Recently, the headquarters base of Bowintec was officially put into operation in the Suzhou Industrial Park. From signing to operation, it took less than seven months.
Bowintec was established with a 50% equity stake from each of Bosch's Boyuan Capital and GALBOT. The headquarters base covers a construction area of nearly 10,000 square meters, with an investment of CNY 1 billion, integrating R&D, manufacturing, durability testing, data collection, and model training.
Its planning is quite specific: production capacity will increase to 4,000 units by 2027, forming an annual production capacity at the ten-thousand-unit level in the coming years. Around the time of signing, it had already signed an agreement with Dijie Industrial, a logistics sorting system integrator, which plans to purchase nearly 2,000 BW10 series robots for express delivery sorting scenarios within three years. The first self-developed industrial robot has also completed prototype verification and small-batch delivery at a Bosch United Automotive Electronics factory.
Placing Bowintec back into the cluster of humanoid robot projects in Suzhou conveys a clear signal: the landing logic of the humanoid robot track is shifting from "where can prototypes be made" to "where can products be mass-produced and delivered." What Suzhou is undertaking is the latter part.
1. Who Is Involved in Suzhou's Humanoid Robot Projects?
1. Bowintec (Suzhou Industrial Park)
Already mentioned above, no repetition here. A structural observation worth supplementing is that this project involves "the scenario provider (Bosch) and the model provider (GALBOT) jointly building an industrialization base in Suzhou," rather than a model company setting up a standalone R&D point or a manufacturing enterprise expanding a single production line. Bosch provides real factory scenarios and manufacturing system experience, GALBOT outputs end-to-end embodied large models and simulation training pipelines, while Suzhou provides the industrial space and supply chain to connect the two.
2. Schaeffler Taicang Embodied Intelligent Robot Digital Factory
This factory, inaugurated by Schaeffler in Taicang on October 20, 2026, covers an area of approximately 7,500 square meters and produces humanoid robot joint modules, including actuators. Six humanoid robots are distributed across five workstations on the production line, completing material preparation and sorting, assembly, inspection, labeling, and packaging. Schaeffler states that this marks the starting point for its mass production of humanoid robot joint components.
As of the third quarter of 2026, Schaeffler's global backlog of orders for humanoid robots stood at approximately EUR 400 million, with two new Greater China customers added in the quarter. It should be noted that the EUR 400 million figure represents order value, not confirmed revenue; in 2025, the robot business accounted for less than 1% of Schaeffler Group's total sales. However, the significance of this project lies in the fact that a Tier 1 automotive supplier is bringing large-scale manufacturing capabilities and existing factory systems into the humanoid robot joint segment, with the competitive focus clearly directed toward unit cost and batch consistency.
3. Suzhou STT Phase III (Wujiang)
Suzhou STT is a national-level "Little Giant" specialized, refined, distinctive, and innovative enterprise, focusing primarily on ultra-miniature precision ball bearings with an outer diameter of less than 9mm. The Phase III project, which broke ground on October 23, 2026, has a total investment of CNY 2.5 billion. Upon completion, it will have an annual output of 1 billion sets of miniature precision bearings and 1 million sets of core components for humanoid robots, focusing on the industrialization of ultra-miniature cycloidal reducers and actuators. After reaching full capacity, it is expected to add CNY 2 billion in annual output value.
The project simultaneously unveiled the "Advanced Precision Metal Materials Innovation Center," aiming to bridge the "materials—components" chain starting from high-end bearing steel. In the most upstream precision transmission link of humanoid robots, the planned annual capacity of 1 million sets points to bulk supply, not prototyping.
4. Linkhou Robot R&D and Manufacturing Headquarters (Wujiang)
Linkhou possesses capabilities in vision modules, joint modules, domain controllers, and complete machine mass production. The Wujiang project has a total investment of CNY 1 billion, covering 57.4 mu, with a planned annual output of 850,000 vision modules, 450,000 core motion control components, and 80,000 embodied robots, expecting to add CNY 1 billion in annual sales. Its positioning is as a "core hub" for both core components and complete machine mass production. This dual capability of "components + complete machines" is harder to replace during the mass production stage than capacity in a single link.
5. FAIRINO (Suzhou High-Tech Zone)
FAIRINO started with collaborative robots, ranking first in both total sales volume and overseas shipments among Chinese collaborative robot brands in 2025. Global orders are expected to reach 18,000 units in 2026, a year-on-year increase of 68%. Its entry point into the humanoid robot track is not complete machines, but humanoid force-controlled arms. The ART series industrial-grade humanoid force-controlled arms launched in August 2026 have signed cooperation agreements for 2,000 units each with Bowintec and Jack Technology.
Agreement scale does not equal final delivery, but the signal is clear: the "arms" of humanoid robots are beginning to enter procurement processes through bulk-level agreements. FAIRINO's global R&D and operations headquarters base, with a total investment of CNY 400 million, broke ground in the Suzhou High-Tech Zone in February 2026.
6. Companies That Have Achieved Mass Production Numbers
Zhishen Technology rolled off its 10,000th industry-grade embodied intelligent robot in 2026, becoming the third domestic embodied intelligence company to join the "ten-thousand-unit mass production club." Its bionic quadruped robots can be produced every four minutes at the fastest rate. Luming Robotics uses Suzhou as its R&D and production base, with monthly sales of its bipedal robot series stabilizing around 200 units. These numbers may not be startling in the broader humanoid robot industry, but they indicate that a group of companies in Suzhou have already moved beyond the stage of "can we make it."
