Decoding 6D Force Sensor: Breakthrough and Changes in the Wave of Humanoid Robots | JD Insights
2025-11-06
(Photo generated by Nano-Banana)
Abstract: The way to reduce the cost of humanoid robot "nerve endings".
Research team | JD Capital Manufacturing Investment Department
Managing Director，Xing Xiaohui xingxh@jdcapital.com
Senior Investment Manager，Xia Han xiahan@jdcapital.com
……
Imagine a scene: on a phone assembly line, a robotic arm is trying to insert a connector smaller than a fingernail into a slot on the motherboard.
In the past, this was almost an impossible task. Traditional industrial robots were like blindfolded, order-following giants whose every action depended on preset trajectory planning. Whenever there was a 0.05mm deviation in the target position—whether from a slight vibration of the conveyor belt or the inherent tolerance of the components—they would unhesitatingly apply brute force, often resulting in connector damage and abrupt production line shutdowns.
But today, that is changing. The new robotic arm will pause slightly as it approaches the slot, then gently "sway" and "test" like a human hand, and smoothly push the connector into place at the right Angle.
Behind this evolution from brute force to clever force in just a few seconds is a core component called the "six-dimensional force sensor". It acts like a nerve terminal implanted in the robot's wrist joint, giving the cold machine for the first time a human-like "feel" (force sense).
This "nerve endings" can simultaneously detect three directional forces (Fx in front/back, Fy left/right, Fz up/down) and three torques (rotational torque Mx/My/Mz, such as the sensation of "twisting" or "pinching"). It comprehensively decodes both the "amount of force applied" and the "source of force", serving not only as the foundation for precise operations but also as the "soul" enabling robots to maintain balance during movement and achieve safe human-robot interaction.
Yet this compact "soul" has long been prohibitively expensive. Overseas tech giants like ATI, leveraging decades of technological leadership, have kept the sensor's price tag in the tens of thousands of RMB. For visionaries like Elon Musk, who dreams of reducing humanoid robot costs to $20,000, the four six-axis force sensors alone for wrists and ankles could cost nearly $10,000 – a figure that seems as far-fetched as a fairy tale.
But now, the rules of this game are being completely rewritten by a group of China's "game-changers". Under the gravitational pull of the massive incremental market for humanoid robots, a "price revolution" centered on cost has already begun. Domestic six-axis force sensor manufacturers are wielding a combination of supply chain advantages, rapid iteration, and scale expectations to drag the price of six-axis force sensors from the "ten-thousand-yuan level" down to the "thousand-yuan level".
This is not a traditional domestic substitution, but more like an industrial singularity triggered by cost. Why is this small sensor so critical? How will this "price butcher" revolution from 10,000 yuan to 1,000 yuan level leverage the entire humanoid robot industry?
I. Giants without a sense of force
To understand the value of a six-dimensional force sensor, consider how "clumsy" a robot would be without it.
From a control theory perspective, a robot without force feedback operates in an "open-loop" system. It functions like a machine that merely executes commands without sensing capabilities – the controller issues instructions and the robotic arm moves accordingly. The system cannot detect whether it has encountered obstacles or gauge the force exerted during object grasping.
The absence of this closed-loop feedback mechanism proves fatal in precision operation scenarios. As previously mentioned, connector insertion and removal require a 0.05mm clearance, demanding robots to achieve exceptional positioning accuracy. This absolute precision dependency makes the system highly vulnerable. Any minor wear and tear accumulated during prolonged operation of robotic arms could lead to reduced positioning accuracy, potentially rendering the entire program inoperable.
In contrast, human hand operation does not rely on high-precision positioning, but uses vision to roughly locate the connector and then insert it by "feel". This is the core value —— of six-dimensional force sensors to provide robots with "feel".
With this technology, robots can emulate human "hand-eye coordination": The eyes (vision sensors) guide the hand to approximate positions, after which the hand (end effector + six-axis force sensor) takes over. By sensing subtle changes in contact forces and torques, it dynamically adjusts its posture to achieve "feel-and-position" and smooth assembly. This approach significantly reduces the precision requirements for the robot's body, enhancing operational stability and robustness.
Alongside the wrist's precision controls, the six-axis force sensor mounted on the ankle plays a vital role. This device continuously monitors the ground's reaction forces acting on the robot's sole, forming the foundation for dynamic balance and stable navigation through complex terrains. While it might allow the robot to perform smooth dance routines on flat surfaces, encountering uneven dirt roads could instantly throw it off balance, resulting in a tumble that would make anyone lose their footing.
The six-axis force sensor is arguably the pivotal component enabling robots to transition from "automated machines" to "intelligent lifeforms". This innovation grants robots the ability to perceive their physical environment and interact with it in a flexible manner. However, developing this "sixth sense" for robots currently requires an enormous investment of resources.
II. The Five Passes and the Invisible Throne
One of the reasons why six-dimensional force sensors are expensive is the complex "soft and hard" technical barriers behind them. To build a high-precision, highly stable sensor, you need to overcome five core challenges.
