U.S. Ban on Chinese Humanoid Robots: A New Front in the AI Race

Event Overview

According to reports, the Trump administration has recently introduced a new policy banning new Chinese humanoid robots from entering the U.S. market, citing the protection of America's domestic AI infrastructure buildout. The term "AI buildout" refers to the systematic infrastructure investment surrounding the artificial intelligence industry, encompassing data center construction, computing chip deployment, AI model training and inference platform development, and the hardware ecosystem required for AI application deployment. Between 2024 and 2025, U.S. tech giants have announced over $300 billion in AI infrastructure investment plans, with robot hardware seen as the "last mile" of this grand investment—extending AI intelligence from the cloud into the physical world. This move is widely regarded as a significant signal that U.S.-China tech competition is expanding from chips and AI models into Physical AI and robotics hardware.

Physical AI refers to the ability of AI systems to not only process digital information but also perceive the physical world through sensors and act upon the physical environment through actuators. Unlike traditional software AI (such as ChatGPT processing text or Midjourney generating images), Physical AI requires AI models to understand the complexities of the real world, including three-dimensional space, physical laws, and mechanical interactions. NVIDIA CEO Jensen Huang has repeatedly called Physical AI the "next frontier" of AI, arguing that enabling AI to understand and manipulate the physical world is far more challenging than understanding language and images. Humanoid robots are the most representative embodiment of Physical AI.

Humanoid robots are widely regarded as the next trillion-dollar market, integrating advanced AI brains, precision mechanics, sensors, and actuators among other key technologies. Multiple research institutions predict that the global humanoid robot market will reach the trillion-dollar scale between 2030 and 2035. Goldman Sachs' 2024 report raised its humanoid robot market estimate to $38 billion (by 2035), with an optimistic scenario reaching $154 billion. Elon Musk has made even more aggressive predictions, suggesting that if Optimus achieves mass production, Tesla's robot business alone could be worth several trillion dollars. The explosive growth logic behind this market lies in: increasingly severe global labor shortages, rapidly advancing AI capabilities making general-purpose robots feasible, and the humanoid form factor being naturally compatible with existing human work environments and tools. The issuance of this ban marks the formal inclusion of this emerging field within the strategic framework of national security and industrial competition.

Why Target Humanoid Robots

Data Security and National Security Considerations

Humanoid robots are not simple mechanical devices. They are equipped with numerous cameras, microphones, radar systems (including LiDAR and millimeter-wave radar), and other sensors capable of continuously collecting sensitive data about surrounding environments, personnel activities, and even factory layouts. Take a typical humanoid robot as an example: its vision system can capture dozens of frames of high-resolution three-dimensional spatial data per second, its microphone array can precisely record environmental sounds, and its torque sensors can detect the physical properties of objects it interacts with—the combined value of this data far exceeds that of ordinary IoT devices. Once deployed at scale in industrial, logistics, or even household settings, these devices could theoretically serve as potential channels for data exfiltration.

The U.S. government's concerns follow the same logic as previous actions against TikTok and Huawei equipment—that intelligent hardware controlled by foreign entities could pose intelligence collection and supply chain dependency risks. Humanoid robots integrate AI reasoning capabilities with physical action capabilities, making their potential security sensitivity even higher than pure software products. Unlike phones or routers that can only passively collect data, humanoid robots with autonomous mobility and manipulation capabilities can actively explore environments and acquire more comprehensive and in-depth information, placing them in a higher risk category in security assessments.

Protecting the Domestic AI Robotics Industry Chain

Another core motivation behind the ban is "protecting America's AI buildout." In recent years, the U.S. has invested heavily in AI large models and computing chips, and the robot body (embodiment) is becoming a crucial vehicle for AI deployment. Embodiment is an important concept in AI, referring to the form in which an AI system acquires a physical body and can perceive and act in the real world—this is considered one of the key pathways toward truly general artificial intelligence (AGI).

Tesla's Optimus has undergone multiple iterations since its debut in 2022, with plans to deploy thousands of units in Tesla factories for actual production tasks in 2025. Figure AI, founded by former Boston Dynamics engineers, has secured over $2.6 billion in funding from investors including Microsoft, OpenAI, and NVIDIA, and its Figure 02 robot is already being piloted at BMW factories. Agility Robotics' Digit robot has partnered with Amazon for testing in warehouse scenarios. Additionally, Boston Dynamics (under Hyundai), Apptronik, 1X Technologies, and other companies are actively advancing commercialization. These American companies are intensively competing in this space.

If lower-priced Chinese humanoid robots with stronger mass production capabilities flood the market, they could seize market share before the U.S. industrial chain matures, undermining the growth potential of domestic enterprises. Therefore, this ban is essentially an industrial protection measure, buying American robotics companies a development window. This follows the same logic as historical U.S. trade protections against Japanese automobiles and Korean semiconductors—using policy tools to create protective market space before domestic industries establish competitiveness.

