It is no longer news that humanoid robots can dance. However, getting them to actually work in factories remains a challenging task.
On October 20, 2026, Schaeffler’s embodied intelligence robot Taicang Digital Intelligent Factory was officially commissioned. On a joint module assembly line, six humanoid robots were introduced into five workstations to participate in production tasks such as material handling and part sorting.
The components they are involved in producing are precisely the core parts required by the humanoid robots themselves.
Upstairs, industrial scenarios such as visual inspection, part assembly, and material replenishment have been moved into training areas, where robots learn and validate different operational skills.
Training upstairs, production downstairs.
This approximately 7,500-square-meter factory connects joint module manufacturing, robot training, data collection, and digital twin systems. The real production lines not only fulfill manufacturing tasks but also provide application validation scenarios for the robots.
An automotive parts giant with decades of manufacturing experience is attempting to connect component production, robot training, and actual industrial applications.
What it aims to capture is not just a business selling core robot components.
Can Automotive Industry Core Competencies Translate to Robotics?
Once inside the factory, the evaluation criteria for humanoid robots change.
Walking and grasping are merely basics; manufacturing clients care more about continuous operational stability, working precision, and cost. These issues happen to be engineering challenges that automotive parts companies have long faced.
Schaeffler did not enter this market from scratch. From bearings, ball screws, and precision reducers to motors, sensors, and drive systems, its accumulated technologies share many commonalities with robotic motion systems.
In April of this year, Schaeffler reached a strategic partnership with ROKAE Robotics, jointly developing integrated joint modules for humanoid robots.
However, developing products that meet performance requirements is only the first step toward scaled manufacturing.
David Kehr, Head of Global Humanoid Robot Business at Schaeffler, believes that products should consider how to achieve scaled manufacturing right from the design stage.
This approach of synchronizing product design with manufacturing processes reflects the long-standing experience of the automotive industry.
Harmonic drives serve as an example.
As key transmission components for humanoid robots, harmonic drives require high machining precision. Schaeffler attempted to apply precision forming technology, which has been used in the automotive sector for about ten years, to manufacture related parts, thereby reducing material consumption and processing steps.
According to David Kehr, precision forming can shorten the manufacturing time for certain processing stages from minutes to seconds. Validation data disclosed by Schaeffler shows that this process can reduce manufacturing costs by over 25% and material consumption by over 75%, with mass production planned to start in Germany first in 2027.
Underlying this is an industrial logic familiar to automotive parts companies: it is not enough to just design the product; one must also consider how to produce it stably at lower cost and higher efficiency.
Of course, the manufacturing advantages of the automotive industry require a certain scale to be fully realized. The humanoid robot market is still in its early stages, and its product structure, performance requirements, and production scale differ from those of automotive parts.
For Schaeffler, what is truly worth migrating is not just specific component technologies, but the entire set of industrial experiences ranging from product design and process development to quality control and scaled manufacturing.
From Selling Parts to Selling Industrialization Capabilities
Compared to manufacturing core components, enabling humanoid robots to truly enter production lines is another complex engineering challenge.
The Taicang Digital Intelligent Factory not only produces joint modules but also uses real production lines as validation fields for robot applications.
“We focus not only on whether the robot can move, but more importantly on whether it can meet takt time, quality, safety, and cost requirements, working stably over the long term.” explained Sun Haitian, General Manager of Schaeffler Embodied Intelligence Robot (Taicang) Co., Ltd.
To this end, the factory set up a training and data collection center on the second floor. Six types of industrial scenarios were selected from the actual needs of Schaeffler’s global manufacturing network. A digital twin system connects the virtual environment with the actual production lines, supporting robot training, process adjustment, and application validation.
A box folding validation conducted by Schaeffler in collaboration with Spirit AI serves as an example.
The robot not only needs to complete the folding but also adapt to boxes of different sizes and operating conditions. According to on-site presenters, to improve the model’s generalization capability, the training center rotates different operators and materials, followed by cleaning, calibration, and filtering of the collected data.
This also brings another potential business opportunity.
“We don’t just produce products; we also produce data. Data can also become an asset.” On-site presenters mentioned that in the future, this data could potentially be commercialized through data trading platforms.
In fact, Schaeffler has included data services in its embodied intelligence business layout. For enterprises with numerous real manufacturing scenarios, the production process can not only validate robots but also continuously accumulate industrial data valuable for training.
Consequently, the value of the Taicang Digital Intelligent Factory extends further beyond component manufacturing itself.
