From Dance Floor to War: China's PLA Is Fast-Tracking Humanoid Robots for Combat

Two days after the World Humanoid Robot Games wrapped up in Beijing last month — an event where robots boxed, danced, fought wushu duels, and occasionally fell apart on camera — the People’s Liberation Army’s official newspaper delivered its verdict. The PLA Daily called for researchers to accelerate the transfer of cutting-edge technologies from laboratories to military training grounds, and to do it for robotic “combatants.”

That editorial was not an isolated opinion piece. According to a Reuters review of more than 100 Chinese military procurement notices, academic studies, patents, official publications, government records, and defense-company materials — much of it previously unreported — China’s defense establishment is systematically accelerating research into the military uses of humanoid robots and planning for their eventual wartime deployment.

What the Reuters review found

The reporting, published September 7 by Eduardo Baptista, documents three parallel tracks of activity. First, Chinese military institutions are actively testing humanoids against battlefield requirements — exposing the machines to conditions closer to combat than the controlled demos that dominate social media. Second, procurement bodies are seeking to acquire robots and the technologies needed to train them. Third, military researchers are studying how humanoids might function alongside human troops, floor by floor and room by room.

The effort gained visible momentum across 2025 and 2026, with the technical focus converging on three hard problems: robot perception, manipulation, and — critically — training data. That last item is telling. The gap between a robot that can dance on stage and one that can operate in a building full of adversaries is largely a data problem, and the PLA appears to know it.

The urban assault scenario

The most concrete vision in the records comes from an August 2025 paper by researchers at the National University of Defense Technology’s test center in Xi’an — NUDT is the PLA’s premier research institution. The study models an urban assault force built around six “combat robots,” a mix of humanoids, robot dogs, and unmanned vehicles, split between two assault teams. Working with ground troops, the robots would clear an enemy-occupied building floor-by-floor and room-by-room.

The authors projected that the equipment for such a force could be available within five to ten years. Notably, the paper did not specify the robots’ rules of engagement, the weapons they would carry, or how they would handle civilian encounters — the questions that matter most to everyone who is not a military planner.

Chinese military commentary goes further. A July 2025 PLA Daily article argued that humanoid missions could eventually evolve “from supporting combat toward primary combat,” and openly discussed how a robot might be authorized to fire on a living target that had been verified by a human. In December, the PLA’s Eastern Theater Command released an AI-generated montage depicting military robots — humanoids and quadrupeds alike — overwhelming Taiwan’s defenses in a hypothetical future conflict.

Why China, and why now

The structural advantage is industrial. Chinese manufacturers accounted for roughly 95% of global humanoid shipments in 2025, according to BofA Global Research. That commercial dominance gives the PLA something no other military has: direct access to a vast, cost-optimizing, rapidly iterating humanoid supply chain while it develops military applications in parallel.

There is also an uncomfortable historical footnote. Reuters reported in August that Unitree — among the world’s leading producers of humanoids and quadrupeds — based designs for its best-selling robot dogs on innovations originally financed by the U.S. military. The Pentagon added Unitree to its list of Chinese military companies in June, calling it a “contributor to the Chinese defense industrial base.”

The Ukraine war looms over all of this. Semiautonomous drones using AI for navigation and targeting have transformed reconnaissance and attack, while ground robots carry supplies and retrieve casualties. Every major military is drawing lessons, and the lesson China has drawn appears to be that autonomous ground systems are the next frontier.

Not science fiction — but not reality either

The caveats in the Reuters report are as important as the findings. Reuters found no evidence that China has deployed an armed humanoid with any operational PLA unit. The robots remain energy-intensive and unreliable outside controlled demonstrations. Battlefield environments pose variables that challenge current machines in unpredictable ways, as Carnegie Mellon mechanical engineering professor Aaron Johnson noted.

Juo-Min Chou, a researcher at Taiwan’s Institute for National Defense and Security Research, offers a sober assessment of the form factor itself. Improvements in balance, perception, and endurance could make humanoids genuinely useful in buildings, tunnels, and ship interiors, where “two hands and two feet could, in theory, combine mobility, climbing and manipulation.” But for simpler reconnaissance or logistics missions, quadruped, tracked, or wheeled robots would generally be cheaper and easier to deploy at scale.

In other words: the humanoid is the aspirational end-state, not the workhorse. The near-term military robot is more likely to have four legs or treads.

The race is two-sided

China is not alone. The U.S. Army launched a competition last year to develop “militarized humanoid capabilities” that could work alongside soldiers, with potential roles including reconnaissance, security, obstacle clearing, hazardous-material operations, and offensive and defensive missions in urban terrain. Up to ten finalists were due to test their systems with U.S. military experts this year, with as much as $1.25 million available for follow-on contracts.

But the competitive asymmetry is stark. The U.S. robotics industry trails far behind China in developing two-legged and four-legged machines, and the Pentagon has no equivalent of China’s 95%-of-global-shipments industrial base to draw on. As UCLA professor Dennis Hong put it: “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.”

What to watch

The trajectory from dance floor to battlefield now has a documented institutional pathway: procurement notices, test protocols, academic modeling, and official doctrine all pointing the same direction. The five-to-ten-year window in the NUDT paper means that decisions being made now — about autonomy rules, training data collection, and human-on-the-loop firing authority — will define what the battlefield of the early 2030s looks like.

The unanswered questions are the ones the papers omit: rules of engagement, weapons fit, and civilian encounter protocols. Those gaps are where policy, ethics, and arms-control debates will have to catch up to engineering. Whether they do before the robots are ready is, increasingly, the question that matters.