National AI Pilot Base in Hangzhou Bridges Gap Between Robotic Prototypes and Industrial Deployment
跨越样机与车间鸿沟，浙江具身智能探寻中试突围之路
The National Artificial Intelligence Application Pilot Base in Hangzhou serves as a national-level platform to validate embodied AI technologies in real-world factory settings. It offers shared testing infrastructure, domestic NPU-based world models that reduce simulation costs by 70%, and standardized data pipelines for manufacturing sectors like textiles and logistics. The base addresses challenges such as high computational costs from imported GPUs, lack of real-scenario data, and the need for composite talent by providing public testing labs, industry standards development, and vocational training centers.
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
robot_bodies
100 robots
open_source_models
3
open_source_dataset_size
300
real_robot_data_collection_hours
2,800 hours
View original evidence
基地汇聚了40个国内外主流开源数据集，总时长超3万小时，完成2800小时真机数据采集；同时探索市场化路径，联合华数传媒等设立浙江具身智能数据运营有限公司，计划年底前完成5万小时具身智能数据集建设，面向全省制造企业、科研机构提供数据服务。
Open source S12planned_dataset_construction_hours
50,000 hours
View original evidence
基地汇聚了40个国内外主流开源数据集，总时长超3万小时，完成2800小时真机数据采集；同时探索市场化路径，联合华数传媒等设立浙江具身智能数据运营有限公司，计划年底前完成5万小时具身智能数据集建设，面向全省制造企业、科研机构提供数据服务。
Open source S12open_source_datasets_count
40
View original evidence
基地汇聚了40个国内外主流开源数据集，总时长超3万小时，完成2800小时真机数据采集；同时探索市场化路径，联合华数传媒等设立浙江具身智能数据运营有限公司，计划年底前完成5万小时具身智能数据集建设，面向全省制造企业、科研机构提供数据服务。
Open source S12Grasp-and-place success rate
94.6%
Reported trials
View original evidence
此外，基地联合浙江大学成立具身智能基础模型实验室，开展具身智能高阶任务规划模型训练，相关成果已在四足机器人上开展应用验证，导航、抓取、放置等多任务综合成功率达94.6%。
Open source S10
- paper: not_reported
What remains unknown
Not established in the collected evidence: Environment, Control, Data origin.
Reported performance applies to the described task. It does not establish general autonomy or deployment readiness.
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
目前，基地集聚了超100款机器人本体、24款本体组件、3款开源大模型、近300TB开源数据集，面向全省创新主体开放共享。
Open source S18
基地汇聚了40个国内外主流开源数据集，总时长超3万小时，完成2800小时真机数据采集；同时探索市场化路径，联合华数传媒等设立浙江具身智能数据运营有限公司，计划年底前完成5万小时具身智能数据集建设，面向全省制造企业、科研机构提供数据服务。
Open source S12
此外，基地联合浙江大学成立具身智能基础模型实验室，开展具身智能高阶任务规划模型训练，相关成果已在四足机器人上开展应用验证，导航、抓取、放置等多任务综合成功率达94.6%。
Open source S10
Source：Robotics — Chinese web discovery · tidenews.com.cn