The first stage involves elastomer design, the mechanical core of sensors typically constructed from high-strength steel, aluminum alloys, or titanium. When subjected to external forces, these components undergo micron-scale deformation. The key challenge lies in structuring the elastomer (e.g., three-rod, four-rod, six-rod, cross-rod, or vertical rod configurations) to enable unique measurable deformation patterns under multidimensional forces. This structural decoupling process represents the foundational step in engineering design and forms the core of patent strategies across manufacturers.
The second critical component is the decoupling algorithm. When force is applied in one direction, structural coupling causes subtle signals to propagate across other axes – a phenomenon known as "crosstalk". This crosstalk serves as the key performance metric for sensors. Top-tier products typically suppress it below 1%, while average models show 2-5% interference. The decoupling algorithm employs sophisticated mathematical models to precisely isolate and reconstruct mixed electrical signals into six independent measurements of force and torque. This technological breakthrough demands both profound theoretical expertise and decades of practical refinement.
The third critical stage involves processing techniques. This includes heat treatment and stress relief processes for the elastomer to ensure long-term operational stability, along with precise bonding of strain gauges. As sensitive components that convert mechanical deformation into electrical resistance changes, these hair-thin elements require meticulous positioning on specific elastomer surfaces and effective curing. Mastering this "know-how" is crucial for achieving optimal sensor accuracy.
The fourth stage is embedded circuit design. With the increasing demand for miniaturization and lightweight of humanoid robots, how to highly integrate the circuit board of data acquisition and signal processing into the sensor is a huge challenge.
The final critical factor determining processing efficiency and product precision is six-axis synchronous calibration technology. Every sensor undergoes rigorous factory calibration to establish precise correlations between deformation measurements and force values. While traditional methods require manual or semi-automatic calibration across all six axes – a time-consuming process that typically takes half a day per axis – industry leaders have developed proprietary automated systems capable of simultaneously applying precise loads across six dimensions with synchronized calibration. These cutting-edge devices, currently unavailable on the market, represent the secret weapon in their competitive portfolios.
Beyond these five strategic gateways, what proved more critical was the niche nature of six-axis force sensors before humanoid robotics took off. These sensors were primarily used in cost-sensitive sectors like aerospace, defense, automotive crash testing, and high-end R&D. This market remained lukewarm, leaving industry players like ATI with little incentive to slash prices and newcomers hesitant to enter the field. The result was a long-term stalemate of low-volume, high-priced products that stuck around.
Ⅲ.The combination of "price butchers"
The wave of humanoid robots, like a huge stone thrown into a calm lake, has completely broken the original market pattern.
When Musk announced his $20,000 target price, all component suppliers heard the ticking of the countdown. For OEMs, performance, cost efficiency, and mass production capabilities are the three core metrics they evaluate suppliers on. In this emerging market where cost sensitivity is paramount, sensor prices in the ten-thousand-yuan range are clearly unsustainable.
This provides a perfect window for domestic manufacturers to break the game. They have launched a set of aggressive "combination punches", aiming directly at the core pain point of price.
First, the supply chain's dimensional reduction strike.
The greatest advantage of domestic manufacturers lies in their reliance on China's mature electronics manufacturing and precision processing industry chain. Whether it is the machining of elastomers or the procurement of components for circuit boards, localized supply chains can provide highly competitive costs and response speeds. Elastomers and strain gauges constitute the majority of material costs. With the anticipated large-scale procurement driven by humanoid robots in the future, there will be significant room for price negotiation. Jiuding Investment predicts that when order volumes increase from dozens to thousands, procurement costs for elastomers alone could decrease by 30%-40%.
Second, the efficiency revolution of automation.
Jiuding Investment's industry research reveals that core manufacturing processes still predominantly rely on manual operations, suffering from low efficiency and high costs. Critical stages including component placement, wiring, and calibration all present substantial opportunities for automation upgrades. Domestic manufacturers are actively developing proprietary automated equipment, particularly calibration systems that can compress time-consuming calibration procedures from hours to mere minutes. This technological advancement not only dramatically reduces labor costs but also paves the way for future large-scale production.
Third, "close combat" agile service.
Currently, the design of downstream robot bodies remains fluid and undergoing rapid iteration. This demands upstream suppliers to possess swift responsiveness and customized development capabilities – precisely the "trump card" of domestic manufacturers. According to industry research by Jiuding Investment, six-axis force manufacturers in China can complete customization within a month, while foreign counterparts typically require about three months and show limited willingness to handle small-batch orders. In this race against time, such differences in response speed prove decisive.
Fourth, pragmatic trade-offs in technology.
Unlike typical domestic alternatives, Chinese manufacturers in the six-axis force sensor sector have not blindly pursued ATI's highest standards across all metrics. The current application scenarios for humanoid robots require less stringent overall precision than military research demands—achieving 0.5% or even 1%FS accuracy is sufficient. Leading domestic companies already demonstrate 0.1% precision, which far exceeds performance requirements. Consequently, they focus on balancing performance and cost through innovative approaches like simplified structural designs and optimized circuit integration to reduce expenses.
