Executive Summary
- Humanoid robots are just beginning to enter commercial use in small numbers, but China is rapidly ramping up production.
- Due to its industrial policies and complete industrial ecosystem, China holds the technological and manufacturing high ground in the humanoid robotics industry.
- As the technology becomes more commercially viable over the next few years, it will be difficult for the US to catch up without restarting economic warfare with China or pursuing similar industrial policies.
Introduction
Humanoid robots are rapidly leaving the realm of science fiction and entering the world of geostrategic competition. Truly general purpose, fully autonomous AI-powered “droids” are still several years away, but Chinese firms plan to sell at least 50,000 of their impressively agile preprogrammed or remotely operated humanoid robots this year – ten times more than in 2025 (The Economist, 2026; Baptista et el, 2026). China’s humanoid robot makers commanded more than 97% of global shipments in the first half of 2026 (Rai, 2026). Commercial deployments remain exceedingly rare, as most humanoids are purchased by scientific research institutes, AI training centers, or large firms for testing and demonstration purposes (The Economist, 2026). But China’s early lead in the mass production of this new technology means that, when humanoid robots do become commercially viable by around 2030, the country’s manufacturers will have a substantial lead over all foreign competition.
Analysis
The US has responded to the rapid growth of the industry, which is still largely propped up by Chinese government money, by banning the import of all humanoid robots and component parts above a certain capability threshold. But banning imports is only an effective strategy if the domestic market has production that would benefit from protection (Palmer, 2026). While the Federal Communication Commission (FCC)’s move was not wrongheaded, it cannot solve America’s fundamental problem in the sector: US firms, aside from a small number of exquisite prototypes, have not been able to produce humanoids at cost or at scale.
China’s state backing for the nascent industry has certainly contributed to its success. As a Mercator Institute report released this spring pointed out, in some cases provincial and municipal governments “subsidize up to 30% of project costs for innovation in automation technologies, or discounts for purchases of humanoids” (Cheng et al, 2026). The humanoid robot training centers which teach humanoid robots to mimic real human actions are estimated to have purchased around 70% of China’s humanoid robot output in recent years; these training centers are themselves typically backed by provincial or municipal governments (The Economist, 2026; Cheng et al, 2026).
The same pattern that has bolstered and later plagued upstart high-tech manufacturing industries in China is beginning to play out in the humanoid robotics sector. Local governments, eager to cultivate their own national champions, have created a partially circular financial structure where state subsidies, cheap credit, and other incentives help humanoid robotics firms get off the ground. The firms sell their expensive and inefficiently produced products largely to state-affiliated buyers at a loss for years until they gain the technical capacity to produce at cost and at scale, allowing them to massively expand production and revenue. The problem is that if too many localities follow this strategy, as they did with electric vehicles, the result is an excessive number of firms in the same industry, price wars, and “involution” (Cheng et al, 2026).
But like China’s electric vehicle industry, the complement to cheap and plentiful state-backed funding is a genuine and rapid process of iterative innovation, technological upgrading, and commercialization (Cheng et al, 2026). There is substantial technical overlap between humanoid robots and electric vehicles, drones, and other “new quality productive forces.” This, along with the density and completeness of China’s industrial ecosystem, may enable today’s loss-making and uneconomical droid firms to keep sprinting until they reach the point where their product is technically capable enough and cheap enough for the mass market (Park et al, 2026). When that day comes, China’s firms may dominate the industry far more completely and rapidly than any other frontier technology (Feng, 2026).
The preponderance of Chinese humanoid robots in many overseas markets will have serious commercial implications for the United States. The mental image of legions of Chinese robots marching at the orders of the People’s Liberation Army is also likely to inspire deep sense of insecurity in Washington. This remains true even if humanoid robots are a less effective combat application than other types of robots, such as quadrupeds and aerial drones. (Palmer, 2026).
The United States does have the power to deal serious setbacks to China’s nascent humanoid robotics industry. Though Chinese suppliers dominate most of the hardware ecosystem, Chinese robotics firms remain dependent on Japanese and German companies for a few extremely complex components, such as actuators (Levin, 2026). Moreover, China’s most advanced humanoids are heavily dependent on purpose-built Nvidia general processing units and Nvidia software (Cheng et al, 2026). But US export controls would likely inspire fierce Chinese retaliation.
