U.S. Army picks IHMC's Alex humanoid and four subsystem suppliers in xTechHumanoid competition

The U.S. Army has selected five winners in xTechHumanoid, a competition designed to assess how commercial humanoid robotics could support future military operations. The programme received 103 white papers; ten finalists were chosen, and nine of them demonstrated their technology at Marine Corps Base Quantico, Virginia, on 9 and 10 September 2026. The winners were announced on 11 September, and the Army published the results on 28 September.

Only one of the winners is a complete robot. The Florida Institute for Human and Machine Cognition (IHMC) was selected with Alex, a humanoid that combines impact-resilient actuators, ballistic protection and a hybrid control model mixing artificial intelligence with human operation. The other four winners address subsystems: ACES Group for integrated power systems with scalable energy storage, Hydroplane for lightweight mobile power generation intended for contested logistics environments, Ras Labs for fingertip tactile sensors that detect force and slip, and Velocity Explorations for command-and-control software that turns military intent into robot actions. Award amounts were not disclosed.

The competition was run by Army FUZE, the Army's programme for funding and testing commercial innovation, together with the Humanoid Community of Collaboration, a group created within the Army in December 2023. The stated aim is to reduce soldiers' exposure to dangerous and physically demanding tasks. The community now plans follow-on work on power and energy constraints, which the competition highlighted as a key limitation for robots operating in the field.

The list of winners is telling. It does not include any of the best-funded commercial humanoid makers tracked in SectorAtlas, such as Tesla, Figure AI or Boston Dynamics; the only complete robot comes from a non-profit research institute. The weight given to batteries, power generation, touch sensing and command software reflects a military view that endurance and control, rather than walking ability alone, are the real bottlenecks. For component suppliers, defence programmes of this kind can offer an early source of revenue that depends less on factory deployments, which remain the main commercial target for most humanoid companies.

Because xTechHumanoid is an assessment rather than a procurement, it does not create a production contract. It does, however, give the Army its first structured comparison of what commercial humanoid technology can already do in a military setting, and it indicates which technical gaps future funding is likely to target.