At the same time, Jiuding Investment also found in the research that the top domestic manufacturers are exploring new technologies, such as replacing the traditional metal strain gauge with a "silicon-based strain gauge + glass micro-firing process" that is cheaper and easier to automate, so as to reserve ammunition for future cost wars.
Through this strategic combination of measures, China's six-axis force sensors have successfully achieved a "price revolution". While foreign competitors still cling to the 10,000-yuan price range, domestic mainstream products have slashed prices to 3,000-4,000 yuan – about half or even lower than international counterparts. This breakthrough unequivocally signals to the industry: The era of premium pricing for six-axis force sensors is now a thing of the past.
Price is also the first domino to knock down industrial change.
Let's break down the economics: A humanoid robot typically requires 2-4 six-axis force sensors installed on its wrists and ankles. If using imported solutions priced at 15,000 yuan per unit, this component alone would cost 30,000-60,000 yuan. By switching to domestic alternatives priced at 4,000 yuan each, costs drop dramatically to 8,000-16,000 yuan – slashing expenses by over 50%.
JD Capital believes that the reduction in costs will enable humanoid robots to achieve large-scale commercial deployment in scenarios requiring frequent robotic grasping and assembly operations, such as automotive manufacturing and logistics sorting. Therefore, once the critical cost bottleneck of "force sensing" is overcome, downstream applications will have the opportunity to open up accordingly.
In the long run, cost reduction will serve as the catalyst for humanoid robots transitioning from B2B applications to B2C (home service) models. A household robot capable of serving tea, water, and folding clothes would remain an exclusive luxury if priced at hundreds of thousands of yuan. Only when costs become sufficiently low can such devices realistically enter ordinary households, much like how cars and smartphones have become commonplace in modern life.
IV. RISKS AND OPPORTUNITIES IN THE MIST
Despite the bright prospects, the road ahead is still foggy and risky for players in the game. In the view of Jiuding Investment, the biggest risk is not the disruption of the technology route, nor the underdevelopment of humanoid robots, but the drastic changes in the competitive landscape.
The application of six-axis force sensors in humanoid robotics is currently in its early stages, with most manufacturers being relatively small-scale. According to research data from Jiuding Investment, the global market size for six-axis force sensors in traditional industrial applications is projected to reach approximately 1 billion yuan by 2024. Looking ahead, as costs are significantly reduced, penetration rates of these sensors are expected to rise further in collaborative robotic arms and polishing scenarios. However, given the ongoing price reductions, the market size for six-axis force sensors in traditional industrial applications is anticipated to see only modest growth.
The real growth driver for the industry lies in the demand for humanoid robots. In emerging scenarios, the market size of six-axis force sensors is directly tied to both the shipment volume of humanoid robots and their penetration rate. Currently, the market base remains relatively small. According to Jiuding Investment's estimates, the annual market size for humanoid robot six-axis force sensors is projected to be less than 100 million yuan in 2025. However, this is expected to grow into a multi-billion-yuan market in the long term, with the sector maintaining rapid growth over the next decade.
However, as the market grows from tens of millions to billions, it also means that today's leaders may be overtaken tomorrow, and the process is full of uncertainty.
So where should the compass be pointing for startups in this race right now?
JD Capital emphasizes that strategic priority should focus on cultivating leading clients and establishing market dominance. For domestic six-axis force sensor manufacturers, the immediate objective is to secure partnerships with 1-2 top-tier robot manufacturers through collaborative R&D. In the early competitive phase, companies that first establish supply chain connections with these industry leaders will gain a decisive advantage in future market expansion. This stage exhibits a pronounced winner-takes-all dynamic where early entrants dominate the field.
Secondly, regarding product development priorities, companies should prioritize cost efficiency while ensuring performance adequacy. Given that most downstream clients currently lack proficiency in using high-precision force sensing data – with algorithms still under iterative refinement – deploying redundant ultra-high precision sensors proves largely unnecessary. Therefore, the strategic focus should lean toward cost optimization to drive market penetration. However, R&D of high-performance products must not be neglected, as this constitutes essential groundwork for industry advancement and future demand upgrades.
Finally, regarding financing strategies, Jiuding Investment emphasizes that companies should pursue maximum fundraising opportunities to build strategic reserves. The robotics industry's boom carries cyclical risks. Capitalizing on market windows to accumulate substantial funds – the golden rule of "stockpiling resources" – is essential for surviving the competition. History shows countless technically superior companies ultimately collapsed due to cash flow crises.
Today, six-axis force sensors – these compact precision components – are emerging as pivotal drivers in the humanoid robotics industry. For domestic manufacturers, this isn't merely a competition for market substitution, but a comprehensive challenge across the entire ecosystem: from design and algorithms to precision machining, automated equipment, and agile market-oriented services. The key lies in securing early-stage victories – the first to achieve such milestones will dominate the discourse of this new era.