China’s military is accelerating research into humanoid robots for future warfare, with defence institutions studying how machines could support soldiers in reconnaissance, logistics and urban combat.
A Reuters investigation published on September 7 examined more than 100 Chinese military procurement notices, academic studies, patents and government records. The documents show growing interest in adapting commercially developed robots for military use.
The findings do not mean China has deployed an army of humanoid combatants. Reuters found no evidence of an armed humanoid serving with an operational People’s Liberation Army unit, and researchers acknowledge that today’s machines remain unreliable in unpredictable battlefield conditions.
Proposal for Conducting Cyber Crisis Drill, Tabletop Exercise (TTEx) & CCMP Readiness Exercise
From Robot Games to Military Training Grounds
China’s World Humanoid Robot Games in August showcased machines running, boxing, dancing and performing increasingly complex movements. The demonstrations highlighted rapid improvements in balance, movement and coordination.
Shortly afterwards, the PLA’s official newspaper urged researchers to accelerate the transfer of advanced robotics technology from laboratories into military training environments.
The connection reflects a broader strategy: technologies developed for civilian purposes can also have military applications. A robot that learns to navigate a factory, carry objects or recover from a fall may eventually be adapted for dangerous military tasks.
China’s manufacturing scale gives its defence sector access to a growing pool of companies, engineers and components. According to research cited by Reuters, Chinese manufacturers accounted for approximately 95% of global humanoid robot shipments in 2025.
That industrial advantage does not automatically translate into battlefield superiority. Military operations require capabilities that are far more demanding than controlled demonstrations.
Six Robots in an Urban Assault Team
One of the clearest examples of China’s ambitions comes from an August 2025 research paper by the National University of Defense Technology.
The study modelled an urban assault operation involving six combat robots, which could include humanoids, robotic dogs or unmanned vehicles. The machines would operate alongside two teams of soldiers while clearing an enemy-held building.
The researchers projected that the necessary equipment could become available within five to ten years. The paper did not specify which weapons the robots would carry or establish rules for their use of force.
This is a proposed future concept, not evidence that such a force is already operational.
Urban warfare is especially difficult for robots. Buildings contain stairs, narrow passages, debris, damaged floors and unpredictable obstacles. Machines must also distinguish between combatants, civilians and other people who may be present.
A robot capable of performing a choreographed demonstration is not necessarily capable of making safe decisions in that environment.
Why Give a Military Robot Two Arms and Two Legs?
Humanoid robots are designed with a body structure resembling a person, usually including two arms and two legs.
The potential advantage is that human environments are built around our bodies. Doors, stairs, tools, vehicles and narrow corridors are generally designed for people rather than wheeled machines.
A humanoid could theoretically climb, manipulate objects and move through spaces where other robots struggle. That makes the design attractive for tasks such as inspecting dangerous structures, handling equipment or entering hazardous areas.
However, a humanoid is not always the best solution.
For transporting supplies or conducting basic reconnaissance, a tracked, wheeled or four-legged robot may be cheaper, more stable and easier to maintain. Military planners must therefore decide whether the added complexity of a human-like body provides a genuine operational benefit.
What Is Embodied AI?
Embodied AI refers to artificial intelligence that can interact with the physical world through a body, such as a robot.
Unlike a chatbot that processes text, an embodied system must interpret information from cameras and sensors, understand its surroundings and control physical movement.
For example, a robot entering a damaged building may need to identify a blocked doorway, choose another route, maintain balance and move an object without falling.
The challenge is that physical environments are unpredictable. A small error in perception or movement can cause a machine to lose balance, damage equipment or fail its mission.
Military researchers are therefore working on perception, manipulation and training data. These capabilities are essential before robots can perform complex tasks reliably outside laboratories.
Remote Control Versus Autonomous Decisions
China is exploring both remotely operated and increasingly autonomous robotic systems.
A remotely operated robot is controlled by a human who receives information from its sensors and directs its movements. An autonomous robot can perform certain tasks independently within the limits of its programming and training.
The distinction becomes critical when weapons are involved.
A robot used to carry supplies presents different risks from a system authorised to identify and attack targets. Questions about human control, accountability and compliance with international humanitarian law become central when machines are given greater freedom to act.
The available research does not establish that China has authorised humanoid robots to make independent lethal decisions in combat.
The Technical Problems Still Standing in the Way
Humanoid robots consume substantial amounts of energy, and their mechanical complexity creates reliability problems.
A battlefield may involve mud, smoke, damaged infrastructure, electronic interference and sudden changes in terrain. Communications can fail, sensors can be obstructed and machines may need repairs far from specialised technicians.
These conditions are much harder than the controlled settings used for sporting or promotional demonstrations.
The military value of a robot will depend on whether it can complete a useful mission consistently, not simply whether it can run quickly or perform an impressive movement.
China’s research indicates a long-term effort to solve these problems, but it does not establish a timetable for operational deployment.
What This Means for India and the Wider Region
China’s military robotics programme is relevant to India because emerging technologies can alter the balance of military capabilities over time.
Robots may eventually support border surveillance, logistics, hazardous-material handling and operations in difficult terrain. Such applications could reduce risks to personnel and improve the ability to sustain operations.
India’s response need not be limited to developing human-shaped combat machines. Investment in reliable autonomous systems, sensors, communications, electronic warfare resilience and domestic robotics manufacturing may be equally important.
The priority should be identifying practical military requirements rather than pursuing humanoid technology solely because it attracts public attention.
What this means for you: The viral videos of robots dancing or fighting should not be mistaken for proof that autonomous robot soldiers are ready for war. The important developments are research funding, military trials, reliable real-world performance and the rules governing human control over potentially lethal systems.
The420 Insight: China’s advantage lies not merely in making humanoids but in the industrial ecosystem that can support rapid experimentation. The military challenge is to turn that capacity into dependable systems that survive real operational conditions. For India, the lesson is to focus on useful autonomy rather than the appearance of the machine. The most consequential advances may come from specialised robots that carry supplies, inspect dangerous areas or assist troops, long before humanoid combatants become practical. Clear rules on human control will be as important as technical progress.
https://www.linkedin.com/company/policetechnology/
