The Global State of Robotics — 2026
Where the money in robotics is, what the humanoid stack really costs, and an honest US-vs-China scorecard.
Robotics in 2026 is not one market but several, and the humanoid hype is only a slice of it. This briefing maps the whole taxonomy — industrial robots (工业机器人), the fast-growing professional-service segment nobody covers, quadrupeds, humanoids (人形机器人), and the newest, most contested category: the “brain” layer of vision-language-action models.
Then it opens up the humanoid itself, where the real costs and limits live: actuation (30–60% of the bill of materials), the magnet problem, the dexterity wall in hands and tactile sensing, power, and the data scaling law behind the VLA brain. Country by country, it profiles the US field (Figure, Tesla Optimus, Agility, Apptronik, Boston Dynamics, 1X) and China’s (Unitree 宇树科技, AgiBot 智元机器人, Galbot 银河通用机器人), and closes with a structural US-vs-China scorecard, a signal-versus-marketing filter, and the bottlenecks ranked.