China’s military establishment is accelerating research into the battlefield use of humanoid robots, mapping out scenarios in which the machines clear buildings alongside infantry, patrol military bases and eventually fire on human-verified targets, according to a Reuters review of more than 100 Chinese military procurement notices, academic studies, patents, official publications and defense-company materials. The findings, based on previously unreported records, show a research effort that gained significant momentum through 2025 and into 2026, even as no armed humanoid has yet reached an operational People’s Liberation Army unit.
The timing of the People’s Liberation Army’s public messaging was pointed. At the World Humanoid Robot Games held in China last month, robots jumped, boxed, danced and ran 100-meter sprints faster than Usain Bolt, drawing cheers for both their advancing capabilities and their occasional on-stage collapses. Two days after the Games concluded, the PLA’s official newspaper, PLA Daily, called for researchers to accelerate the transfer of the technology on display from laboratories directly to military training grounds for robotic “combatants.”
A Review Spanning Labs, Bases and Budgets
The Reuters review found that military institutions across China are testing humanoids against battlefield requirements, working to acquire the robots and the systems needed to train them, and studying how the machines might operate alongside human troops. The documents show the research concentrated heavily on three technical areas: robot perception, physical manipulation and the training data needed to make both functional outside of controlled demonstrations.
China’s commercial position gives its military a substantial head start. Chinese manufacturers accounted for approximately 95% of global humanoid robot shipments in 2025, according to BofA Global Research, giving the PLA direct access to a rapidly scaling domestic industry even as it pursues its own military-specific applications separately.
An Urban Assault Force On Paper
Chinese researchers have already modeled what a humanoid-integrated urban assault force could look like. An August 2025 paper produced by researchers at the National University of Defense Technology’s test center in Xi’an described six “combat robots” — a mix of humanoids, robot dogs and unmanned vehicles — split between two assault teams tasked with clearing an enemy-occupied building floor by floor and room by room alongside ground troops. The paper modeled equipment its authors projected could be fielded within five to 10 years, though it did not specify rules of engagement, what weapons the robots would carry, or how they would handle encounters with civilians.
Analysts say the physical design of humanoids is what makes them attractive for exactly this kind of close-quarters, structurally complex environment. Juo-Min Chou, a researcher at Taiwan’s Institute for National Defense and Security Research, said improvements in balance, perception and endurance could make humanoids useful in buildings, tunnels and ship interiors, noting that “their two hands and two feet could, in theory, combine mobility, climbing and manipulation.” Chou added that for simpler reconnaissance or logistics missions, quadruped, tracked or wheeled robots would generally remain cheaper and easier to deploy at scale.
From Academic Forums To Live Simulations
Chinese military interest in humanoids’ combat applications began building in 2024, when the National University of Defense Technology and China’s Academy of Military Sciences co-hosted a defense-technology forum that included a dedicated session on military applications for humanoid systems. Neither institution responded to Reuters’ requests for comment.
By June 2025, that interest had moved into live simulation. A university competition hosted by NUDT featured an event simulating “infiltration behind enemy lines,” in which robots were required to enter an area, map it and locate targets. A separate portion of the competition envisioned robots performing identity checks, discipline inspections and security patrols at military bases. The research university described the competition in part as a proving ground for humanoids to eventually “move toward the battlefield.”
PLA commentary published around the same period signaled ambitions well beyond support roles. A July 2025 article in PLA Daily argued that humanoid missions could eventually evolve “from supporting combat toward primary combat,” and discussed circumstances under which robots might be authorized to fire on a living target after that target had been verified by a human operator.
The Procurement Trail
Chinese military procurement records from the same window show the research moving from theory toward acquisition. In May 2025, the PLA issued a tender to purchase a humanoid robot that included installation and technical training. Four months later, a separate procurement notice listing a National University of Defense Technology email address as its point of contact sought an “embodied humanoid robot intelligent perception and dexterous operation system.” Reuters found the records contained few specifics and could not determine whether the tenders were ultimately filled or which companies, if any, supplied the robots.
