Only 7,000 Humanoid Robots Sold Worldwide: The Reality Behind the AI Robotics Boom, China’s Ambitions, Factory Pilots and the Race to 1.2 Million by 2030

Humanoid robots can run, box, dance, carry parts, and dominate technology demonstrations—but how many are actually being sold and put to work? In this episode of The Daily AI Chat, we examine a striking new reality check from Reuters: only about 7,000 humanoid robots were sold worldwide in 2025 for industrial and professional-service use.The figure comes from the International Federation of Robotics, or IFR, and represents one of the first industry-wide measurements of a technology attracting billions of dollars in investment. It is a useful baseline because the excitement surrounding humanoid robots often makes the market appear much larger and more mature than it is. By comparison, approximately 542,000 conventional industrial robots were installed worldwide in 2024, while an estimated 199,000 professional-service robots were sold for transportation, hospitality, cleaning, and other tasks.The most revealing detail is what buyers are doing with those 7,000 humanoids. Many were not purchased to perform productive labor. Instead, research institutions and technology companies bought them to collect real-world movement and interaction data that can be used to train and improve artificial-intelligence models. The robots are therefore serving partly as data-gathering platforms—helping developers teach future machines how to understand physical environments, manipulate objects, and operate around people.Carmakers are among the most closely watched early adopters, but their deployments remain small. According to IFR secretary general Susanne Bieller, manufacturers are typically running pilots with single-digit or occasionally double-digit numbers of humanoids inside their plants. That is a long way from the large-scale replacement of factory workers imagined in many headlines and promotional videos.At the same time, forecasts for the next several years are enormous. Bank of America Global Research estimates that 90,000 humanoid robots could ship in 2026 and that annual shipments could climb to 1.2 million by 2030. Reaching that trajectory would require a rapid transition from research projects and controlled pilots to reliable, economically useful machines deployed at scale.We explore what counts as a humanoid robot under the IFR definition, why legs are not required, and why autonomy in environments designed for humans matters more than appearance alone. We also discuss the categories excluded from the tally—including consumer, military, and separately classified medical robots—and why those boundaries affect the market numbers.China plays a central role in the story. It is already the world's largest market for industrial robots and has showcased increasingly capable humanoids in high-profile running and boxing demonstrations. Yet impressive demonstrations are not the same as durable products. Commercial success will depend on reliability, safety, battery life, dexterity, maintenance costs, integration with existing workflows, and whether the machines can deliver a clear return on investment.The 7,000-unit baseline does not prove the humanoid boom will fail. It shows that the industry is still at an early and unusually speculative stage. Investors and customers are betting that data collection, AI progress, falling hardware costs, and factory experimentation will converge into practical machines. The critical question is whether that conversion happens quickly enough to justify today’s valuations, capital spending, and million-unit forecasts.Listen for a clear breakdown of the current market, the gap between pilots and production, the role of embodied AI, the importance of China, and the milestones that will reveal whether humanoid robotics is becoming a real industry or remaining a compelling demonstration.Source: Reuters, September 21, 2026. Reporting by Toby Sterling. Editing by Alexander Smith.