JD.com Unveils Embodied AI Data Collection at Wuzhen Summit
让机器人在真实世界中学习！京东乌镇峰会披露具身数采已覆盖100余个细分场景
At the 2026 World Internet Conference Wuzhen Summit, JD.com Technical Committee Chairman Cao Peng announced plans to build the world's largest embodied AI data collection center. The company aims to collect over 10 million hours of high-quality human real-world scenario video data and over 1 million hours of robot operation data within two years. Currently, JD's embodied data collection covers over 100 niche production service scenarios with more than 100,000 real task operations collected.
Source: Robotics — Chinese web discovery · Read original article ↗
English translation unavailable. Choose the original article to read it.
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
human real-scenario video data hours
>10,000,000 hours
View original evidence
围绕“云、数、模、端、场、链”六大模块构成的超级AI供应链，京东计划建成全球最大的具身智能数据采集中心，两年内目标采集超过1000万小时高质量人类真实场景视频数据，并同步采集超过100万小时机器人本体运行数据。
Open source S3robot body operation data hours
>1,000,000 hours
View original evidence
围绕“云、数、模、端、场、链”六大模块构成的超级AI供应链，京东计划建成全球最大的具身智能数据采集中心，两年内目标采集超过1000万小时高质量人类真实场景视频数据，并同步采集超过100万小时机器人本体运行数据。
Open source S3real-world operational tasks
100,000
Reported trials
simulation evaluation efficiency multiplier
3.2
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
围绕“云、数、模、端、场、链”六大模块构成的超级AI供应链，京东计划建成全球最大的具身智能数据采集中心，两年内目标采集超过1000万小时高质量人类真实场景视频数据，并同步采集超过100万小时机器人本体运行数据。
Open source S3
目前，京东具身数采已覆盖100余个细分生产服务场景，采集超10万真实作业任务。
Open source S4
在已完成的测试中，仿真评测与真实评测结果的一致性约为90%，评测效率达到真机评测的3.2倍。
Open source S14
在推理与模型服务方面，京东云自研推理引擎，大幅提升高并发场景下的吞吐与资源利用效率，典型任务推理成本最高降低80%。
Open source S19
Source：Robotics — Chinese web discovery · sohu.com