Beijing has clearly made humanoid robots a priority. The Chinese term “embodied intelligence” suffuses Beijing’s key policy documents, including the 2025 Government Work Report and national, provincial, and municipal 15th Five-Year Plans. Embodied intelligence refers broadly to any interface between AI and the physical world; the ongoing and substantial financial support provided to humanoid robotics firms at a time of fiscal belt tightening demonstrates that humanoids are a central part of China’s embodied intelligence ambitions (PRC State Council, 2025). Any move to “suppress” the development of this industry could lead Beijing to reimpose critical mineral export controls or other forms of retaliation intended to exert maximum pressure on Washington. The Ministry of Commerce’s vehement response to the US FCC’s ban on humanoid robot imports underscored Beijing’s sensitivity to this issue (MOFCOM, 2026).
Conclusion
Humanoid robots may be one of the main economic security battlefields in US-China competition within the next few years. Unlike most of the AI stack and connected technologies, here, China has a commanding lead in productive capacity and the innovation-to-commercialization pipeline. Though China’s dependence on a few crucial foreign components and software remain key vulnerabilities, any US attempt to take advantage of this leverage would likely trigger Chinese retaliation serious enough to impair America’s own tech-security ambitions. As a result, there may be little the US can do to tip the productive balance in its favor without risking economic warfare with China or copying pages from China’s own industrial policy playbook.
Bibliography
Baptista, Eduardo; Jiang Xihao; Li Qiaoyi. (2026) “Robots Can Outrun Humans, But Can They Plug in a Cable?” Reuters. 23 August. Available at: https://www.reuters.com/world/asia-pacific/robots-can-outrun-humans-can-they-plug-cable-2026-08-23/
Billman, Savannah. (2026) “China Readies for a Record Number of Humanoids.” The Wire China. 9 August. Available at: https://www.thewirechina.com/2026/08/09/china-readies-for-a-record-number-of-humanoids/
Cheng, Wendy; Arcesati, Rebecca; Junusova, Altynay. (2026) “Embodied AI: China’s Ambitious Path to Transform Its Robotics Industry.” Mercator Institute for China Studies (MERICS). 30 April. Available at: https://merics.org/en/report/embodied-ai-chinas-ambitious-path-transform-its-robotics-industry
Feng, Coco. (2026) “Commercial humanoid robots in China may soon do laundry, make beds, care for elderly.” South China Morning Post. 21 May. Available at: https://www.scmp.com/tech/article/3354371/commercial-humanoid-robots-china-may-soon-do-laundry-make-beds-care-elders
Levin, Henrietta. (2026) ‘Embodied AI and the Transformation of China’s Robotics Industry, Pekingology[Podcast]. 23 July. Available at: https://www.csis.org/podcasts/pekingology/embodied-ai-and-transformation-chinas-robotics-industry
Palmer, James. (2026) “The US Can’t Catch Up to China’s Robots by Banning Them.” Foreign Policy. 25 August. Available at: https://foreignpolicy.com/2026/08/25/china-robot-race-humanoids-drones-us-ban-restrictions/?tpcc=china_brief&utm_source=Sailthru&utm_medium=email&utm_campaign=China%20Brief%2008252026&utm_term=china_brief
Park Ju-min; Chen, Laurie; Baptista, Eduardo. (2026) “China Robot Makers Seek to Turn Humanoid Hype into Useful Work.” Reuters. 19 August. Available at: https://www.reuters.com/world/asia-pacific/china-robot-makers-flock-beijing-show-seek-path-mass-adoption-2026-08-19/
Rai, Saritha. (2026) “China Humanoid Makers Hold 97% of Global Shipments, Report Says.” Bloomberg. 10 August. Available at: https://www.bloomberg.com/news/articles/2026-08-10/china-humanoid-makers-hold-97-of-global-shipments-report-says
The Economist. (2026) China is Training up Thousands of Humanoid Robots.” The Economist, 23 August. Available at: https://www.economist.com/business/2026/08/23/china-is-training-up-thousands-of-humanoid-robots