Dropping from 500,000 yuan to 10,000 yuan, why are dexterous robotic hands standing on the eve of mass production?

A price war in the robotics industry is spreading from complete machines to the "hands".

At the 2026 World Robot Conference, high-degree-of-freedom dexterous hands that used to sell for 400,000 to 500,000 yuan or even higher have seen their price range drop all the way to the 10,000-yuan level. Yuequan Bionics launched a 10,000-yuan product, Xinuo Future quoted 128,000 yuan for its high-DOF integrated arm-hand product, and many enterprises began to configure the degree of freedom and performance according to different scenarios such as industry, scientific research, and special operations.

This signifies a critical change: Dexterous hands are shifting from "being developed for demonstration" to "being made at different price points for different customers".

Capital is also placing bets in advance. In 2025, the total financing of China's domestic dexterous hand track reached about 168.77 billion yuan, while in the first half of 2026 alone, the financing amount exceeded 250 billion yuan, nearly 50% higher than the full-year figure of last year.

However, the "eve of mass production" does not mean that full-scale explosion will definitely arrive in 2027. The judgment that 2027 will become the core year of rapid volume growth is currently mainly put forward by industrial enterprises. What truly determines whether this "hand" can move from exhibition booths to factories and households is whether cost, tactile sensing, micro-actuators and AI brains can all cross the commercialization threshold at the same time.

Why are "hands" becoming more important as robots get closer to commercialization?

Over the past two years, the robotics industry has first solved the problem of "whether the robot can move".

After bipedal robots learned to walk, run and even somersault, the problem the industry really encountered has become more realistic: after entering the factory, what exactly can it do?

Moving boxes is relatively easy, while the truly difficult tasks include tightening screws, inserting and removing parts, sorting flexible objects, picking up eggs, and operating tools.

All these tasks cannot be done without hands.

A dexterous hand that can truly replace human operations not only requires five fingers to move, but also complete different actions such as grasping, pinching, clamping, rotating, pushing and pulling, and adjust the force in real time according to the hardness of the object at the same time.

This is why the industry often refers to dexterous hands as the "last centimeter" for robots to interact with the real world.

More importantly, dexterous hands do not rely on a specific robot form.

Bipedal humanoid robots need hands, wheeled dual-arm robots need hands, and AGVs and even fixed robotic arms can also be equipped with dexterous hands.

Therefore, even if there is still debate about whether future robots will be "two-legged" or "four-wheeled", the demand for end effectors is relatively certain.

The complete machine route may diverge, but a functional "hand" is almost a common demand.

This is also one of the reasons why dexterous hands are likely to form a large-scale industry earlier than general-purpose humanoid robots.

From 400,000-500,000 yuan to the 10,000-yuan level, more than the price has changed

The price drop is probably the most noteworthy signal for investors at this WRC.

In the past, high-DOF dexterous hands were more like laboratory equipment: low output, expensive components, requiring a lot of manual debugging, and customers mainly came from universities, scientific research institutions and robotics enterprises.

Even with the strongest performance, such products can hardly enter ordinary industrial scenarios.

Because factories calculate a different set of accounts.

If the total annual cost of a manual position is more than 100,000 yuan, while a pair of robotic hands costs hundreds of thousands of yuan, plus the cost of the robot body, maintenance, algorithm training and downtime risks, it is difficult for enterprises to get a reasonable return on investment.

After the emergence of 10,000-yuan products, the logic has begun to change.

Scientific research scenarios can purchase high-DOF and high-performance versions; general industrial grasping can choose products with relatively low degrees of freedom and more controllable costs; future household robots may require lighter, safer and cheaper products.

This means that dexterous hands have begun to see product stratification just like auto parts and consumer electronics.

Product stratification is often an important sign that an industry is evolving from scientific research equipment to industrial commodities.

However, after the price drops, another problem will become more important:

Can the performance remain stable after the price becomes low?

