China is rapidly scaling the manufacture of humanoid robots, with domestic output projected to exceed 100,000 units in 2026 and Chinese companies already accounting for the vast majority of global shipments.
This production surge has prompted questions about whether the United States is falling behind in the race to commercialize general-purpose humanoid machines.
According to China’s Ministry of Industry and Information Technology, national humanoid robot production is on track to surpass 100,000 units this year. In the first half of 2026 alone, global shipments reached approximately 19,100 units, a 272% year-on-year increase, with Chinese manufacturers capturing more than 97% of the total.
Leading firms include AgiBot (also known as AGIBOT), which recently reported building its 20,000th robot overall, and Unitree Robotics, which has shipped thousands of units. Other players such as UBTECH have opened dedicated factories capable of producing more than 10,000 robots annually on high-speed lines.
Morgan Stanley has repeatedly raised its forecasts for Chinese shipments, most recently projecting 50,000 units for 2026, citing faster commercialization, falling component costs, and a maturing domestic supply chain. Chinese makers benefit from integrated manufacturing ecosystems, lower costs for actuators and sensors, and strong government support aimed at deploying robots in factories, logistics, and service settings.
In contrast, US and Western companies remain at earlier stages of volume production.
Tesla has ramped Optimus output at its Fremont facility to several hundred units per week for primarily internal use and data collection, with ambitions for higher volumes later. Figure AI, Agility Robotics, Apptronik, and others have focused on targeted commercial pilots such as Figure’s work at BMW plants rather than high-volume consumer or general shipments. Combined output from major American firms is estimated in the low thousands for 2026, far below Chinese levels.
Analysts say that China’s advantage stems from its manufacturing scale and ability to iterate quickly on hardware at lower cost, creating a data flywheel from real-world deployments. US firms often emphasize advanced AI software, higher reliability for industrial tasks, and partnerships with major automakers or logistics companies.
However, the disparity in unit volume means Chinese companies are accumulating far more operational hours and sensor data, which could accelerate improvements in physical AI models.
Challenges persist on both sides. Many Chinese robots are still used for research, demonstrations, or lighter service roles rather than continuous heavy industrial work, and finding sustained commercial buyers remains a hurdle. US developers face higher costs, supply-chain constraints for specialized components, and the difficulty of scaling complex, software-intensive systems.
But the gap in mass production is clear. China currently leads in sheer volume and manufacturing capacity. Whether this translates into long-term technological dominance will depend on real-world performance, reliability, and the ability of American companies to leverage strengths in AI and systems integration as production scales.