Chuguang Yueqian completes seed funding of hundreds of millions of yuan; Xiaomi Strategic Investment and BV Baidu Ventures participate

Leaderobot.com reported on Oct. 16, 2026, that Chuguang Yueqian had completed a seed round worth hundreds of millions of yuan. The investors included Chentao Capital, Xiaomi Strategic Investment, BV Baidu Ventures, Xinhe Capital, Yunji Tianze and Shanghai Qijixing, the report said. It also said the company was founded less than five months ago and that its founder, Zhang Ying, had been a core technology lead at Horizon HSD.

Chuguang Yueqian is said to have been founded by Zhang Ying, a former core technology lead at Horizon HSD, and to have received investment from firms including Xiaomi Strategic Investment and BV Baidu Ventures. The company’s approach brings end-to-end development experience from autonomous driving into robotics, suggesting that development methods for AI operating in the physical world may be a factor in investment decisions.

What stands out in the latest news about Chuguang Yueqian is less the funding amount than the combination of its investors and its founder’s background. The investor group, which includes Chentao Capital, Xiaomi Strategic Investment, BV Baidu Ventures, Xinhe Capital, Yunji Tianze and Shanghai Qijixing, points to support for developing AI that operates in the physical world, rather than an idea-stage project alone. However, the main source describes a seed-stage company and does not discuss production capacity or product supply arrangements. The development connects with a trend seen in our related articles, [Chuguang Yueqian raises hundreds of millions of yuan in seed funding; cumulative contract value reaches hundreds of millions of yuan](/news/nexform-shingen-karada-suusenman-motono-shiido-choutatsu-rui-0c462c) and China’s embodied-AI training grounds expand amid shortage of useful data: investment is flowing into embodied intelligence and Physical AI, but the substance of implementation is increasingly under scrutiny. In this case, the focus is on a person with end-to-end implementation experience in autonomous driving moving into robotics. Zhang Ying’s background suggests an approach that handles perception and control together being brought to Physical AI, but how far it can be generalized will need to be assessed through testing on individual tasks. From an industry-structure perspective, the news indicates that the point of competition is not computing infrastructure or hardware alone, but the ability to connect model design, real-world data and evaluation methods. The company’s performance at RoboDojo and its real-robot tea-making demo suggest that, beyond lab-based metrics, it is becoming important to show continuous operation and adaptation to the environment. However, the original source does not make clear what data Simate-beta was trained on, what boundary conditions it has, or how the seed funding will be allocated across research and product development. The next things to watch are the tasks being targeted, the consistency of evaluation metrics and the use of the funding.

In Japan, a key question is how companies and research institutions with experience in automotive AI and robot control will connect with the competition to develop systems that handle perception and control together. The main source reports funding and the movement of talent, not an advantage in finished products. The development therefore appears less like a direct threat to Japanese companies and more like a stage at which differences may emerge in evaluation methods and the creation of real-world data.