Real mass production does not mean making 100 hands, but that after making 10,000 hands, the force, accuracy, service life and control performance of each hand are basically consistent.

This is the core test of whether 2027 can really become a year of volume growth.

Why is a small "hand" so difficult to manufacture?

The most tricky part of dexterous hands is that they require a large number of components to be packed into a very small space.

Motors, lead screws, reducers, sensors, drivers and circuits all need to be integrated into a structure close to the size of a human hand.

This means that the components must be small enough, and output sufficient force at the same time; they must be precise, and durable as well.

Take the hollow cup motor as an example.

It determines the finger driving performance, but the space in the robot palm is limited, so the size of the motor must be continuously reduced, while ensuring the response speed, torque and service life.

Another example is the miniature lead screw.

Previous industry surveys show that high-precision planetary roller screws are still a key weak link in China's domestic robot supply chain. The smaller the size, the more difficult it is to control the accuracy, processing and batch consistency.

Therefore, the price drop of dexterous hands from hundreds of thousands of yuan to the 10,000-yuan level is essentially not a simple price reduction by enterprises.

A round of supply chain industrialization must take place behind it:

The originally laboratory-style small-batch parts need to be transformed into industrial products that can be produced automatically and standardizedly.

This is also why capital is gradually spreading from robot complete machines to upstream links such as hollow cup motors, miniature lead screws, integrated joints and sensors.

Why does capital heavily bet on "tactile sensing" first?

If motors and lead screws solve the problem of "whether the fingers can move", then sensors solve the problem of:

Whether the robot knows how much force it is applying.

When a human picks up an egg, he does not calculate how many Newtons of force are needed in advance, but adjusts the force automatically according to the tactile feedback of the fingers.

For robots to do the same thing, force sensors and tactile sensors are required.

At this year's WRC, tactile sensing has even become one of the core directions of dexterous hand competition.

The original report cited data from GGII, pointing out that Bluedot Touch's share in China's domestic six-axis force sensor market for humanoid robots reached 72.6%. The company has built an automated production line with a designed annual output of 1 million sets of joint force sensors and 200,000 sets of end six-axis force sensors; its disclosed delivery cycle has been shortened from 6-8 weeks for some overseas products to 2-3 weeks.

These figures are more noteworthy than a robot rock-paper-scissors performance.

Because only when core components start to plan production capacity at the level of "hundreds of thousands of sets and millions of sets", can the industry truly prepare according to the logic of large-scale manufacturing.

What capital is betting on is not actually "whether robots can shake hands".

It is that every robot that truly enters the production environment in the future may need a large number of perception and execution components.

But the biggest bottleneck is shifting from "hand" to "brain"

A more interesting change is taking place.

In the past, the industry was worried that the hardware could not meet the requirements.

Now some enterprises have been able to make dexterous hands perform quite rich anthropomorphic actions, but a new problem has emerged:

The robot does not know what action to perform at what time.

The relevant person in charge of Xinuo Future stated during WRC that from a hardware perspective, its products can already reproduce about 90% of the common actions of human hands, but for complex and high-precision tasks, the industry still relies heavily on teleoperation, teaching or trained strategy models. In other words, the capability of high-DOF hardware is getting stronger and stronger, but the model has not fully released these capabilities.

This is also the biggest difference between embodied intelligence and ordinary industrial automation.

Traditional robotic arms can be pre-programmed:

Move to point A, clamp the object, and move to point B.

But in household or complex factory environments, it is impossible to pre-program all actions.

Robots need to judge by themselves how to grasp an unknown object after seeing it, adjust the force by themselves when the object slides, and re-plan the action when the position changes.

This requires a large amount of real-world data.

Human operation data cannot be 100% directly migrated to robots. Different dexterous hands have different degrees of freedom, transmission modes and joint structures, and the same human action may be completely different when converted to a robot.

Therefore, the stronger the hardware, the higher the importance of data and algorithms.