China's Industrial Advantages in Humanoid Robots

You may not have noticed, but this ban actually reflects the fact that China has already developed considerable competitiveness in the humanoid robot field. In recent years, Chinese companies such as Unitree, UBTECH, and Fourier Intelligence have launched multiple humanoid and quadruped robot products with outstanding cost-performance ratios, with some products priced at tens of thousands of yuan—far below comparable European and American products.

Specifically, Unitree was founded in 2016, starting with quadruped robots. Its G1 humanoid robot is priced at approximately 99,000 RMB (about $16,000), making it one of the lowest-priced products of its kind globally, with capabilities including walking, carrying objects, and performing backflips. UBTECH went public on the Hong Kong Stock Exchange in 2023, becoming the world's "first publicly listed humanoid robot company," with its Walker series already undergoing actual deployment testing in new energy vehicle factories. Fourier Intelligence's GR series targets rehabilitation medicine and general-purpose scenarios, using modular design to reduce costs. Furthermore, tech giants like XPeng, Xiaomi, and Huawei have all entered the field. The number of Chinese humanoid robot companies exceeded 60 in 2024, forming one of the world's densest industrial clusters.

China holds significant advantages in supply chain completeness, manufacturing cost control, and rapid iteration capabilities. The core components of humanoid robots include servo motors (providing joint power), reducers (precision transmission), torque sensors (sensing force magnitude and direction), IMU inertial measurement units (posture sensing), and high-energy-density batteries. Traditionally, the precision reducer market was monopolized by Japan's Harmonic Drive and Nabtesco, but in recent years Chinese companies such as Leaderdrive and Laifual Drive have achieved technological breakthroughs and significantly reduced costs. The maturity of China's motor industry chain (Inovance Technology, Kinco, etc.) and lithium battery industry chain (CATL, BYD, etc.) provides a solid foundation for low-cost mass production of humanoid robots. This highly localized supply chain enables Chinese robots to enter global markets at lower prices. This "hardware cost-performance" advantage is the direct reason for U.S. alarm.

Potential Impact and Industry Reactions

Impact on the Global Robot Supply Chain

The ban could reshape the global humanoid robot supply chain landscape. On one hand, American companies will gain relative protection in the domestic market. On the other hand, it may also push U.S. companies to accelerate building a "de-Sinicized" component supply system, which could raise costs and slow mass production timelines in the short term. Notably, even U.S. domestic robot companies currently depend on the Chinese supply chain to varying degrees—from motors and batteries to various sensors and structural components, Chinese manufacturing holds a significant share in the global robot component market. Complete supply chain decoupling means U.S. companies will need to find or cultivate alternative suppliers, a process that could take years.

For Chinese companies, the closure of the U.S. market means shifting focus toward domestic, European, Southeast Asian, and Belt and Road Initiative markets. The global landscape may consequently bifurcate into two camps. China's domestic market itself holds enormous potential—manufacturing upgrades, population aging, and emerging service demands are creating scaled demand for humanoid robots, which can partially offset the loss of the U.S. market.

Intensification of U.S.-China Tech Decoupling

From a broader perspective, this ban is a continuation of the U.S.-China tech decoupling trend in the robotics field. Following semiconductors, AI chips, and EDA tools, Physical AI has become the new focal point of competition. EDA (Electronic Design Automation) tools are essential software for chip design, with the global market dominated by over 75% market share held by U.S.-based Synopsys, Cadence, and Germany's Siemens EDA. Since 2022, the U.S. has imposed export controls on advanced chips and EDA tools to China, aiming to prevent China from acquiring the ability to design advanced-node chips. Now, this control logic has further extended to the humanoid robot domain.

This decoupling could both stimulate independent innovation within each side's respective ecosystem and lead to the fragmentation of global technology standards and redundant investment. Historical experience shows that fragmentation of technology ecosystems often brings efficiency losses—the separate aerospace standards developed by the U.S. and Soviet Union during the Cold War still create compatibility challenges for international cooperation today. In the robotics field, divergence in interface standards, communication protocols, and safety certification systems could create a "parallel universe" structure in the future global market.

Conclusion: The AI Race Enters Its Physical Phase

The issuance of the humanoid robot ban marks the expansion of U.S.-China AI competition from the virtual realm of algorithms and models into the tangible physical world. When AI acquires a "body," its strategic significance is no longer limited to data and computing—it directly impacts manufacturing, labor, and national competitiveness. According to the International Federation of Robotics (IFR), global manufacturing faces a shortfall of millions of workers, and humanoid robots are seen as the ultimate solution to fill this gap. Whoever first achieves scaled mass production and deployment of humanoid robots will gain a decisive advantage in global manufacturing competition.

For industry observers, this event reminds us that future tech competition will increasingly revolve around the convergence of "AI + hardware." Whoever can build a complete, self-reliant, and efficient industrial chain in Physical AI will hold the initiative in the next technological revolution. This ban may be just the beginning of a long-term contest. From a longer time horizon, the competition over humanoid robots could, like the automobile and semiconductor industries of the 20th century, profoundly reshape the global economic landscape and geopolitical order.

Note: This article is based on publicly available reports and analysis. Specific policy details are subject to official documentation.