NextFin News - China’s robotics boom is running into the same obstacle that has frustrated the field for decades: the hand. The latest signal comes from LinkerBot, a Chinese start-up that focuses on dexterous robotic hands rather than humanoids as a whole. Its founder, Zhou Yong, says making a robotic hand is “one hundred times more difficult” than making a humanoid, and the company is building around that idea instead of treating manipulation as a secondary feature.
The numbers behind the effort are now large enough to matter. LinkerBot says it makes about 5,000 hands a month and plans to double that figure. Zhou says the company is aiming for a $6bn valuation, while also targeting a price of just $1,000 per hand, down from the tens of thousands of dollars that mass-market prosthetic hands can cost today. Those figures make the hand look less like a research accessory and more like a standalone industrial product category.
That matters because the robotics industry has been stuck on the same truth for years: walking, balancing and waving are not the hard parts. The hard part is manipulating the physical world with enough precision, repeatability and sensitivity to be useful in ordinary work. The Guardian’s reporting from Beijing shows that this is exactly where the center of gravity is shifting. China is not just producing more humanoid robots. It is increasingly concentrating on the one subsystem that determines whether a humanoid can actually do work.
The broader backdrop is also telling. More than half of the factory robots installed each year are in China, giving the country a deep manufacturing base, a dense supplier network and a huge local market for robotics companies to test and scale products. But the same reporting also notes that the use cases for humanoids remain minimal, and the International Federation of Robotics concludes that “True multipurpose humanoids are far off yet.” That tension explains why hands, not whole robots, are becoming the focus.
It is an important distinction. A humanoid can draw attention by looking human-sized and human-shaped. But it cannot become economically useful unless it can pick up irregular objects, grip tools, adapt to weight shifts and handle fragile materials. That is the role of the hand. A robot with weak manipulation can still be a demo; a robot with capable hands starts to look like a machine that can earn its keep.
China’s advantage is that it can turn that insight into scale quickly. LinkerBot was founded in 2023, and in only a short span it has grown into one of the clearest examples of how fast the country’s robotics sector can move from concept to production. The company’s ambition is not just to sell hands to humanoid makers. It also wants mass-market prosthetic hands, suggesting that the same engineering stack could serve both industrial robotics and medical devices. That crossover is where the commercial logic becomes most interesting.
Still, the hard part is not simply making a hand that moves. It is making one that can do useful work consistently. Zhou’s own framing captures the engineering challenge: the hand’s dexterity is far greater than other body parts, but its physical volume is much smaller, which leaves less room for motors, sensors and control systems. That compression problem is the reason the hand sits at the intersection of hardware and software, and why it has become the field’s most stubborn bottleneck.
That bottleneck helps explain why so many humanoid efforts have seemed impressive but incomplete. Robots that can dance, gesture or perform tightly scripted motions still have not cleared the threshold into broad utility. The missing piece is not only intelligence. It is the ability to interact with the world through a hand that can feel, adapt and repeat tasks reliably. Until that happens, humanoids remain more convincing as showcases than as workers.
The Hand Is The Bottleneck
The central point in the current robotics cycle is that form factor is not the same thing as functionality. A humanoid body may satisfy the public imagination, but a robot becomes useful only when it can manipulate objects the way people do. That means the hand must handle precision, compliance, force and feedback at once. The engineering difficulty is amplified by the sheer concentration of complexity in such a small component.
Zhou’s comment that the hand is “one hundred times more difficult” than the humanoid captures the asymmetry. The torso and legs can be built to move in predictable patterns. The hand, by contrast, must adapt to the real world. It needs to grasp a plastic cup differently from a metal wrench, and it needs to do so without crushing either. That kind of dexterity depends on software that can interpret touch and position in real time, not just on mechanical design.
The source material also points to a deeper constraint: the hand is where hardware and software become inseparable. The mechanism must be small enough to fit into a practical device, but strong enough to work repeatedly, and sensitive enough to provide control signals. That means every improvement in one area can expose weaknesses in another. More force can create more wear. More sensors can create more complexity. More freedom of motion can create more control problems. The hand is a system of trade-offs.
That is why the hand has become a better barometer of progress than the humanoid shell itself. A robot that can stand upright and move around may still be far from useful if it cannot hold, twist, turn or place objects accurately. The hand is the part that turns mobility into labor.
This is also why the industry’s benchmarks are changing. The old test was whether a robot could look human-like enough to impress an audience. The new test is whether it can survive contact with the messy, varied and repetitive tasks that define real work. In that test, the hand is decisive.
“Human hands are the most important ability of human beings,” Zhou Yong said.
“If we focus on this one point, it is easier to realise many human skills,” Zhou Yong said.
Those remarks are more than branding. They amount to a strategy statement: focus narrowly on the hardest subsystem and let the rest of the product family follow.
Why China Is Racing Here
China is a particularly important place for this race because the country already sits at the center of industrial robotics deployment. More than half of the factory robots installed each year are in China, which means local companies can build against a large and active customer base. That matters for hands because the best way to improve manipulation is to put devices into environments where they are used constantly, fail quickly and can be iterated on.
It also matters because China’s robotics ecosystem is built for speed. A company can move from prototype to production more quickly when it sits near component makers, assembly capacity and large manufacturing customers. That does not guarantee success, but it does shorten the distance between a good idea and a marketable product. For dexterous hands, that speed is a real advantage.
Another reason China is central is that the market structure rewards component suppliers as well as robot makers. If a hand can be sold across multiple platforms, it can capture value even before humanoids become mainstream. That creates a path to revenue that is less dependent on a single robot brand winning the market.
The commercial appeal is obvious. LinkerBot says it can eventually bring the price of a hand down to $1,000. If that target is credible, the hand could move from premium robotics into much broader adoption, including prosthetics. The problem is that the same forces that drive cost down can also threaten durability and performance. Every buyer will want the same thing: a hand that is cheap enough to scale, but reliable enough to use.
That is where the real test begins. Robotics has long been full of prototypes that looked like products but never survived the hard economics of deployment. A hand that breaks, drifts out of calibration or loses sensitivity after repeated use is not a solution. It is a demo with a price tag.
So the Chinese race is not simply about building more machines. It is about proving that a highly complex part can be made in volume without losing the characteristics that make it valuable. That is a harder problem than adding more robots to factory floors, and it explains why the field has found itself back at the same fundamental question: what exactly makes a robot able to do human work?
What The Market Should Watch Next
The next phase of the story will not be decided by dramatic humanoid demonstrations. It will be decided by whether dexterous hands can cross the gap from promising hardware to dependable industrial tool. The key indicators will be production scale, price, durability and repeatability. LinkerBot’s current output of about 5,000 hands a month, and its plan to double that, are important because they suggest the company is already trying to industrialize rather than remain in prototype mode.
The company’s $6bn valuation target is also a useful marker of how investors may be thinking about the category. That number implies that dexterous hands are no longer being treated as a niche accessory. They are being valued as an enabling technology with multiple routes to market. Whether that valuation logic holds will depend on whether the product can actually deliver on the job.
For the broader robotics sector, the lesson is sharper. Humanoids will continue to attract attention, but the real breakthrough may come from a component that is easier to overlook. The hand is where the promise of embodied AI becomes physical, practical and measurable. It is where a robot stops being a spectacle and starts becoming a tool.
NextFin News - China’s bet is that the hardest part of the robot is also the part that can unlock the whole market. If that proves right, the next robotics winner may not be the best-looking humanoid. It may be the company that solved how to make the hand work.
Explore more exclusive insights at nextfin.ai.
