RoboticsBusiness & Market 🇨🇳 14.08.2026 11:02

Delta Intelligence and WuJi Tech Sign Strategic Cooperation to Standardize Whole-Body Coordinated Dexterous Manipulation Data Collection

Delta Intelligence, a native humanoid robot foundation model company, and WuJi Tech, a dexterous hand hardware company, have signed a strategic cooperation agreement. They will focus on data collection and model training for whole-body coordinated dexterous manipulation, aiming to launch a unified data standard and model training baseline within the year.
Delta Intelligence (Delta Intelligence) and WuJi Tech have officially signed a strategic cooperation agreement on August 14, aiming to collaborate on data collection and model training for whole-body coordinated dexterous manipulation, building a complete pipeline from data collection to model training to dexterous operation. They plan to release a joint 'Whole-Body Coordinated Dexterous Manipulation Data Standard' and 'Model Training Baseline' within the year. The cooperation addresses the issue that existing dexterous manipulation data mostly only records hand degrees of freedom, neglecting the coordination with torso posture, arm trajectories, and visual focus, leading to unrealistic whole-body postures. Delta Intelligence specializes in native humanoid robot foundation models (HFMs) and native 3D world engines, with full-stack capabilities. WuJi Tech focuses on dexterous hand hardware, including direct-drive high-DOF dexterous hands and lightweight data collection gloves. The two companies will collaborate on two main directions: standardizing the data collection system using WuJi Tech's lightweight data gloves and Delta's foundation model training capabilities, and deep adaptation of hardware and algorithms for deployment. This collaboration aims to solve the data scarcity problem and reduce development costs for the industry, providing a common foundation for applications in industrial manufacturing and commercial services.
Abbreviations
HFMs = Humanoid Foundation Models — Гуманоидные фундаментальные модели
Source: QbitAI 量子位 — original
Our earlier posts on this topic ↓
Fresh news