At the World Humanoid Robot Games at the National Speed Skating Oval in Beijing, China on August 26, 2026, the two-legged machines were displayed in a civilian setting. Two days after the Games ended, the People’s Liberation Army’s official newspaper urged researchers to move advanced technology from laboratories into military training environments for robotic ‘combatants’.
The call reflects a broader push by China’s defence establishment to adapt humanoid robots for war. A review of more than 100 Chinese military procurement notices, academic studies, patents, official publications, government records and defence-company materials shows that the People’s Liberation Army, or PLA, has been building capabilities for their use in wartime. Military planners have considered roles including clearing buildings, reconnaissance and, eventually, primary combat alongside ground forces.
The work gathered pace in 2025 and 2026, with efforts concentrated on perception, manipulation and the collection of training data. The PLA has issued multiple tenders for humanoid robots since May 2025. One tender that month sought a humanoid robot together with installation and technical training.
In June 2026, the PLA allocated approximately $300,000 for a system to gather and label camera, radar and motion data for humanoids. The data was to include information on terrain that the machines could cross. Such work is a basic, if unglamorous, part of making a robot function outside the polished confines of a demonstration floor.
Military institutions have also begun setting out operational concepts. An August 2025 paper by researchers at the test centre of the National University of Defense Technology, or NUDT, in Xi’an described six ‘combat robots’ assigned across two assault teams. The proposed group combined humanoids, robot dogs and unmanned vehicles.
The paper envisaged equipment of that kind becoming available within five to ten years. It gave no account of rules of engagement, weapons or how the machines would respond to civilians. Those omissions matter because the boundary between a machine that carries equipment and one authorised to use force is the central issue in any combat role.
PLA Daily addressed that question in July 2025, when it wrote that humanoid missions could develop “from supporting combat toward primary combat”. The article considered a system under which a robot could be permitted to fire at a living target once a human had verified it. The available material does not establish whether the PLA intends humanoids to receive autonomous authority to use lethal force, or whether any firing function would remain under human control.
The military’s interest has extended into training. An NUDT competition in June 2025 included a simulated infiltration exercise in which robots entered an area, mapped it and found targets. Other scenarios at the competition involved identity checks, discipline inspections and security patrols on military bases. These exercises suggest that military planners are considering both combat and routine security functions.
State-owned defence company Norinco has promoted its Fuxi humanoid for sentry work, reconnaissance and hazardous assignments. The company says the machine can be remotely operated by a human controller. In December, the PLA’s Eastern Theater Command released an artificial intelligence-generated video showing humanoid and quadruped robots breaking through defences in a hypothetical conflict involving Taiwan.
China’s industrial position gives the military programme a large civilian base to draw from. Chinese manufacturers accounted for about 95% of global humanoid shipments in 2025, according to BofA Global Research. The brief records do not show how closely civilian robotics companies and military procurement agencies are coordinating, though military academies, defence contractors and official publications are all contributing to the discussion of applications and training.
There are substantial limits to the technology. Reuters found no evidence that China has deployed an armed humanoid with an operational PLA unit. Current robots use considerable energy and remain unreliable beyond controlled demonstrations. Wang Yonghua, a researcher at China’s Academy of Military Sciences, wrote in a November commentary that movement, perception and intelligence remain unresolved problems.
Wang nevertheless argued that technical advances, cheaper production and industrial-scale manufacturing could change the calculation. If those conditions are met, he wrote, “humanoid robots will stream onto the battlefield.” For now, the PLA’s programme is better understood as a concerted effort to prepare for that possibility than as evidence that humanoid soldiers have arrived.
Comments (0)