Elon Musk says building human-like robot hands is “shockingly difficult.” TrendForce’s latest research shows why: Tesla’s pursuit of near-human dexterity has made manufacturing the toughest bottleneck for Optimus. Its hands still depend heavily on manual assembly, and tactile sensor reliability is still being improved. Tesla is responding with a modular, swappable sensing glove planned for its 2027 model.
Dexterous hands decide whether humanoid robots can move from demos to real-world work. As the end-effector, the hand directly determines whether a robot can grasp, pinch, use tools, and perform fine manipulation. Over the long term, dexterous hands are expected to account for 20% to 30% of a humanoid robot’s total hardware cost, second only to the body actuation system, with tactile sensors, coreless motors, reduction gears, and lead screws among the highest-value components.
TrendForce projects that global humanoid robot sales will exceed 2 million units by 2030, potentially driving the dexterous hand market to around US$6 billion. Capital is already moving. As of 1H26, China had nearly 510 companies related to dexterous hands. While investment in China’s humanoid robot sector still centers on the robot body, capital is increasingly flowing into key components such as dexterous hand modules, sensors, and chips. For instance, three core dexterous hand module makers, LinkerBot, AGILINK, and GeniTech, raised a combined RMB 5.3 billion (about US$790 million) in 1H26.
This article takes a closer look at the technical routes and progress of specialized dexterous hand makers and robot makers such as Tesla, 1X and Boston Dynamics.
Related report: Humanoid Robot Market Bulletin - Oct. 8, 2026