xinghai Tu Open World Dataset has been downloaded over 500,000 times.
The company has established a full-chain, high-efficiency platform for physical world data collection and model training. In 2025, building on its mature real-world data acquisition system, Xinghai Tu innovatively introduced new collection methods such as UMI, further enhancing the efficiency, diversity, and accuracy of multi-form robot scenario data. By 2026, the company entered a phase of large-scale training with hundreds of thousands of hours of high-quality data, continuously iterating and optimizing its embodied foundational models through massive scenario datasets.
It is reported that Xinghai Tu has already received orders for several thousand units, with customers including top global universities, research institutions, and industry-leading enterprises. The company has achieved full-stack in-house research and development, covering core modules, whole-machine design, high-quality data, end-to-end models, and scenario solutions. Its R1 Pro and R1 Lite platforms have also become core development platforms for leading laboratories, such as Stanford University's Fei-Fei Li team and Physical Intelligence (PI).