On September 15, at a Gasgoo seminar, Gasgoo partner Wang Xianbin pointed out that the commercialization of humanoid robots is still in its early stages, with global robot shipments this year at about 60,000 units and China at about 50,000 units. In the complete machine BOM, the actuation system accounts for nearly half, and domestic production of core components is expected to drive prices down by 40% to 70%. The overlap between the automotive and robotics supply chains reaches 60% to 70%, providing a path for component companies to reduce costs through migration and reuse their overseas expansion capabilities, but small-scale pilots will not begin until next year, with the short term focusing more on capability building and validation.

Supply Chain Overlap of 60%-70%: The Cost-Reduction Logic and Going Global Extension of Auto Parts Companies Migrating to Robotics

On September 15, at the Embodied Perception Fusion and Multimodal Large Model Innovation Seminar hosted by Gasgoo, Wang Xianbin, Gasgoo Partner and Vice President of the Research Institute, made judgments on the industrialization pace, cost structure, and supply chain migration path of embodied intelligence and humanoid robots. His core view: although industry enthusiasm is high, commercialization is still in its early stages; what truly determines the industry's progress is cost reduction, data closed loops, and industrial chain collaboration, and the migration of the automotive supply chain will become an important driving force.

Industry Status: A Time Gap Between Enthusiasm and Commercialization Pace

Wang Xianbin first started with industry scale. This year's WIC exhibition area is about 100,000 square meters, and WRC about 60,000 square meters, with over 3,000 products and hundreds of companies participating. But in terms of shipments, it is estimated that this year's global robot volume will be about 60,000 units, and China about 50,000 units—a clear contrast between this order of magnitude and the exhibition enthusiasm.

Scene distribution also reflects early-stage characteristics. Entertainment performances and scientific research still account for a relatively high proportion. Although scenarios such as 3C, automotive, and logistics have made notable progress, they account for about 15% to 20%. Different scenarios have different product requirements: industrial and logistics scenarios require safety, reliability, long-term repetitive work, and precise operation and generalization capabilities; home scenarios, due to diverse people and objects, have higher requirements for vision, force perception, sensing, and large models; special scenarios face reliability challenges in dark and complex geographical environments.

Based on this, Wang Xianbin judges that humanoid robots are still in a small-scale pilot stage, will only enter small-scale pilots next year, and in the future will gradually expand from relatively structured ToB industrial scenarios to unstructured scenarios.

BOM Structure and Cost Reduction: Execution System Accounts for Nearly Half, Localization Drives Prices Down

Cost reduction is the first main thread in Wang Xianbin's discussion. He pointed out that in the BOM cost of the complete machine, the execution system accounts for nearly half; the perception part includes six-axis force sensors, IMU, 3D vision, and other components.

The main driver of price decline comes from localization. Wang Xianbin said that as scale expands, the localization rate of core components (reducers, motors, sensors, perception components, etc.) will rise rapidly, and prices are expected to drop by 40%-70%. The impact of this magnitude on the complete machine cost curve is a key variable for judging when robots will be ready for large-scale deployment.

It is worth noting that cost reduction is not an improvement in a single link. Hardware precision, cost control, and integration with large models were listed by Wang Xianbin as key areas that still need breakthroughs. In other words, price reduction is only a precondition and does not automatically equate to opening up scenarios.

Supply Chain Overlap of 60%-70%: A Feasible Path for Automotive Companies to Migrate

The most industry-significant data in this speech is that the overlap between automotive and robotics supply chains reaches 60% to 70%. Wang Xianbin believes that migrating automakers' cost reduction and technology pilots to robotics is a feasible path.

This overlap means that the process capabilities, mass-production experience, and quality systems accumulated by auto parts companies in links such as reducers, motors, sensors, and perception components have a foundation for reuse in the robotics field. For parts suppliers that have experienced multiple rounds of auto price wars and are accustomed to the annual price reduction rhythm, robotics is a new incremental market that has not yet been fully priced, but its entry threshold is not at the conceptual level; rather, it lies in whether the cost structure already verified in the automotive industry can be transferred over.

From the perspective of industrial collaboration, this is consistent with Wang Xianbin's emphasis on "industrial chain collaboration": cost reduction in robotics will not be achieved by a single company, but must rely on the scale capabilities of the existing supply chain system.

Technical Variables: The Evolution of Large Models Reshapes Industrial Division of Labor

At the technical level, Wang Xianbin gave a set of directional judgments. He believes that the rapid evolution of general large models will bring pressure to embodied intelligence "brain" companies, but will benefit companies doing "cerebellum," body, and visual-tactile perception algorithms.

Domestic general models have begun to land vertically. Against this background, those that can succeed in the future will mainly be leading tech companies, some automakers, and body manufacturers with resources and talent; however, companies that deeply cultivate vertical fields still have opportunities. This division-of-labor judgment echoes the aforementioned supply chain logic—the body, execution, and perception links are closer to manufacturing capabilities, and this is exactly the existing advantage range of the automotive industry chain.

Going Global Perspective: Path Reuse Space Brought by Supply Chain Migration

For practitioners concerned about the overseas expansion of China's new energy vehicles, the above judgments provide an indirect but still clear clue.

Chinese auto parts companies have already formed certain manufacturing and delivery networks overseas over the past few years along with the layout of complete vehicle capacity. If the overlap between automotive and robotics supply chains indeed reaches 60% to 70%, then this part of the capability theoretically has room to extend to the robotics business. Its significance lies in that going global does not require building production lines and customer relationships from scratch, but may reuse the existing manufacturing foundation and overseas service system.

But caution is needed. Wang Xianbin's judgment is equally clear: robot commercialization is still in its early stages and will only enter small-scale pilots next year, with scenarios also concentrated in relatively structured ToB industrial fields. At this stage, supply chain migration is more about capability reserves and pilot verification, rather than an already-formed overseas increment. For auto parts companies, observable indicators include whether the price decline brought by the localization of core components can continue, whether the pilot scale of industrial scenarios can expand, and whether body and perception companies can form stable orders in vertical fields.

In this sense, the relationship between automotive and robotics is, in the short term, closer to capability reuse rather than business replacement. What truly determines the speed of migration are still the three words Wang Xianbin repeatedly emphasized: cost reduction, data closed loops, and industrial chain collaboration.