en.Wedoany.com Reported - On May 9, 2026, Beijing XiaoYu Intelligent Technology Co., Ltd., a general-purpose embodied intelligence technology company, officially announced the completion of hundreds of millions of yuan in Series B+ funding. This round of financing was jointly invested by three industrial capitals: BAIC Capital, Fosun RZ Capital, and C&D Emerging Industry Equity Investment. Existing shareholders Huaye Tiancheng and Guizhou Science and Technology Innovation Angel Fund co-invested, while Didi Chuxing and Xiaomi co-founder Li Wanqiang continued to make additional investments. Gengxin Capital served as the exclusive financial advisor and also participated in the investment. With this, XiaoYu Intelligence's investor lineup now spans the four core industrial sectors of consumer electronics, automotive, shipbuilding, and construction.

Founded in 2023 by Qiao Zhongliang, former head of Xiaomi's MIUI R&D, XiaoYu Intelligence's core team comes from leading tech companies such as Xiaomi, ByteDance, and Huawei. The company takes "One Brain, Multiple Forms" as its core technical pathway, dedicated to building a highly generalizable, general-purpose embodied intelligence brain that can adapt to control robot bodies of different forms, such as single-arm, wheeled dual-arm, and humanoid robots. This technical philosophy is rooted in Qiao Zhongliang's over a decade of software architecture experience at Xiaomi—he once led the system transformation of "develop once, deploy across multiple terminals," enabling the same MIUI system to be deployed on phones, watches, TVs, and other terminals. After founding XiaoYu Intelligence, he extended this universality mindset to the physical world: using a single general-purpose "brain" to drive robot "bodies" of different forms. However, Qiao Zhongliang is well aware that a general-purpose brain must first successfully run a data flywheel in a single scenario before cross-form migration is possible. According to his deduction, once "one brain, one form" is achieved in a certain scenario, the difficulty of cross-body adaptation for that scenario is expected to drop to around 10%.

Industrial welding was chosen by Qiao Zhongliang as the breakthrough point. After consulting the Ministry of Human Resources and Social Security for data on job gaps and conducting field research in shipyards and construction sites, he reached a clear industrial judgment: China's welder shortage for individual job types exceeds 10 million, the average age of welders has reached 45, the training cycle lasts as long as one to two years, and the working environment causes significant physical damage to operators. Welding happens to be one of the most complex scenarios for physical feedback—variables such as molten pool dynamics, metal shrinkage and expansion, and fume interference collectively constitute an information-rich environment, demanding extremely high capabilities in "seeing accurately, aligning correctly, and controlling steadily." Qiao Zhongliang's judgment is concise and direct: if a general-purpose brain can achieve a closed loop in welding, the most difficult scenario to run through, migration to adjacent processes like riveting, grinding, and spraying will naturally follow.

This judgment has already received preliminary validation on the industrial side. Starting from the fourth quarter of 2025, the welding intelligent agent developed by XiaoYu Intelligence entered customer production line deployment, with over a hundred prospective purchasing units. The company also reached a cooperation agreement with Tangshan Panasonic to jointly develop a large-model intelligent welding robot, and has completed mass production delivery of a hundred-unit order to a leading enterprise in the heavy industry sector. Just before the announcement of this funding round, XiaoYu Intelligence signed a strategic cooperation agreement with China Construction Science and Industry Intelligent Technology Co., Ltd. The two parties will jointly promote the large-scale application of 1,000-unit-level embodied intelligent welding robots in the construction steel structure field within the next year. Feng Qingchuan, head of China Construction Science and Industry Intelligent Technology, pointed out, "Steel structures come in a wide variety, and developing specialized programs for different working conditions involves heavy debugging workloads and long cycles; the industrial site environment is complex, demanding extremely high system stability and anti-interference capabilities. This is also why we chose to explore together with XiaoYu Intelligence—general-purpose embodied intelligence has the potential to fundamentally solve the problems of high adaptation costs and insufficient flexibility inherent in traditional solutions."

Qiao Zhongliang candidly admitted at the signing ceremony, "When we apply general-purpose embodied intelligence to complex industrial sites, we deeply feel a huge chasm: there is a natural barrier between those exquisite algorithm demonstrations in the lab and the requirement for round-the-clock, high-reliability, low-cost operation in the factory. We don't want to be 'lone heroes'; instead, we are eager to build an ecosystem and set standards together with like-minded partners, joining hands to face global competition." Concurrently, the Beihang-XiaoYu Industrial Embodied Intelligence Joint Laboratory, co-established by XiaoYu Intelligence and Beihang University, was officially inaugurated. Relying on Beihang's School of Computer Science and Engineering and the State Key Laboratory of Complex & Critical Software Environment, the lab focuses on core technology research in areas such as spatial perception native foundation models, fine-grained planning for long-horizon tasks, and high-fidelity simulation data generation.

Starting from welding, Qiao Zhongliang's expansion path is "fan-shaped"—extending to upstream and downstream processes like riveting, grinding, and spraying, covering core manufacturing sectors such as heavy industry, automotive, and consumer electronics. "In the next decade, fewer than 10 general-purpose embodied intelligence companies globally will remain to enter the final round," his judgment is clear. "The lack of a super hit product is the core pain point of the current industry, and the ability to create a hit product will become a key indicator for measuring embodied intelligence companies."

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com