Beijing Ruierman Intelligent Technology Co., Ltd. (hereafter "Ruierman") has officially kicked off its A-share listing process. According to the CSRC website, the company has signed a listing guidance agreement with Guotai Haitong Securities, targeting a ChiNext board listing on the Shenzhen Stock Exchange. The guidance period is set to commence in August 2026. However, shortly after the initiation of this IPO guidance, news emerged that Beijing's first humanoid robot data training center has recently ceased operations.
Located in Shougang Park, Shijingshan District, Beijing, the center was inaugurated in March 2025. Spanning 3,000 square meters, it had deployed 108 robots across ten major realistic scenarios including home healthcare, special operations, new retail, and automotive assembly. Local authorities in Shijingshan had projected the center would facilitate annual data transactions exceeding one million entries and cultivate a 5-billion-yuan robot data service market within three years. Crucially, Ruierman served as the core equipment and technology provider for this data training center.
Public records show that Ruierman was established on January 8, 2018, with a registered capital of 60.4 million yuan. Controlling shareholder Zheng Suibing holds a 42.96% stake in the company. Positioning itself as a global system-level infrastructure platform for the embodied intelligence era, Ruierman's core offerings include the GLN remote operation network, fully self-developed joint modules, and ultra-lightweight humanoid robotic arms. The company has launched three humanoid robot models. It achieved operational profitability in 2025 and completed a strategic financing round of nearly 500 million yuan in March 2026.
Recently, Ruierman has withdrawn from the Beijing project. According to a report from Beijing Daily, a representative from Zhongguancun Tongli, the project operator, stated that following the original partner's relocation, the center is undergoing renovations aligned with Shijingshan District's overall embodied intelligence industry layout. The focus will shift towards 4D Gaussian light field and world model technology routes. In August 2025, Ruierman's official WeChat account published a post about hosting an Open Day at the Shijingshan data training center, where it ambitiously declared its goal to create a "China solution" for embodied intelligence robot training grounds.
The Open Day event drew nearly 40 representatives from industry, academia, and research institutions, including Shijingshan District Deputy District Mayor Yu Jie, Beijing Institute for General Artificial Intelligence, the National Robot Evaluation Center headquarters (also known as Shanghai Robot Industry Technology Research Institute), Shandong University's School of Control Science and Engineering, Pengcheng Laboratory, Shanghai AI Laboratory, and Changzhou Testing & Inspection Institute for Standards and Certification (also known as the National Intelligent Industrial Robot Industry Metrology Testing Center). The event also invited partner enterprises like Alibaba Damo Academy, JD.com, Beijing Haitian Ruisheng, Max, Beijing Data Exchange, RoboGround, Zhixing Robot, YuanKe Vision, and Zhongguancun Tongli to tour the park and experience the training center's innovative applications firsthand.
Given that Ruierman's headquarters are also located there, it remains to be seen whether this shutdown and adjustment of the data center will impact the company's cooperative relationships with government bodies and industry partners.
Ruierman cites this as a model upgrade in response to investor concerns about the shutdown, the company's board secretary explained that the Beijing center's adjustment marks a significant upgrade in its data strategy. Describing the Beijing center as the 1.0 version of its data collection efforts, which used a centralized deployment model, the company acknowledged it successfully completed the validation journey from zero to one, accumulating tens of millions of trajectory data segments. However, it fundamentally remained a "laboratory scenario."
Ruierman believes that for embodied intelligence to genuinely "do real work," data must originate from real production and living scenarios, not simulated environments. Based on this assessment, the company has shifted its focus to the Changzhou Embodied Intelligence Data Experiment Platform (Data Collection Center 2.0). This platform, officially operational in May 2026, leverages the GLN (Global Labor Network) remote operation network. The plan is to deploy over one hundred wheeled humanoid robots into real work environments to complete more than 1,000 types of authentic operational tasks. Currently, over 30 robots have been deployed on a larger scale in real-world settings like factory production lines and warehouse logistics. Company representatives stated that the 2.0 version signifies a new phase, marking the transition of data collection from "effect demonstration" to "large-scale data delivery from real-world scenarios."
This adjustment aligns with the policy direction of the "2026 Special Action for Humanoid Robots and Embodied Intelligence Real-Scene Training," jointly initiated by the Ministry of Industry and Information Technology and the State-owned Assets Supervision and Administration Commission. For Ruierman, which is currently in the IPO guidance phase, its data system is a critical component of its core technology narrative. ChiNext review processes impose high requirements on hard-tech enterprises regarding technological iteration, business sustainability, R&D investment, and commercial model execution. This shift in data infrastructure means the company will face restructuring in data acquisition costs, collection paths, and R&D resource allocation.
Moving data sources from self-built large-scale simulation venues to dispersed real-world business scenarios offers the advantage of data being closer to actual industry needs. However, the distributed collection model also introduces new challenges in areas such as maintenance, operations, and data quality control.