From Reconnaissance to Combat: China’s Plan for Humanoid Robots in Future Warfare

Amy Adelaide
10 hours ago
9 min read

China’s rapid progress in humanoid robotics is beginning to intersect with an area where technological competition carries unusually high strategic consequences, military applications. What recently appeared on a sports stage, humanoid robots running, boxing, wrestling and performing coordinated movements, is increasingly being examined by Chinese defense researchers as a potential foundation for future battlefield systems.

The transition is still at an early stage. There is no evidence that China has deployed armed humanoid robots as operational People’s Liberation Army units, and today's machines remain constrained by energy requirements, reliability problems and the difficulty of operating autonomously in unpredictable environments. Yet procurement activity, military research, academic competitions and defense-industry development indicate that China is not treating humanoid robotics solely as a commercial technology.

The emerging objective is broader, develop robotic systems capable of operating alongside soldiers, entering hazardous environments, performing reconnaissance and eventually taking on increasingly demanding combat functions.

From Humanoid Entertainment to Military Technology

The World Humanoid Robot Games held in Beijing in August 2026 provided a striking demonstration of how quickly the technology has progressed. Robots competed in events involving running, martial arts, tug-of-war and other physical tasks. Some machines demonstrated impressive speed and increasingly sophisticated balance.

For the commercial robotics industry, such events demonstrate improvements in locomotion, coordination and mechanical control. For military planners, however, the same capabilities have a different significance.

A robot capable of maintaining balance while moving rapidly over uneven terrain could potentially become a platform for reconnaissance. A machine capable of manipulating objects with two arms could interact with doors, equipment and obstacles designed for humans. A robot capable of navigating stairs and confined spaces could theoretically operate inside buildings, tunnels or ships.

The military value therefore comes less from the spectacle of humanoid movement itself and more from the possibility of reproducing human mobility in environments where conventional robotic platforms encounter limitations.

The People's Liberation Army Daily's call for faster movement of advanced technologies from laboratories toward military training environments illustrates this changing perspective. China appears increasingly interested in converting advances in commercial robotics, artificial intelligence and embodied intelligence into military capabilities.

Why Militaries Are Interested in Humanoid Robots

The fundamental argument for military robotics is straightforward, reduce human exposure to dangerous environments.

Robots can potentially be sent into areas where soldiers face extreme risks, including contaminated zones, damaged buildings, tunnels, mine-threatened areas and heavily contested urban environments.

Humanoid machines offer an additional advantage because the physical world has largely been designed around human bodies.

Buildings contain stairs, doors, ladders, handles, vehicles and tools intended for people.

A humanoid robot with two legs and two arms could theoretically use infrastructure that would be difficult for a wheeled or tracked machine to navigate.

This does not mean humanoids are automatically superior to other robotic platforms. In many situations, conventional designs remain more practical.

Platform	Potential military strength	Major limitation
Wheeled robot	Efficient movement on prepared surfaces	Poor performance on complex obstacles
Tracked robot	Stability and difficult-terrain mobility	Less suited to human-designed environments
Quadruped robot	Strong balance and terrain adaptability	Limited human-style manipulation
Humanoid robot	Mobility plus two-handed manipulation	High complexity, energy demand and reliability challenges
Drone	Rapid aerial reconnaissance and reach	Limited endurance and payload constraints

The most compelling role for humanoids may therefore be situations where mobility and manipulation must be combined.

China’s Urban Warfare Concept

Chinese military research has already explored scenarios in which mixed robotic forces could support troops during urban operations.

A 2025 study from researchers associated with the National University of Defense Technology modeled an assault force involving six robotic systems, potentially including humanoids, quadrupeds and unmanned vehicles. The systems were envisioned operating alongside human troops during the clearance of an occupied building.

The concept is significant because urban warfare is one of the most demanding environments for autonomous machines.

A building can contain stairs, narrow corridors, rooms, debris, unexpected obstacles and civilians. GPS signals may be unreliable or unavailable. Communication links can be obstructed. Lighting conditions can change dramatically. Objects may be partially hidden or damaged.

A robot designed for such an environment needs more than strong motors. It requires perception, navigation, dexterous manipulation, situational awareness and robust decision-making.