However, moving from single-scenario validation to scaled promotion, the embodied intelligence industry still faces many issues to resolve.
Feng Tianlin, General Manager of JD Industrial Electrical Automation Business Unit, pointed out that current embodied intelligence solutions lack standardization, with highly dispersed technical routes posing obstacles to scaled promotion. Chen Runze, Vice President of R&D at Sudo Technology, directed the issue towards the evaluation system. He believes that the industry needs effective benchmarks to avoid companies blindly pursuing individual technical indicators while neglecting system performance in real tasks.
These views also highlight the value of industrial engineering experience.
Schaeffler’s special equipment business has a history of over 70 years, accumulating experience in production equipment development, automation integration, and customized production solutions. Now, it is attempting to extend these capabilities to humanoid robot testing, system integration, and scenario deployment.
A key question among them is: If a robot application is successfully validated in Taicang, can it be used in another factory?
When asked by Gasgoo, Sun Haitian admitted that cross-scenario migration remains a difficult problem being explored by the industry. He hopes to gradually shorten the current deployment cycle, which may take months, to weeks, and eventually to one or two weeks.
This clarifies the industrialization capability Schaeffler aims to establish: enabling validated robot skills, production processes, and deployment experiences to be reused across different factories, reducing the cost of redevelopment and debugging for each project.
From providing robot components, to helping robots enter production lines, to accumulating reusable industrial data, Schaeffler is expanding its role in the embodied intelligence industry chain.
How Big Can the New Business of an Auto Parts Giant Get?
Notably, Schaeffler’s humanoid robot business has begun to secure market orders of a certain scale.
On the day the Digital Intelligent Factory was commissioned, Schaeffler disclosed that it received new orders from two Chinese customers in the third quarter of this year, bringing the cumulative order value for the group’s humanoid robot business to CNY 400 million.
In terms of global layout, Schaeffler has established partnerships with companies like Humanoid and Hexagon Robotics; in China, it collaborates with partners such as Leju Robotics, ROKAE, Spirit AI, and JD Industrial on component development, industrial applications, and supply chain coordination.
Behind these collaborations lies a relatively unique identity for Schaeffler: it not only supplies robot components but also procures and uses robots.
According to relevant cooperation plans, Schaeffler intends to deploy approximately 2,000 Humanoid robots in phases across its global manufacturing bases by 2032; the cooperation with Hexagon Robotics involves deploying at least1,000 AEON humanoid robots during the same period.
The globally distributed manufacturing network not only provides application scenarios for robots but also allows Schaeffler to understand actual needs from the perspective of end-users.
Jochen Schröder, Chief Operating Officer of Schaeffler Group, believes that robot products must not only meet technical requirements but also satisfy the economic requirements of projects.
This dual identity enables Schaeffler to feed problems discovered on the industrial floor back to its component and system development teams. It understands both the manufacturing issues suppliers need to solve and the cost and efficiency concerns directly faced by robot users.
Meanwhile, the commercial division of labor in the humanoid robot industry is still forming.
Sun Haitian noted that procurement models vary among different finished machine enterprises. Some choose to procure components for self-integration, while others prefer more complete joint modules.
This diversity offers Schaeffler market opportunities at different levels. From core components and joint modules to system integration, industrial deployment, and data services, its product and engineering capabilities are seeking new application spaces.
The Chinese market serves as an important starting point for this business.
From joint module co-development with ROKAE, to industrial task validation with Spirit AI, to the operation of the Taicang Digital Intelligent Factory, Schaeffler is connecting local robot technologies, component manufacturing, and real industrial demands.
Leveraging its global manufacturing network, the product development, application validation, and deployment experiences accumulated in Taicang also have the opportunity to serve other markets.
This makes the Taicang Digital Intelligent Factory not just a component production base for the Chinese market, but also a significant carrier for Schaeffler to explore robot industrialization applications and connect Chinese innovation with the global manufacturing system.
Returning to the Taicang Digital Intelligent Factory.
Downstairs, humanoid robots participate in manufacturing joint module components; upstairs, they continue learning new industrial skills and accumulating real-scenario data.
The automotive industry already possesses a mature manufacturing system, while the humanoid robot industry is still exploring how to achieve scaled production and application.
What Schaeffler is attempting in Taicang is to transform the manufacturing technologies, industrial scenarios, and engineering experiences accumulated by the former into products and services usable by the latter.
This might be more noteworthy than the act of robots participating in manufacturing robot components itself.