This is why a quite contrasting situation has emerged in today's dexterous hand industry:

The hands are getting closer and closer to mass production, but the brains still need to be improved.

250 billion yuan of financing in half a year, what exactly is capital betting on?

Capital certainly will not enter the market after all problems are solved.

In the first half of 2026, the total financing amount of the dexterous hand track has exceeded 250 billion yuan, and the valuation of many enterprises has risen rapidly. After Xinuo Future completed the 500 million yuan A+ round of financing in July this year, its post-investment valuation exceeded 1 billion US dollars.

On the surface, investment institutions are betting on the sales volume of dexterous hands.

The deeper logic is actually:

Dexterous hands are likely to become standardized high-value components in the robotics industry.

It is somewhat similar to chips and auto parts.

There may not be only one robot enterprise in the final market, but as long as a large number of robots need "hands", the core actuators, sensors and control systems can cover multiple complete machine brands.

This is also why the upstream supply chain is getting more and more attention.

However, the 250 billion yuan of financing also means that competition will intensify rapidly.

At present, the degrees of freedom of dexterous hands range from single digits to dozens, the prices range from several thousand yuan, ten thousand yuan to hundreds of thousands of yuan, and routes such as rope drive, direct drive and connecting rod coexist. The industry has not yet formed a unified "standard hand".

What will truly determine the value of enterprises in the future will not only be the financing amount or the number of degrees of freedom.

It is the ability to achieve three things:

Being able to produce in batches, having continuous willingness from customers to purchase, and keeping the cost declining.

Whether 2027 will be the "first year of mass production" ultimately depends on orders

Therefore, for the industry judgment that "2027 will become the core year for the large-scale growth of dexterous hands", the market should pay attention to it, but also remain rational.

The positive side is already very clear.

Prices have begun to decline, the localization of core components is accelerating, million-level sensor production capacity is under construction, the demand of complete machine enterprises for dexterous operation is increasing, and capital is also entering rapidly.

These are all pre-signals of standard industrialization.

But the final test criterion is only one:

Real orders.

If dexterous hands are sold in large quantities to scientific research institutions and robot manufacturers for training, it is a completely different commercial stage from when they enter the automobile, 3C, warehousing and even household scenarios in large quantities.

The former proves that someone is willing to test the technology.

The latter proves that the economic account is feasible.

Therefore, this "war of hands" has entered a very critical transition period.

The first half of the competition focuses on:

Who has more degrees of freedom, and whose actions are more human-like.

The next stage of the competition will truly focus on:

Who can turn hundreds of thousands of yuan of laboratory products into industrial products priced at tens of thousands of yuan or even the 10,000-yuan level; who can stably produce tens of thousands or hundreds of thousands of sets; whose products break down less, last longer, and whose AI models can truly give full play to their capabilities.

The drop from 500,000 yuan to the 10,000-yuan level is only the first admission ticket.

The sign that dexterous hands are truly standing at the eve of mass production is not whether they can play rock-paper-scissors, nor whether they can pick up an egg.

It is that one day, when factory purchasers see it, they will no longer ask:

"What can this thing do?"

Instead, they will directly start calculating:

"How many of these do we need to buy for one production line?"

If this step really happens, 2027 is expected to become the key year for dexterous hands to transform from a technical hotspot to a large-scale industry.

References: Science and Technology Innovation Board Daily, Shanghai Securities News, GGII, public financing and production capacity data of enterprises

Risk Warning: This article is analyzed based on public industrial and financing data, and does not constitute investment advice. "Rapid growth in mass production in 2027" is the expectation of industrial practitioners, not a definite result. Dexterous hands still face multiple uncertainties such as ununified technical routes, mass production consistency of core components, cost, real scenario data and algorithm capabilities.

This article is from WeChat official account "BT Finance" (ID: btcjv1), Author: Shuyan, published with authorization from 36Kr.