Gasgoo Munich- Global Tier 1 giants are fast becoming a critical link in the embodied AI supply chain.
An analysis by Gasgoo reveals that since 2025, foreign Tier 1 suppliers—including Bosch, Schaeffler, Valeo, and Aptiv—have been racing to double down on embodied AI.
Yet unlike startups navigating the path from zero to one, these giants aren’t simply chasing a trend. Instead, they are making calculated moves grounded in a century of industrial heritage. Their goal: to expand their boundaries amid the wave of physical AI and redefine their ecosystem—evolving from core players in the auto supply chain into key architects of the embodied AI value chain.
Global Tier 1 Suppliers Make Intense Moves in Embodied AI
Participation from global Tier 1 suppliers in the embodied AI sector continues to climb.
Image Source: Qianxun Intelligence
Recently, Bosch China signed a strategic partnership agreement with Qianxun Intelligence. Over the next two years, leveraging Bosch’s factories and logistics centers in China, the two will collaborate on data collection, model training, industrial deployment, and core component supply. The aim: to close the technical loop from "real-world data to embodied models to real-world scenarios."
This marks Bosch’s latest move in the embodied AI space.
Prior to this, Bosch had already invested in companies such as Linghou Robotics, Yinshi Robotics, Farobot, and Yimu Technology through its investment arm, Bosch Capital. It also established a wholly-owned platform, "Bosch Yuan Qishi," dedicated to embodied AI, and formed a joint venture, Boyin Hechuang, with GalaxyBot. These initiatives provide end-to-end support from technical incubation to product deployment, accelerating the scaling of embodied AI in industrial settings.
Today, Bosch’s presence spans core areas including full-stack hardware, dexterous hands, collaborative robots, multimodal perception, data collection, and model training. Through Boyin Hechuang, it has already begun supplying embodied AI robots externally, forming an initial commercial closed loop.
This means Bosch currently wears multiple hats in the embodied AI sector: capital operator, core component supplier, scenario provider, and end-user. It has constructed a closed validation loop where demand originates from its own production lines, products are deployed internally first, and mature solutions are then rolled out to the wider market.
The greatest value of this model lies in solving the "cold start" problem—the toughest hurdle in commercializing new technologies.
Schaeffler’s goals are even more direct and ambitious: by 2035, it aims to deploy thousands of humanoid robots across its global factories and generate 10% of total group revenue from this sector.
To that end, Schaeffler established a dedicated embodied AI robotics company in China this February, focusing on the R&D and production of core components and key subsystems for humanoid robots. It aims to be a "shovel seller"—a critical infrastructure provider—in the embodied AI space.
Image Source: Schaeffler
Beyond internal development, Schaeffler has partnered with Leju Robotics and ROBO to jointly drive humanoid robots toward large-scale industrial application.
ZF LIFETEC and Omow, for their part, have chosen to partner deeply with Wujie Dynamics to jointly explore the scalable application and continuous evolution of embodied AI in advanced manufacturing scenarios.
Founded in 2025 by Zhang Yufeng, former president of Horizon Robotics’ smart vehicle division, Wujie Dynamics focuses on building a "general brain" and "manipulation intelligence" for robots. Despite being a newcomer, its first-generation robot platform has already achieved substantive breakthroughs in both industrial manufacturing and commercial services—likely the reason two major Tier 1 giants have simultaneously bet on this startup.
After all, partnering with a team that has both technology and practical execution capabilities is far more efficient than building from scratch. Moreover, these partnerships highlight the open attitude of foreign Tier 1 suppliers toward their transformation into embodied AI, as well as the urgency to pivot quickly.
By comparison, Aptiv and Valeo are currently more focused on adapting existing automotive solutions for the embodied AI sector.
Aptiv, for instance, has developed the PULSE™ fusion perception system for embodied AI. This solution is already deployed in collaborative robots and autonomous mobile robots, enabling safe and efficient human-machine collaboration in complex warehouse environments.
Valeo, meanwhile, reportedly aims for non-automotive revenue to account for roughly 10% of total group revenue in the future—a figure that includes its robotics business.
To achieve this, Valeo is adapting its automotive sensor suites—including ultrasonic sensors, cameras, millimeter-wave radars, and LiDAR—for the robotics industry. Additionally, it offers high-performance actuators, thermal management systems, and power management systems to robot makers.
