Dexterous hands at scale
Linkerbot is producing nearly 5,000 dexterous robotic hands a month and claims 80% of the high-DOF market.
Humanoids · Supply Chain · Components · China Robotics · Dexterous Hands
Linkerbot is interesting because it sits in a narrow part of robotics that usually gets treated as a subsystem.
The specialist supplier position it is building already exists elsewhere in the robotics stack. Harmonic Drive dominates strain wave gearing across competing platforms. Maxon supplies precision motors without betting on which OEM wins. The advantage in both cases came from serving the whole market rather than one customer. Linkerbot is tracing the same path in dexterous manipulation; the company claims over 80% global market share in high-DOF robotic hands, manufactures joint modules, motors and reducers in-house, and is accumulating manipulation data across a widening customer base.
The engineering problem it is solving is underappreciated relative to the attention the full humanoid body receives. A human hand has approximately 27 degrees of freedom. Most commercial robotic grippers have between one and six. That gap is the difference between a robot that can move a box and one that can thread a needle, turn a screw into a recessed hole, or handle deformable materials without crushing them. Many companies building humanoids have either deprioritised the hand or described it as one of their hardest subsystems.
Source: Linkerbot
Reuters In an exclusive report this week shared that Linkerbot produces nearly 5,000 high-DOF hands per month across five factories in Beijing and Shenzhen, and plans to scale toward 10,000. The company delivered over 10,000 units across 2025. Customers include Cambridge, Stanford, Peking University and Tsinghua, alongside undisclosed industrial operators.
That implies a sharp shift in production tempo. Linkerbot delivered over 10,000 hands across 2025; it now says it is producing nearly half that number every month. If the current run-rate holds, the company has moved from early volume into something closer to supplier-scale manufacturing.
More interesting than the production figures is how customers are actually deploying the product. Many mount Linkerbot hands onto existing industrial robotic arms rather than full humanoid bodies. Zhou described the logic directly in a Reuters interview; Chinese factory owners are pragmatic, and for most factory work, two arms and a pair of dexterous hands are enough. That reframes the addressable market. This means Linkerbot's volumes are not tied only to humanoid OEM sales cycles, and a large share of its demand may come from customers that never buy a full humanoid body. The factory arm market is larger and more immediate than the humanoid market, and Linkerbot is already selling into it.
LinkerSkillNet is the second layer of the business; a multimodal data collection system that converts human manipulation into standardised reusable robotic capabilities, currently holding over 500 skills, with Zhou expecting the library to double every six months. The stated goal is one million hands mastering one million skills. Each hand deployed in a new environment feeds real-world data back into the platform. Hardware volume and manipulation data compound together, which is the structural bet underlying the valuation.
On funding, Linkerbot raised RMB 1.5 billion, approximately $210 million, in February 2026, then closed a Series B+ last week at a $3 billion valuation backed by Zhongguancun Science Park Fund, Bank of China Asset Management and Fosun Capital, with Ant Group and HongShan as earlier investors. The company is already targeting $6 billion in its next financing. For context, Unitree, one of the highest-volume full humanoid manufacturers, filed for a Shanghai IPO in March to raise about $610 million. It was previously reported that Unitree had been eyeing an IPO valuation of up to $7 billion.
Alex Zhou founded Linkerbot in 2023, though the idea had been forming since 2018-2019 when he began tracking the early signals of embodied intelligence. He graduated from Huazhong University of Science and Technology's Youth Class at fourteen, a selective programme for academically exceptional teenagers that has produced a notable share of China's technology founders. His previous company reached 300 million users; fewer than 1% were Chinese. That distribution shaped how he thinks about building markets, and it shows in how Linkerbot is positioning now; manufacturing its own components, serving research institutions and industrial operators simultaneously, and accumulating data across all of them rather than betting on one customer or one platform winning.