Chinese startups are racing to solve one of the hardest problems in robotics: creating fully dexterous robotic hands. These companies aim to transform humanoid robots from dancing gimmicks into useful products capable of everyday tasks.
The complexity of human hands
Human hands are exceptionally complex, requiring intricate neurological instructions for tasks like tying shoelaces or buttoning a shirt. No machine has truly replicated this tool, but advances in artificial intelligence (AI) are bringing companies closer.
China's government has emphasized "embodied AI" in development plans. In May, the Chinese Communist party's journal Qiushi reported that embodied-intelligence robots are among sectors opening up new trillion-yuan markets.
Chinese startups lead the charge
A new suite of Chinese startups is leveraging China's manufacturing advantages and enthusiasm for embodied AI. LinkerBot, founded in 2023 by Zhou Yong, produces about 5,000 hands monthly and plans to double that figure, chasing a $6bn valuation. Zhou says making a robotic hand is "one hundred times more difficult" than making a humanoid, with dexterity ten times that of other body parts but only one-tenth the volume.
Zhou aims to make mass-market prosthetic hands for amputees at a fraction of the current price, targeting $1,000 per hand versus tens of thousands of dollars today.
Hardware advantages in China
China's cheap, sophisticated manufacturing supply chain gives it an edge in hardware. The electric vehicle industry has produced companies capable of making components like lithium-ion batteries and miniaturized motors at scale. Pan Yunzhe, founder of Wuji Technology in Shenzhen, says sourcing components in China was easier than in the US, where he graduated in 2018. "It was really impossible to do hardware in the United States because the supply chain problem is just so constraining," he says.
Software challenges remain
The more challenging problem is software: teaching hands how to perform tasks. Nathan Lepora, a professor of robotics and AI at the University of Bristol, says: "The challenge of making these hands is getting solved now. Controlling them, now that's a whole different game … nobody knows how to do that."
Startups are using teleoperation—controlling machines remotely—to harvest data for training spatial intelligence models. Unlike large language models, 3D models lack abundant data sources. Wuji's flagship product, the Wuji glove, is a sensor-filled wearable that collects movement data and information about pressure and touch, essential for tasks like cracking an egg without crushing it.
Industry growth and future vision
China has registered over 1 million robotic companies, with 2025 registrations up 40% year-on-year. The dexterous hand industry surpassed 50bn yuan ($7.4bn) in 2025, up from 13bn yuan in 2024, according to Chinese media. Pan says focusing on hands is critical because "the problem of manipulation is much more important than the problem of locomotion."
LinkerBot's Zhou dreams of a factory where robotic hands build more robotic hands, a self-perpetuating loop with minimal human input. "We are not creating robots to replace labour," he says. "We are creating robots so that humans can live a better and more prosperous life."