2. Suzhou's Foundation in the Humanoid Robot Track
The overall scale of Suzhou's robotics industry is indeed significant. In 2025, there were over 700 enterprises related to the city's robotics industrial chain, including more than 100 key enterprises in the field of embodied intelligent robots, with an industrial scale of CNY 159.55 billion. The gross output value of "Robotics + Artificial Intelligence" enterprises above designated size in Wuzhong District reached CNY 117.43 billion, gathering over 1,600 related enterprises, accounting for nearly 40% of the city's total robotics industry revenue. There are over 150 embodied intelligence-related enterprises in the Suzhou Industrial Park, with revenue of CNY 23.4 billion in 2025. As of the end of 2025, the Suzhou High-Tech Zone had over 100 embodied intelligence-related enterprises, with core output value exceeding CNY 10 billion.
In these figures, humanoid robots are a subset of embodied intelligence, but Suzhou's industrial density supports both forms equally: reducers, servo motors, controllers, precision transmissions, vision modules, joint modules—the core hardware required for humanoid robot mass production—is supplied with considerable density in Suzhou and its surroundings.
3. Why Humanoid Robots? Why Suzhou?
The landing of different links such as humanoid robot complete machines, joint modules, precision transmissions, and force-controlled arms in Suzhou points to very specific reasons.
First, the major cost component of humanoid robots lies in parts, and Suzhou is a dense zone for parts. Reducers are one of the core components with the highest cost proportion in humanoid robots. Leaderdrive holds the top market share for harmonic reducers in China and ranks second globally. In 2025, harmonic reducer sales reached 425,000 units, with a capacity of 640,000 units; after the launch of new generation projects, capacity will increase by another 1 million units. Inovance and Veichi Electric continue to layout in servo motors and controllers. Suzhou STT places the capacity for 1 million sets of core humanoid robot components directly in Wujiang. To reduce costs for humanoid robots, these components must first be supplied stably and cheaply within close proximity.
Second, the "10-Kilometer Manufacturing Circle" compresses engineering iteration time. There is a frequently cited saying in Wuzhong District: the servo motors, reducers, and controllers required for a robot body can basically be fully supplied within a ten-kilometer radius. The person in charge of Shuangzi Zhiqing stated more directly: almost all core components needed for R&D can find suppliers within a 10-kilometer radius, and the entire process from drawing to complete machine for the first product is completed within Wuzhong District. Moving humanoid robots from prototypes to mass production requires extensive design changes, sample trial production, and yield ramp-up. Shortening the supply chain radius by a notch compresses the iteration cycle by a segment. This capability cannot be built up quickly through policy documents alone.
Third, scenarios and data are the actual raw materials for training the "brains" of humanoid robots. Suzhou has 160,000 industrial enterprises and massive manufacturing scenarios. Bowintec relies directly on Bosch factories for industrial scenario model training and delivery verification; Schaeffler replicates selected scenarios from its global factories in the Taicang digital factory for humanoid robot skill training. The first phase of the Suzhou City Embodied Intelligent Robot Comprehensive Innovation Center covers ten major training scenarios including home, catering, and logistics, having attracted 18 enterprises and accumulated over 1.4 million high-quality data entries. The intelligence level of humanoid robots largely depends on the density of real-scenario data. What Suzhou provides are ready-made training grounds and initial customers, rather than demonstration lines built from scratch.
4. Implications for the Industry and Boundaries of the Suzhou Model
Bowintec went from signing to mass production delivery in less than a year, Suzhou STT invested CNY 2.5 billion in ultra-miniature cycloidal reducers and actuators, Schaeffler brought the automotive industry's large-scale manufacturing system into joint modules, and FAIRINO's humanoid force-controlled arms entered procurement with 2,000-unit agreements—these actions collectively point to the fact that downstream customer requirements for humanoid robots have moved beyond "can it move" to "can it move stably, cheaply, and in large volumes." Mass production capability is becoming one of the hardest bottlenecks in the first half of industrialization.
However, the Suzhou model has clear boundaries that should not be obscured by the excitement of dense landings. Suzhou's strengths are concentrated in reducers, servo motors, precision transmissions, joint modules, vision modules, complete machine mass production, and scenario verification, i.e., the hardware and manufacturing layers. In areas such as general-purpose embodied large models, cognitive decision-making algorithms, and high-end motion control software, Suzhou still lacks leading algorithm enterprises with national influence, and some high-end sensors and industrial software rely on external supply.
This does not affect Suzhou's status as a hub for scaled manufacturing and commercial landing of humanoid robots, but it reminds the industry: Suzhou solves the problem of "making," not "thinking." For humanoid robot startups, placing mass production and supply chain links in Suzhou while keeping frontier algorithm R&D in other cities is becoming a pragmatic division-of-labor combination.
Bowintec going from signing to operation in less than seven months is a microcosm of this division-of-labor logic. It shows that when humanoid robots enter the "submitting answers" stage, whether a city can provide sufficiently dense component supply, enough real scenarios, and fast enough engineering iteration rhythms determines the speed of project landing more than whether it can provide a beautiful laboratory.