Key facts
- X Square Robot has achieved a post-investment valuation exceeding 20 billion yuan ($2.9 billion).
- The company has completed four funding rounds, including Series B+, B++, and C.
- It is the first embodied intelligence company in the Guangdong-Hong Kong-Macao Greater Bay Area to reach this valuation.
- X Square Robot focuses on developing foundation models and hardware for embodied AI.
- Key investors include major internet companies like Meituan, Alibaba, ByteDance, and Xiaomi, as well as state-backed funds.
X Square Robot, a Chinese startup specializing in embodied artificial intelligence, has achieved a significant milestone by reaching a post-investment valuation of over 20 billion yuan ($2.9 billion). This valuation was attained across four completed funding rounds, with the latest being a Series C. The company, founded in 2023 and based in Shenzhen, is noted as the first embodied intelligence firm in the Guangdong-Hong Kong-Macao Greater Bay Area to achieve this valuation.
The surge in investment highlights a broader trend of capital migration from robot hardware to the AI models that power them, a shift particularly evident in China's robust robotics hardware supply chain. X Square Robot's funding history includes lead investments from major internet companies such as Meituan, Alibaba, ByteDance, and Xiaomi, with Xiaomi's strategic investment arm showing particular conviction by reinvesting across multiple rounds.
Recent funding rounds have seen the entry of state-level capital, including the National Artificial Intelligence Industry Investment Fund and China Insurance Investment, alongside industrial strategics like Honor and Chery Automobile. These investments signal a national strategic priority for embodied intelligence and provide a de-risking floor for the company's valuation. Industrial investors are motivated by potential supply-chain integration and direct deployment use cases in sectors like home services, automotive manufacturing, and logistics.
Technologically, X Square Robot has developed the WALL-B embodied AI foundation model, built on its World Unified Model (WUM) architecture. This model integrates perception, language, action, and physical prediction within a single network for enhanced multimodal understanding and spatial reasoning. The company has also open-sourced components like WALL-OSS-0.5 and WALL-WM, demonstrating strong autonomous task completion capabilities. A proprietary model-driven data pipeline automates data collection and cleaning, facilitating rapid model iteration and the generation of high-quality datasets for complex scenarios. The company is actively deploying its hardware and model stack across household, industrial, and logistics environments.