This makes it clear: a core objective for Tier 1 giants entering embodied AI is to "sell shovels." While their paths may differ, the destination is the same. After all, no matter who ultimately wins the embodied AI race, they cannot bypass the need for reliable core components.
In terms of application scenarios, these Tier 1 giants have, without exception, locked their core focus on the industrial manufacturing sector—the arena they know best.
More specifically, they are all targeting the "tough nuts" that traditional industrial robots cannot crack.
Image Source: Boyin Hechuang
Boyin Hechuang, for example, focuses on key links that traditional automation struggles to cover—such as complex component handling, precision assembly, and flexible quality inspection. It develops flexible, high-precision solutions like dexterous manipulators and single-arm robots.
This also illustrates that in industrial settings, whether embodied AI truly looks like a human is less important than whether it can solve production pain points that traditional industrial robots could not address.
The partnership between ZF LIFETEC and Wujie Dynamics similarly targets technical bottlenecks in critical areas like complex component handling, precision assembly, and soft object manipulation—tasks where traditional automation falls short.
The reason is straightforward: traditional automation excels at standardized, repetitive, and structured tasks. However, it struggles when faced with flexible operations, complex decision-making, and non-standard part handling.
And that is precisely the irreplaceable core value of embodied AI in the industrial sector.
Threefold Crossover Logic: Technology, Scenario, and Data
The most direct logic behind global Tier 1 suppliers’ intense crossover into embodied AI is the commonality of technology.
"Intelligent driving and robotics share the same technological roots. Core technologies like perception, planning, and control can be directly repurposed—this is a massive opportunity," notes Gu Jianmin, CTO of Valeo China.
For Valeo, therefore, the convergence of automotive and robotics technologies is a natural evolution.
Currently, Valeo is leveraging its scale and technological advantages in the automotive sector—particularly its deep expertise in perception systems—to provide the "eyes" and "cerebellum" for the broader robotics market.
Image Source: Aptiv
Yang Xiaoming, President of Aptiv China and APAC, also believes that the underlying logic of perception in embodied AI is highly consistent with that of intelligent driving. Solutions from automakers can be directly applied to humanoid robots; Aptiv’s PULSE™ fusion perception system is a direct result of this philosophy.
Similarly, Schaeffler’s core components and key subsystems for humanoid robots—such as rolling bearings, planetary roller screws, precision reducers, motors, and sensors, along with rotary actuators, linear actuators, thermal management, and battery management modules—leverage decades of technical accumulation in the automotive and industrial sectors. This heritage has allowed it to rapidly establish a comprehensive product lineup and mass production capabilities.
Furthermore, in Yang’s view, the automotive industry holds mature advantages in industrial cost reduction and mass production iteration. Aptiv’s long-standing expertise in large-scale intelligent manufacturing processes—especially digital production—can bring numerous opportunities to improve quality and lower costs in the manufacturing of embodied robots.
However, technological commonality only addresses the question of "can it be done."
What truly gives global Tier 1 suppliers a unique edge in the embodied AI race is a more hidden "ace": their network of manufacturing plants worldwide. These factories serve as natural "training grounds" for the advancement of embodied robots.
"Embodied robots are AI-driven with high technical barriers, and AI iteration relies heavily on data support. In actual deployment, these robots continuously generate massive amounts of valid data, which is crucial for achieving breakthroughs in cost-performance ratio. Our digitally enabled, intelligent factories provide a comprehensive data application environment for the deployment of embodied robots. This is one of our key advantages," Yang explains.
The partnership between Bosch and Qianxun Intelligence, for example, utilizes factories and logistics centers to directly collect data, train models, and then redeploy them back into real-world scenarios for verification—ensuring high data quality and security.
Particularly in actual production, variables such as lighting changes on the line, equipment vibration, material tolerances, and cycle times are difficult to replicate in simulation. These variables are precisely the most scarce "nutrients" for training embodied AI models—an advantage many embodied AI startups lack.
A research report from Goldman Sachs even states bluntly that current robot pre-training relies heavily on simulation and synthetic data. Yet, an accuracy rate of 80% to 90% in a simulated environment often plummets to below 50% in real-world scenarios.
The importance of large-scale, high-quality real-world data is self-evident. Behind this data lie decades of tacit knowledge held by Tier 1 suppliers—precision craftsmanship, quality judgment, and failure modes. This forms a competitive barrier that other players in the value chain cannot easily surmount.
