China wants to solve the hardest problem in robotics – making hands

2026-07-06 · Source: AI (artificial intelligence) | The Guardian · Field: Technology & Digital — Robotics & Autonomous Systems, Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, medium

Summary

China is aggressively pursuing the development of highly dextrous robotic hands, considered the most challenging problem in robotics, to advance "embodied AI" and transform humanoid robots from novelties into practical tools. Companies like LinkerBot, founded in 2023, produce 5,000 hands monthly, aiming for a \$6 billion valuation and prosthetic hands at \$1,000. The Chinese dextrous hand industry reached 50 billion yuan (\$7.4 billion) in 2025, a significant increase from 13 billion yuan in 2024, driven by China's robust manufacturing supply chain, which facilitates component sourcing. While hardware development is progressing rapidly, the primary hurdle remains software: teaching hands complex manipulation tasks. This requires extensive data collection for 3D models, often through teleoperation or sensor-filled wearables like Wuji Technology's glove, to capture nuanced movement, pressure, and touch information, which is currently scarce.

Key takeaway

For AI Scientists and Robotics Engineers focused on developing practical humanoid robots, recognize that dextrous manipulation remains the primary bottleneck, far exceeding locomotion challenges. Your efforts should prioritize advanced software for hand control, specifically exploring innovative data collection methods like sensor-filled wearables to capture critical pressure and touch information. This approach is essential to move beyond basic movements and enable robots to perform complex, real-world tasks, transforming them into truly useful products.

Key insights

Developing dextrous robotic hands is the critical, most difficult hurdle for practical humanoid AI, with China leading the charge.

Principles

• Robotic hand dexterity is 100 times harder than humanoid locomotion.
• Hardware development for robotics benefits significantly from robust supply chains.
• Software control for manipulation requires vast, nuanced 3D spatial data.

Method

Training robotic hands involves extensive data collection via teleoperation or human-worn sensor devices like the Wuji glove to capture movement, pressure, and touch.

In practice

• Mass-market prosthetic hands at \$1,000 are a near-term goal.
• Robots performing household tasks like folding laundry or cooking.

Topics

• Robotics
• Humanoid Robots
• Dextrous Manipulation
• Embodied AI
• Robotic Hands
• AI Training Data
• Manufacturing Supply Chain

Best for: Investor, Research Scientist, Entrepreneur, Robotics Engineer, AI Scientist, Director of AI/ML

Related on AIssential

• Are robots nearing their ChatGPT moment? – podcast — China's massive investment and rapid advancements are pushing robotics towards a "ChatGPT moment" in dexterity and real-world utility.
• Can a Prosthetic Hand Deliver Humanoid Dexterity? — Integrating human-proven AI bionic hands into humanoids enhances factory-ready durability and dexterity.
• Video Friday: Robotic Motion Discovery Reveals Unusual Behaviors — Robotics innovation spans humanoid loco-manipulation, domestic assistance, space operations, and open-source hardware for research.
• Why Touch Is Becoming a Big Deal Again in Robotics — Touch sensing is crucial for robots to understand physical interaction, enabling robust manipulation and adaptive control beyond vision.
• AI is gonna take your job and your girl. — Advanced robotic hands like the Linker Hand L30 demonstrate increasing dexterity for complex manipulation tasks.
• Why Do Humanoid Robots Still Struggle With the Small Stuff? — Despite AI advancements, humanoid robots still struggle with fundamental physics like force control for reliable real-world interaction.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI (artificial intelligence) | The Guardian.