Human hands, those nimble, nerve-filled appendages constituting the most flexible component of the human skeleton, are exceptionally complex. Tasks as mundane as tying shoelaces or buttoning a shirt demand a sophisticated orchestration of neurological instructions and precise choreography, yet they are performed largely without conscious thought. For millennia, no machine has managed to replicate this quintessentialquintessential/ˌkwɪntɪˈsɛnʃəl/L3典型的；精髓的；最完美的代表Representing the most perfect or typical example of a quality or class. human tool—until now. As artificial intelligence accelerates, a cohort of Chinese companies, leveraging the country’s manufacturing prowess and governmental enthusiasm for “embodied AI,” contends it is on the verge of surpassing this final, formidableformidable/ˈfɔːrmɪdəbəl/L3令人敬畏的；难以克服的；强大的Inspiring fear or respect through being impressively powerful, intense, or capable; difficult to overcome. hurdle in robotics. These firms aim to transform humanoid robots from dancing novelties into genuinely useful products.
China’s fixation on robotics is not merely cultural spectacle; it is underpinned by profound demographic and economic imperatives. With an ageing and shrinking workforce, technologists and policymakers view humanoid robots as pivotal to unlocking future economic potential. The Chinese Communist Party’s theoretical journal, Qiushi, recently highlighted “embodied-intelligence robots” as a sector capable of opening up new trillion-yuan markets. Notwithstanding this fervour, the International Federation of Robotics cautioned last September that true multipurpose humanoids remain distant. The principal obstacle, as Elon Musk observed, is the hand: it represents the majority of a robot’s engineering difficulty, given its requirement for dexteritydexterity/dɛkˈstɛrɪti/L3灵巧；熟练；敏捷（尤指手部）Skill and ease in using the hands or body; mental adroitness or cleverness. ten times that of other body parts within a volume one-tenth their size.
Entrepreneurs like Zhou Yong of LinkerBot and Pan Yunzhe of Wuji Technology are betting that China’s manufacturing ecosystem can solve both hardware and software challenges. Zhou, inspired by Steve Jobs’ dictum on focus, launched LinkerBot in 2023 and now produces 5,000 hands monthly, targeting a $6bn valuation; his ambition includes mass-market prosthetic hands at $1,000 each, a fraction of the current cost. Pan, who returned from the US citing insurmountable supply chain constraints, argues that the problem of manipulation supersedes that of locomotion: humanoids are useless until they can wield tools. The dextrous hand industry in China has already surpassed 50bn yuan ($7.4bn), up from 13bn yuan in 2024, reflecting a surge in registrations of over one million robotics companies.
The more intractableintractable/ɪnˈtræktəbəl/L3难处理的；棘手的；难以控制或改变的Hard to control, deal with, or solve; stubborn or unmanageable. challenge, however, lies in software: teaching robotic hands to perform tasks that humans find intuitive. Teleoperation, the remote control of robotic appendages, requires hundreds of training hours for even simple actions like packing groceries. Researchers are consequently developing sensor-laden wearables, such as Wuji’s glove, which capture movement, pressure, and tactile feedback—data essential for training spatial intelligence models. Unlike large language models, which feast on vast online text corpora, three-dimensional data sources remain scarce. As Nathan Lepora of the University of Bristol notes, controlling these hands is a “whole different game” for which no definitive solution yet exists.
Despite these obstacles, Chinese entrepreneurs remain resolute. Zhou envisions a self-perpetuating factory where robotic hands build more robotic hands, minimising human input. Further down the line, adequately dextrous robots could become fully-fledged household helpers, not to replace labour but to enable a more prosperous life. That said, the gap between ambition and execution remains vast; the fundamental problems of capturing human movement and tactile sensation are, as Pan concedes, “super complicated and not solved yet.” Nevertheless, with government backing, an agile supply chain, and a surging market, China is arguably best positioned to crack the code of the human hand—the most exquisite and elusiveelusive/ɪˈluːsɪv/L3难以捉摸的；难以实现的；逃避的Difficult to find, catch, achieve, or remember; evasive or subtle. of robotic frontiers.
