CHINGMU MotionDecode Data Open Program

1,000 Hours of High-Quality Human Motion Data – Free to Use

On June 16, 2026, the 2026 Zhangjiang EAI Embodied AI Developer Conference & Supply Chain Summit was held at the Zhangjiang Science Hall in Pudong, Shanghai. Themed "Embodied AI, Chain the Future," the event brought together 30+ industry experts, 100+ leading enterprises across the core supply chain, and 2,000+ professionals. As a global leader in motion capture, CHINGMU made a strong presence — not only with cutting-edge Embodied AI tech demos on-site but also with the launch of the MotionDecode Data Open Program, offering 1,000 hours of high-quality human motion data free to applicants worldwide, injecting powerful data momentum into Embodied AI foundation model training and robot motion learning.

01丨   MotionDecode Data Open Program: Powering Embodied AI with 1,000+ Hours of Data

As Embodied AI accelerates toward industrial deployment, "data-driven" has become an industry consensus. Enabling robots to understand human motion and adapt to the real world is the fundamental prerequisite for this leap.

MotionDecode is CHINGMU's high-quality multimodal motion data system built for robot enterprises, Embodied AI algorithm teams, university research labs, simulation platforms, and model development teams. The first release offers 1,000 hours of high-quality human motion data free of charge, covering core use cases including robot training, motion generation, simulation validation, and Embodied AI model learning — driving the industry from isolated tech exploration to collaborative data ecosystem building.

MotionDecode Key Features:

• Larger Scale: 3,000 hours built to date – 1,000 hours free in Phase 1 – with an estimated 500,000 hours/year production capacity

• Higher Precision: Captured by CHINGMU's optical mocap system – sub-mm spatial positioningand joint angle error < 0.5°

• More Modalities: Body｜Hand｜Ego｜Exo｜Object 6D｜Scene

• Broader Scenarios: Home｜Industrial｜Office｜Outdoor

• Easier to Use: Supports 10+ standard formats including BVH, FBX, CSV

MotionDecode will significantly lower R&D barriers for the Embodied AI industry, providing a unified, high-quality training data benchmark for global researchers and enterprises. It fills the industry gap in large-scale, high-precision, full-scenario human motion data, empowers breakthroughs in full-body motion control for humanoids, dexterous hand manipulation, and human-robot collaborative interaction, and accelerates Embodied AI from validation to mass deployment – helping China build dual advantages in both data and technology on the global stage.

Join the MotionDecode Early Access & Feedback Program. Phase 1 (1,000 hours of high-quality human motion data) is now open for free application via QR code, with continuous updates. More sample data will be released on GitHub, Hugging Face, and the CHINGMU official website – stay tuned.

02 丨 Industry Leaders Gather to Explore Embodied Data Innovation

On the morning of June 16, the 2026 Zhangjiang EAI (Zhangjiang Embodied AI Supply Chain Conference) officially opened. At the opening ceremony, CHINGMU, as a representative of robotics industry application enterprises, joined over a dozen partner companies in the Zhangjiang Robot Valley Ecosystem Alliance collective signing ceremony. This marks a further refinement of Zhangjiang's Embodied AI robotics industry ecosystem, strengthening regional supply chain synergy and building a more solid foundation for more robot products to move from R&D validation to large-scale deployment.

At the main forum's panel discussion, Zhang Haiwei, Founder & CEO of CHINGMU, joined industry experts for an in-depth conversation on the theme "The 'Data Foundation' and Evolution Engine in the Embodied AI Era." Zhang emphasized that high-quality motion data directly determines the effectiveness of robot fine manipulation and intelligent interaction – and that CHINGMU will continue to focus on optical mocap innovation to provide a solid data infrastructure for the Embodied AI industry.

In the afternoon, the "Embodied AI Data Infrastructure & Technology Innovation" sub-forum – co-hosted by Shanghai Zhangjiang Group and the National & Local Co‑built Humanoid Robot Innovation Center, and co-organized by CHINGMU and Zhejiang DexRobot – was held concurrently. Six speakers from industry and academia took the stage: Zhang Haiwei (CHINGMU Founder & CEO) , Zhou Chen (DexRobot CEO) , Gong Xiangfeng (FAIR Robot CEO) , Ye Qi (Zhejiang University, "100 Talents" Professor) , Shen Keli (Shanghai University, Embodied AI Expert) , and Ji Yaohui (DexDecode CTO) . They shared cutting-edge insights across the full chain – from data acquisition and dexterous manipulation to motion capture and standardized execution.

Immediately after, the panel discussion "Digital Foundation, Embodied Future – Building Universal Data Infrastructure for the Embodied AI Era" was moderated by Dr. Shen Keli, with Zhang Haiwei, Zhou Chen, Gong Xiangfeng, and Ji Yaohui joining as panelists. They explored cross‑disciplinary perspectives spanning data services, core hardware, mocap technology, and full‑system deployment. The consensus: Embodied AI progress depends on systemic synergy of data supply, hardware capability, algorithm iteration, and ecosystem development. High‑quality, multimodal real‑world data is the core production asset for industrial deployment, and robust data infrastructure is the essential prerequisite for moving from validation to mass adoption.

Beyond the forum, CHINGMU's presence at the C43 Black Tech Zone drew significant attention. The booth featured dedicated data acquisition and robot teleoperation demo zones, showcasing the full data factory collection workflow of the high‑precision optical mocap system, and visually demonstrating how motion data translates into fine robot control commands. The immersive experience gave attendees a close‑up view of the real‑world application value of Embodied AI data infrastructure.

Data as a foundation – intelligence for the future. Innovation through collaboration – leadership through change. This conference, through a complete mapping of data acquisition, dexterous hands, multimodal interaction, mocap, and standardized execution, fully revealed the underlying logic of how data infrastructure drives Embodied AI into the real world. The launch of MotionDecode marks a new phase for the Embodied AI industry: data‑driven, ecosystem‑built. Moving forward, CHINGMU will continue to deepen technology R&D and data production capacity, partner with upstream and downstream players to refine Embodied AI data standards and ecosystem, and drive Embodied AI from exploration to mass deployment – empowering industries across the board.