
Tech • IA • Crypto
The global race for autonomous military robots is accelerating, amid industrial promises, strategic rivalries, and ethical concerns.
The company Foundation Future Industries, launched in 2024, is developing the humanoid robot Phantom MK1, designed to use the same weapons as soldiers. Its promoter, former Marine Mike LeBlanc, argues for a “moral imperative” to send machines into combat. The company explicitly targets military uses, contrary to many U.S. players who رفض militarization.
The company claims $24 million in contracts with the U.S. military, but these largely come from inherited agreements after acquiring Boardwalk Robotics and via intermediaries like IHMC. No major new funding has been won directly, despite positioning as a military supplier.
Robots are already tested in industry, with more than 24,000 cars produced in 2025 thanks to these machines. Foundation plans factories capable of producing 5,000, then 50,000 units per year. Target costs could fall below $20,000 at scale, versus about $300,000 for a military version today.
Despite demonstrations, prototypes remain unstable. Frequent falls highlight current limits. According to expert Rodney Brooks, it may take over a decade to master complex environments. Challenges include perception, object manipulation, and mechanical reliability.
The war in Ukraine marks a turning point, with up to 9,000 drones launched per day. Autonomous systems can strike without human input when communications are jammed. Armed ground robots have already forced soldiers to surrender, illustrating a shift where machines become primary actors.
Beijing fully integrates these technologies into its military “intelligentization” strategy. Armed quadruped robots, nicknamed “wolves,” combine LiDAR sensors, cameras, and onboard intelligence. They can coordinate in swarms, reach 15 km/h, and operate as a networked collective system.
China dominates the supply chain, notably with about 80% of the global LiDAR market. Its civil-military fusion model enables rapid integration of commercial innovations into the military, accelerating operational deployment.
Systems like DeepSeek can analyze up to 10,000 scenarios in 48 seconds, versus 48 hours for humans. In the U.S., projects like Fury Orchestrator aim to automate the entire lethal chain, from detection to strike.
Regulatory safeguards are weakening. Some restrictions on military AI use have been lifted in the United States. The debate pits “human in the loop” against “human on the loop,” the latter allowing machines to act with minimal supervision.
Around 60 countries support non-binding principles to govern these weapons. The UN and the Red Cross call for a treaty requiring meaningful human control. Major powers remain divided, blocking strict regulation.
The humanoid sector could reach $29.5 billion by 2036, with nearly 1.8 million units per year. Average prices are expected to drop from $114,700 in 2024 to $37,000 in 2030, accelerating adoption.
Between rapid innovation, geopolitical rivalry, and a regulatory vacuum, combat robots are moving from prototypes to the battlefield, raising urgent questions about human control in future wars.