That explains why Chinese military interest is expanding beyond robot hardware toward the data and artificial intelligence systems required to operate these machines.

Perception and Training Data Could Become the Real Competitive Advantage

Humanoid robotics depends on a combination of mechanical engineering and artificial intelligence.

A robot must understand what its sensors are seeing, estimate its position, recognize objects, predict how its body will interact with the environment and coordinate many joints in real time.

This makes training data particularly important.

In June 2026, a PLA procurement document reportedly allocated approximately $300,000 for a system intended to collect and label camera, radar and motion data relevant to humanoid robots, including information about traversable terrain.

The significance of such investment goes beyond the individual procurement.

For an embodied AI system, physical-world data can be used to improve perception, navigation and manipulation. A robot that has encountered thousands of variations of doors, stairs, surfaces, obstacles and environmental conditions can potentially develop more robust behavior than a machine trained primarily through controlled laboratory demonstrations.

China's enormous commercial robotics sector could provide an additional ecosystem for this development. According to the material reviewed by Reuters, Chinese manufacturers accounted for approximately 95% of global humanoid shipments in 2025.

That commercial scale matters strategically because military robotics does not develop in isolation. Improvements in actuators, batteries, sensors, motors, control systems and manufacturing can emerge from commercial demand and subsequently become useful for defense applications.

The Move Toward Teleoperation

Full autonomy is not necessarily the immediate objective.

One of the most practical approaches is teleoperation, where a human operator controls the robot remotely. This allows the machine to provide physical access to dangerous environments while retaining human decision-making.

Norinco's Fuxi humanoid represents this direction. Company materials describe potential applications including sentry duties, reconnaissance, patrol functions and dangerous missions, while its teleoperation capability allows a person to control the robot remotely.

Teleoperation offers a critical compromise.

Instead of requiring artificial intelligence to independently interpret every battlefield situation, the robot can perform physical actions while a human remains responsible for important decisions.

This could be particularly valuable during the transitional period between today's robotics and genuinely autonomous military systems.

However, teleoperation introduces its own vulnerabilities. Military communications can be disrupted, delayed or jammed. An operator may have limited situational awareness compared with a person physically present in the environment. Bandwidth requirements can also become substantial when robots transmit multiple high-resolution sensor streams.

The future military value of humanoids may therefore depend on a hybrid architecture combining autonomous movement with human supervision.

The Russia-Ukraine War Is Accelerating the Robotics Debate

The evolution of warfare in Ukraine has demonstrated the strategic importance of unmanned and increasingly autonomous systems.

Drones have transformed reconnaissance, targeting and attack operations, while ground robots have been used for logistics and casualty-related missions. Artificial intelligence is increasingly being integrated into systems that must operate under conditions where communications, visibility and human access can be constrained.

These developments provide military planners with real-world evidence about the advantages and weaknesses of autonomous machines.

China has closely studied the conflict and its implications for future warfare. The lesson is not necessarily that humanoids are the next logical step. Rather, it is that autonomous systems are becoming increasingly important and that militaries need platforms capable of operating in environments where human access is difficult.

Humanoids are one possible component of that broader robotic ecosystem.

Infiltration, Security and Reconnaissance

Chinese military research has also considered missions extending beyond direct combat.

A 2025 competition associated with NUDT reportedly included an exercise simulating infiltration behind enemy lines. Robots were expected to enter an area, map it and locate targets.

Other envisioned applications included identity verification, security patrols and military-base monitoring.

These missions may prove more realistic in the near term than autonomous offensive operations.

Reconnaissance does not necessarily require a robot to make irreversible decisions about the use of force. A machine can gather imagery, map terrain and identify potential objects while leaving final judgments to human personnel.

Similarly, dangerous inspection and security tasks could exploit the robot's physical mobility without requiring complete battlefield autonomy.

This suggests a possible progression in military humanoid deployment:

Training and experimentation
Remote-controlled logistics and inspection
Reconnaissance and surveillance
Hazardous-environment operations
Human-machine teaming
Increasingly autonomous support missions
Potentially more complex combat roles

The transition between these stages will depend heavily on reliability, cost, communications and artificial intelligence performance.

The United States and China Are Entering a New Robotics Competition

China is not alone in exploring military humanoids.