The spending continued into this year. In June 2026, the PLA budgeted approximately $300,000 for a system designed to collect and label camera, radar and motion data for humanoids, including data on the types of terrain the robots could traverse. Defense scholars say systems like this — focused on perception, manipulation and training data rather than the robot bodies themselves — are essential if humanoids are to move beyond scripted demonstrations.
Norinco’s Fuxi And The Remote-Control Workaround
By 2026, humanoid development in China had expanded from military academia into the country’s formal defense industry. Norinco, a state-owned defense conglomerate, has built a humanoid called Fuxi. Company materials published in August described the full-size robot as capable of sentry duty, all-weather reconnaissance, intelligent patrol and other roles considered dangerous for human soldiers. Norinco did not respond to a request for comment.
Fuxi can reportedly be paired with a Norinco teleoperation system that allows a human operator to control the humanoid remotely — a design choice that sidesteps what many robotics experts describe as the single biggest obstacle to military humanoids: full autonomous decision-making. Malcolm Davis, a senior analyst at the Australian Strategic Policy Institute, said he does not currently view humanoid robots as practical battlefield systems, “but in five to 10 years, they could very well be.”
Washington Is Watching The Same Technology
China is not the only military exploring humanoids for combat roles. The U.S. Army launched a competition last year to develop “militarized humanoid capabilities” that could eventually operate alongside soldiers, with potential missions including reconnaissance, security, obstacle clearing, hazardous-material operations, and offensive and defensive missions in urban terrain. Up to 10 finalists were expected to test their systems with U.S. military experts this year, with as much as $1.25 million available for follow-on contracts. Neither the Pentagon nor the U.S. Army commented for the Reuters report.
Even so, the U.S. robotics industry trails China significantly in two-legged and four-legged robot development. Reuters reported last month that Unitree, one of the world’s leading producers of humanoids and quadrupeds, based key design elements of its best-selling robot dog on innovations originally financed by the U.S. military. Separately, industry analysts tracking the broader defense-robotics sector project global military spending on unmanned and robotic systems will exceed $30 billion annually by 2027, a trajectory driven in part by early commercial military-humanoid ventures already courting Pentagon-adjacent contracts.
Unresolved Questions Of Law And Ethics
Any move toward autonomous humanoid deployment carries unresolved legal and ethical questions. The laws of armed conflict require forces to distinguish combatants from civilians and prohibit attacks on wounded or surrendering fighters. The International Committee of the Red Cross has cautioned that autonomous weapons systems may struggle to reliably interpret signs of surrender, since such judgments often depend on context, conduct and intent that current systems are not built to evaluate.
Both militaries publicly maintain that human oversight remains a requirement. The U.S. military mandates that autonomous weapons undergo legal review and realistic testing, and that commanders and operators exercise appropriate judgment over the use of force. China’s defense ministry said in March that AI-powered weapons should remain under human control. Those stated positions have not slowed China’s research into humanoids’ military applications.
“Humanoid Robots Will Stream Onto The Battlefield”
Even China’s own researchers acknowledge the technology is not there yet. Wang Yonghua, a researcher at China’s Academy of Military Sciences, wrote in a November commentary that humanoids still face unresolved problems in movement, perception and intelligence, and that technical breakthroughs, lower manufacturing costs and industrial-scale production would all be needed before humanoids become widespread military systems. If those conditions are met, Wang wrote, “humanoid robots will stream onto the battlefield.”
That assessment tracks with independent analysts, including Carnegie Mellon University mechanical engineering professor Aaron Johnson, who noted that battlefield environments present variables that challenge robots’ capabilities in ways controlled demonstrations do not, and Reuters found no evidence any armed humanoid has been deployed with an operational PLA unit to date. Still, the underlying appeal of the technology to militaries was summarized by UCLA mechanical and aerospace engineering professor Dennis Hong: “Going where we don’t want humans to go is one of the most compelling reasons to develop robots in the first place. If losing a robot means avoiding the loss of a human life, then the robot can become expendable.”