The $5 Trillion Humanoid Market: From Research Labs to Real-World Deployment
The robotics industry is currently standing at the precipice of a historical technological milestone. Drawing on Michelangelo’s The Creation of Adam, Aska Liu described the present moment as the beginning of an era in which humanoids move from research prototypes into practical use, calling it the "creation of humanoids". Liu, founder and CEO of EnduX, previously led Insurtech, AI-platform, and voice-bot businesses at OneConnect Financial Technology, part of Ping An Group, after working on technology-driven transformation at Salesforce and McKinsey.
Her company, the robotics distribution and deployment company EnduX, curates leading humanoid robots from top Chinese manufacturers for global distribution, providing clients with use case evaluations, hardware selection, and deployment integration services.

The foundational driver behind this robotics boom is a severe global physical labor shortage. Driven by structural changes such as an aging population, shifting generational preferences, and evolving international labor mobility policies, Liu argued that humanoid robots could become an important response to shortages in physical labor. Citing a 2025 industry report by Morgan Stanley, Liu notes that the humanoid robotics industry is projected to become a market worth approximately $5 trillion by 2050.
Historically confined to academic research, the humanoid sector is experiencing a massive paradigm shift toward industrial application. Speaking at Hannover Messe, Liu highlighted that the industry is explicitly shifting its focus from locomotion to manipulation. Because locomotion technology is maturing rapidly, developers are prioritizing dexterity and object manipulation. This transition is heavily supported by a departure from traditional programming methods toward the integration of artificial intelligence. Unlike traditional industrial robotic arms or "cobots" that are rigidly programmed for a single fixed task, AI-trained humanoids are designed to assess their surroundings dynamically. While achieving true generalization across multiple tasks remains a current software limitation, this AI integration is intended to allow robots to adapt to changing situations and execute complex, varied tasks.
This rapid technological advancement is already operational across multiple sectors. In the entertainment space, humanoids have evolved from moving like rigid machines to fluidly mimicking "kung fu masters" at the Chinese New Year Gala in just a single year, showcasing mind-blowing progress. Humanoids have also taken on prominent roles in major public and industrial facilities. UBTECH’s Walker C served as an embodied-AI tour guide at the China Pavilion during Expo 2025 Osaka. Furthermore, UBTECH’s Walker S has been tested at Zeekr’s smart factory in Ningbo for material handling and logistics tasks.
In heavy industry, the combination of visual recognition and mobile "pick and place" technology has enabled a growing number of pilots and early commercial deployments. Agility Robotics’ Digit has been deployed in GXO-operated logistics operations to move totes between autonomous mobile robots and conveyors, while a G1 model robot actively loads batteries in car manufacturing facilities. Within warehouse logistics, the Kepler K2 has been deployed to load wheel hubs from containers onto conveyor belts, while the Galbot G1 sorts and places car parts into different compartments. More complex integrations are also being explored: Boston Dynamics' Atlas has been demonstrated performing part sequencing—a common use case for manufacturers—and multiple AgiBot G2 robots have been deployed on Longcheer’s tablet-production line to perform automated inspection tasks, safely surrounded by human workers.
The Business-to-Consumer (B2C) market is equally active. The Galbot G1 robot operates in 24/7 unmanned pharmacy warehouses across 24 cities in China, fulfilling over 300,000 orders. Additionally, Galbot operates more than 100 automated retail booths across 20 Chinese cities.
Despite these breakthroughs, Liu cautions that the technology is still in its early stages and faces several critical constraints. Widespread deployment is currently limited by overall automation levels, payload capacities, software generalization limits, and high initial costs, although prices are rapidly decreasing. Uptime and reliability must undergo long-term testing, and regulatory frameworks—particularly in Europe—are still maturing.
The ultimate engineering hurdle is what Liu calls the "last mile" of humanoid robotics: dexterity. Replicating the incredible flexibility and dedicated capability of the human hand requires immense ongoing development. Nevertheless, with industry leaders already actively testing and piloting these systems, Liu emphasizes that the question for businesses is no longer where humanoids will work, but which models to deploy and how to integrate them. The time to start developing internal capabilities to welcome this transformative technology is right now.
