The U.S. Army has pursued militarized humanoid capabilities intended to operate alongside soldiers, with potential roles including reconnaissance, security, obstacle clearing, hazardous-material operations and missions in urban terrain.

The competition illustrates an important strategic development. Humanoid robotics is becoming part of a broader race involving artificial intelligence, autonomous systems, advanced manufacturing and military human-machine teaming.

China's commercial advantage in humanoid and quadruped manufacturing gives it a significant industrial foundation. The United States, meanwhile, retains major strengths in artificial intelligence research, defense technology, advanced computing and military systems integration.

The competition will therefore not be determined simply by which country produces the fastest robot.

The decisive advantage could come from whoever can integrate robotics with sensors, AI models, communications, manufacturing, batteries, battlefield networks and command systems most effectively.

The Biggest Obstacles Are Still Technological

Despite impressive demonstrations, humanoid robots remain far from being reliable autonomous soldiers.

Energy consumption is a major problem. Bipedal locomotion requires continuous control and substantial power, while military systems also need endurance, payload capacity and ruggedization.

Reliability is another challenge. A robot that performs successfully during a controlled competition may behave very differently in mud, dust, rain, darkness, smoke, debris or electronic interference.

Perception is equally difficult. Recognizing a person or object in a laboratory is not the same as interpreting an ambiguous battlefield scene containing civilians, combatants, damaged structures and rapidly changing circumstances.

There is also a fundamental cost question.

For reconnaissance and logistics, cheaper wheeled, tracked or quadrupedal platforms may often provide greater capability per dollar. Humanoids make the strongest case where human-like manipulation and mobility provide a unique operational advantage.

The Ethical and Legal Problem Is Even Harder

Military humanoids raise questions that extend beyond engineering.

If a robot is merely carrying supplies or inspecting a dangerous building, human control may remain relatively straightforward. The situation becomes substantially more complicated when an autonomous system participates in the use of lethal force.

International humanitarian law requires combatants to distinguish between civilians and military targets and imposes restrictions concerning wounded or surrendering fighters.

Determining surrender, intent and civilian status can require contextual judgment that remains difficult for artificial intelligence.

This is why human oversight is likely to remain an important component of military autonomous systems, at least for the foreseeable future.

The distinction between a remotely operated machine and a genuinely autonomous weapon is therefore strategically and ethically significant.

What the Next Five to Ten Years Could Bring

The most plausible near-term development is not an army of autonomous humanoid soldiers.

Instead, the technology is likely to progress through specialized roles where the advantages of robotic physical presence outweigh the disadvantages of complexity.

Military organizations could increasingly experiment with humanoids for hazardous inspection, reconnaissance, logistics, facility security, obstacle negotiation and remote operations.

As batteries, actuators, sensors, AI models and manufacturing processes improve, more demanding applications could become technically feasible.

The crucial milestone will be reliability outside demonstrations.

A military robot does not need to perform one impressive task once. It needs to operate repeatedly under uncertain conditions, tolerate damage and environmental variation, maintain communications where possible, conserve energy and provide predictable behavior.

That is a much higher standard.

Conclusion

China's growing interest in military humanoid robotics represents an important convergence of commercial robotics, artificial intelligence and defense modernization.

The technology is not yet ready to replace soldiers on the battlefield. Current machines remain constrained by energy requirements, mechanical reliability, autonomous perception and the extraordinary complexity of real-world military environments.

But the strategic direction is becoming clearer.

China is building an industrial ecosystem capable of producing humanoid machines at scale, while military researchers are investigating the perception, manipulation, training data and operational concepts required to adapt them for defense.

The most consequential development may not be a humanoid robot carrying a weapon. It may be the gradual construction of a military ecosystem in which humans and machines divide dangerous, repetitive and physically demanding tasks.

For analysts tracking the future of artificial intelligence and defense technology, the rise of military humanoids deserves attention because it represents another stage in the broader evolution of autonomous warfare. As Dr. Shahid Masood and the expert team at 1950.ai examine the intersection of AI, robotics and geopolitical competition, humanoid systems offer a particularly revealing case study of how commercial technological breakthroughs can migrate into strategic domains.

The central question is no longer whether robots can walk like humans. It is whether they can reliably understand, navigate and operate within the complexity of the human battlefield, and what happens when that capability becomes practical.