Humanoid robots set to reshape manufacturing

A Roland Berger report estimates the humanoid robotics market could generate $300 billion to $750 billion in annual revenue by 2035, potentially expanding to $4 trillion by 2050—comparable to the global automotive industry. The shift is driven by structural labor shortages and declining working-age populations, which could drop by 22% in some regions by 2050, outpacing traditional automation solutions. Unlike conventional industrial robots, humanoid systems are designed to integrate into existing human-centric infrastructure without requiring factory redesigns. Future operating costs may drop to around $2 per hour, transforming manufacturing and logistics, though mass commercialization remains years away. The study identifies AI architecture as the primary bottleneck, with humanoid robots relying on vision-language models and vast real-world sensor data for training. Access to this data has become a competitive advantage, while regulatory frameworks and safety standards lag behind hardware advancements by 3–5 years. Current safety protocols, designed for fenced-off automation, are inadequate for dynamic, human-sharing workspaces, requiring new testing and certification. Chinese developers are leading in deployment, producing around 15,000 humanoid robots last year, accelerating cost reductions and training capabilities. Meanwhile, North American and European firms prioritize AI-first approaches backed by venture capital. Early industrial applications are expected in repetitive tasks like unpacking, transporting items, and warehouse logistics, with broader capabilities expanding as software improves. The report advises manufacturers to collaborate with robot developers by providing real-world factory training environments. Durability, safety, and liability concerns will dictate deployment speed, with harmonized regulations needed before large-scale adoption. Opportunities for early adopters include cost savings and operational efficiency gains in labor-constrained sectors.