It is worth noting that beyond enabling the technical closed loop of "real data collection—model iteration—scenario verification," Tier 1 giants are accelerating the formation of a commercial closed loop by deploying embodied robots in their own factories first. This not only drives faster scaling of embodied AI but also creates a "second growth curve."
Image Source: Farobot
Ding Hao of Bosch Capital explicitly points out that their strategy involves leveraging their own industrial resources and real-world scenarios to empower partnerships and incubation. This increases their level of control and participation, improving the safety margin while pursuing excess returns.
This is especially urgent given the current macroeconomic slowdown, intensified competition in the automotive sector, and the pressure of the intelligent EV transition. Many foreign Tier 1 suppliers are facing growth headwinds in their traditional businesses and are in dire need of opening new frontiers to build fresh growth engines.
And right now, is there any sector more enticing—or better suited to absorb the industrial accumulation and resource advantages of global Tier 1 suppliers—than embodied AI?
China: The "Testing Ground + Accelerator" for Embodied AI Deployment
As global Tier 1 suppliers intensify their deployment of embodied AI in China, the country is becoming a source of innovation for the global industry.
Regarding Boyin Hechuang, for instance, Bosch China has clarified that it will conduct joint R&D and commercial exploration in the embodied AI robotics sector with Bosch Capital and GalaxyBot, gradually driving financing expansion and global layout.
As one of the first tangible results of this partnership, Boyin Hechuang has signed a strategic memorandum of understanding with United Automotive Electronic Systems Co., Ltd. (UAES) to jointly establish "RoboFab," a joint laboratory for embodied AI robotics.
Image Source: Boyin Hechuang
Recently, Boyin Hechuang secured a major order for 2,000 BW10/BW10-Lite embodied AI robots from Dijie Industrial, officially kickstarting its commercial closed loop.
Through its partnership with Wujie Dynamics, ZF LIFETEC has already launched specific scenario applications in its Chinese factories. Looking ahead, the two parties may deepen cooperation in areas such as scene reconstruction, test verification, and global quality standards by establishing joint laboratories, with plans to first establish benchmark projects in China.
On this foundation, the two companies aim to form globally replicable practical experience, thereby driving the establishment of standards and paradigms for the large-scale application of embodied AI in global manufacturing.
Additionally, Wujie Dynamics’ embodied AI robots have entered Omow Group’s Chinese factories, where they have begun trial operations and iterative evolution on assembly lines, efficiently supporting various dexterous manipulation tasks in flexible production.
All these deployment projects point to one reality: China is becoming the frontline for the deployment of embodied AI in industrial scenarios.
The logic behind this is straightforward: China possesses the world’s most complete robotics supply chain, the densest cluster of factories, and the most proactive policy support. This provides a testing ground and acceleration space for embodied AI—from R&D to scaled deployment—that other markets cannot match, making China the core engine driving the development of global industrial embodied AI.
This path mirrors the one foreign Tier 1 suppliers have taken in the Chinese automotive market over the past decade: evolving from "in China, for China" to "in China, for the world."
Yet clearly, in the realm of embodied AI, global Tier 1 suppliers have recognized the frontier value and innovative leadership of the Chinese market much earlier.
However, as robots begin to enter Tier 1 factories in batches to collect real-world scenario data, a thorny issue surfaces: who owns the data?
Data from automotive production scenarios is extremely valuable, containing decades of accumulated process know-how and quality control experience. Such data is inherently closed and exclusive.
Moreover, in the long run, if algorithms determine how fast a robot can go, data determines how far it can go. The ability to acquire large-scale, high-quality real-world data will not only determine the positioning of different participants in the industry, but also—depending on whose hands the data is in—determine who truly survives to the end.
This scene has already played out in the autonomous driving sector over the past few years. Back then, automakers and autonomous driving algorithm companies engaged in a prolonged tug-of-war over data ownership and model iteration rights. In the end, the winners were those capable of building their own data closed loops.
Today, a similar game may be unfolding in the embodied AI sector. Only this time, the scenario owners holding massive amounts of real-world data are the global Tier 1 suppliers.
Conclusion
The entry of robots into factories is not merely a matter of replacing workstations. Every day, they collect data, train models, and iteratively optimize. The knowledge accumulated in this process will eventually become a new class of assets that are the hardest for manufacturing enterprises to replicate.
In this sense, the intense moves by global Tier 1 suppliers in the embodied AI sector appear to be a technology race on the surface, but behind the scenes, it is a long-term strategic play for data sovereignty and industrial voice.
