Gasgoo Munich- Recently, Yiren Technology has partnered with JD Group, China Machinery Engineering Group, Suqian High-tech Zone, and Suqian Technician College to establish an Embodied AI Research Institute. The partnership aims to build a comprehensive ecosystem for the embodied intelligence industry.
Once operational, Yiren Technology and its partners—including JD Group and China Machinery Engineering Group—plan to construct a full-chain ecosystem covering data supply, R&D, scenario application, and talent support. The goal is to align real-world scenarios with data required for robot algorithm training.
The institute will serve as a central hub for local large-scale data collection, coordinating on-campus training production lines with distributed collection points in urban communities.
Image Source: Yiren Technology
By establishing a "large-scale data collection" model in Suqian High-tech Zone, Yiren Technology, JD Group, and China Machinery Engineering Group are overcoming the limitations of traditional, closed data factories. The initiative opens standardized data collection channels to students, residents, industrial workers, and the flexible workforce.
Using lightweight collection equipment, participants can capture multimodal data on human movement and environmental perception in real-world settings, ranging from homes and workshops to industrial parks and logistics hubs.
As a major player in China's embodied AI data sector, Yiren Technology specializes in multimodal data collection, annotation, and simulation training. The company holds a full suite of proprietary core technologies and solutions for industrial deployment.
JD is rapidly expanding its embodied data collection operations. In March, the company announced plans to mobilize hundreds of thousands of people, including more than 100,000 internal employees and 500,000 external industry professionals. In Suqian alone, more than 100,000 residents are expected to participate, covering over 100 specific scenarios in homes, offices, factories, logistics centers, shops, restaurants, medical facilities, and sanitation sites.
JD aims to accumulate 5 million hours of real-world video data within one year, increasing the volume to over 10 million hours in two years. Concurrently, it plans to collect 1 million hours of data directly from robots.
The company's Suqian data collection center can currently accommodate nearly 10,000 concurrent operators. JD is also expanding into community living and industrial production, establishing a dual-track model combining "community crowdsourcing" with "fixed-point industrial collection."
