China’s robot-hand unicorn Linkerbot is hunting a $6bn valuation

Two years after a Beijing engineer started building dexterous hands inspired by a Japanese cartoon, his company holds 80% of the global market and is doubling its valuation in months.

Robotic hands are not, conventionally, the part of a humanoid robot that investors get excited about. The legs walk, the arms lift, and the head, increasingly, talks. The hands have always been the difficult part, fiddly bits of engineering with too many moving pieces and not enough commercial pull. For most of the past decade, capable, dexterous hands have been research projects, not businesses.

Linkerbot, a Beijing-based startup founded in 2023, is testing whether that has changed.

According to a Reuters report on 4 May, the two-year-old company is preparing to raise its next funding round at a $6bn valuation, double the $3bn it secured in a Series B+ closed only days earlier.

Reuters describes Linkerbot as the global market leader in high-degree-of-freedom dexterous hands, with more than 80 per cent share by volume and the only operation in the world capable of shipping above 1,000 units a month. Production, the company told Reuters, is on track to scale from roughly 5,000 units monthly to 10,000.

That is a striking pace for a category that, on most observers’ priors, was not a market until very recently.

From Doraemon to dexterity

Linkerbot’s chief executive, Alex Zhou Yong, told the South China Morning Post that the idea began in childhood with Doraemon, the Japanese cartoon cat whose magical pocket produces an inexhaustible supply of gadgets. Zhou, who studied robotics, eventually concluded that the trick was not the pocket but the hands. A robot that can manipulate objects with anything close to human dexterity does not need an infinite toolkit; it can use ours.

The company’s flagship Linker Hand series spans six to 42 degrees of freedom, the engineering measure of independent movement axes, and uses, according to Reuters, all the major actuation methods used in the field. The lightweight O6 model weighs 370g and is rated to handle a 50kg load. Higher-DoF variants, such as the L30, are aimed at research labs and humanoid integrators willing to pay for finer-grained control.

In demonstrations, the hands turn screws at speed, grasp soft objects without crushing them, thread a needle, and carry out high-precision assembly tasks. None of these is a particularly novel feat for a research prototype, earlier robot hands have managed similar tricks, and Linkerbot is far from the only firm chasing the problem; Swiss startup mimic is using generative AI to teach a humanoid hand by watching humans work. The point, in Linkerbot’s pitch, is that the demos are now happening on a production line.

Early backers include Ant Group, the financial-services arm of Alibaba, and HongShan Group, the Chinese spin-out of Sequoia Capital. The latest round added the state-backed Zhongguancun Science Park Fund, Bank of China Asset Management, and Fosun Capital. The mix, private internet capital alongside state vehicles and a major bank, is a fair summary of where Chinese deeptech funding is concentrated in 2026.
It also reflects a wider pattern.

It also reflects a wider pattern. Investor interest in Chinese humanoid robotics has surged this year, helped along by viral demonstrations including the Beijing humanoid robot half-marathon in April, where a Tien Kung Ultra unit beat the human world record by seven minutes, and a televised performance featuring Unitree’s machines.

Unitree itself filed for a Shanghai listing in March at a target valuation of around $7bn, having last raised privately at roughly $3bn. Other Chinese players, including Galbot, AgiBot, and AI2 Robotics, are clustered in the $2bn-$3bn range.

The intensity of the talent race in Chinese robotics has reached the point where rival UBTech has dangled an $18m package for a chief AI scientist. By that yardstick, a $6bn valuation for a company that makes only the hands is an aggressive bet.

It is also a bet that goes against a striking gap in the market. Figure AI, the leading US humanoid startup, raised at a $39bn valuation in September 2025, despite shipping a fraction of the volume its Chinese counterparts manage. The disparity reflects, in part, what each side is being priced as.

American investors treat humanoid companies as artificial-intelligence platforms; Chinese investors price them more cautiously, as industrial hardware businesses. Linkerbot, which sells the hands rather than the whole robot, sits awkwardly between those framings.

Hands that build hands

The company’s expansion is more concrete than most pitches at this stage of the AI cycle. Reuters reports more than 400 employees and five factories across Beijing and Shenzhen. The most quietly ambitious detail, mentioned almost in passing, is that some of those factories are being designed as intelligent production lines on which Linkerbot’s robotic hands assemble more robotic hands.

If that pans out, it would be one of the first commercial demonstrations of dexterous manipulation moving from research into closed-loop manufacturing, akin to the Siemens-Nvidia factory deployment in Erlangen but turned inward, and a useful internal proof point for would-be customers.

Linkerbot also says it has built what it calls LinkerSkillNet, described as the largest real-world dataset of dexterous manipulation skills in operation, with more than 500 captured behaviours. The figure has not been independently audited, but the broader claim, that the company is accumulating proprietary training data of the kind that will shape next-generation manipulation models, is consistent with how rivals describe the strategic prize.

Whether $6bn is too much for a hand maker, however dexterous, will depend on which market thesis turns out to be correct. If humanoid robots remain industrial niche products, Linkerbot’s component-supplier role caps its upside, however large its market share.

If, on the other hand, the field follows the trajectory China’s investors increasingly assume, with humanoids moving into logistics, services, and eventually domestic environments, as 1X is already attempting with its NEO home robots, hand suppliers with deep manufacturing capability and proprietary data could capture a disproportionate share of the value chain.

There are also softer risks: the geopolitical exposure that comes with state-fund investors and a Beijing-Shenzhen footprint at a moment of renewed US scrutiny of Chinese deeptech, and the possibility that humanoid platforms eventually integrate hand designs in-house.

Tesla, which is preparing Shanghai Gigafactory for Optimus mass production, has signalled it intends exactly that, and Meta’s acquisition of Assured Robot Intelligence hints that platform players want the whole robot, not a third-party bill of materials. Linkerbot’s bet is that being early, large, and the only volume supplier in its category will be enough to outrun those risks.

Investors will price that bet in the next round. For now, a company that started with a cartoon cat and a hunch about hands is one of the more closely watched stories in the global robotics market, and one of the few in which the headline numbers, market share, monthly volume, and employee count are not just promises.