Chinese companies are rapidly advancing in the development and deployment of humanoid AI robots, shifting them from novelty acts to workers in factories and other industries, according to an analysis by Morgan Stanley. This development fuels concerns about artificial intelligence's potential impact on employment and China's growing dominance in technological innovation.
Embodied AI refers to artificial intelligence that has a physical form, such as robots, drones, or self-driving cars, in contrast to purely digital AI.
The implications for global investors and markets are significant, especially if China gains a substantial lead in the "robot race." A previous incident in 2025 saw China's DeepSeek release powerful AI at a fraction of the cost of U.S. competitors, causing a temporary $1 trillion drop in stock values, including a nearly $590 billion decline for chip designer Nvidia. While that market reaction was short-lived, future disruptions could be more enduring.
China appears to have an advantage in humanoid robot production, with companies scaling up operations faster than elsewhere, partly due to government subsidies. Plans are in place to utilize robots in over 100 "high-value" applications by the end of the current year, beginning with industry and logistics. China's access to rare earth minerals necessary for robot construction further bolsters its position.
Western companies are also investing heavily in humanoid robotics. Tesla is developing its Optimus robot, which CEO Elon Musk has suggested could become the company's most significant product. Boston Dynamics, now owned by Hyundai, has created the Atlas humanoid robot, capable of working on assembly lines, and its robotic dog, Spot, used for patrolling industrial sites. In the UK, Google DeepMind is developing AI software for robots like Atlas, while the London-based startup Humanoid has become Europe's first robotics "unicorn" with a valuation exceeding $1.35 billion. Engineered Arts, based in Cornwall, has developed the Ameca robot, known for its social interaction capabilities.
The potential for widespread job displacement is a primary concern, as robots could perform tasks that are difficult or dangerous for humans, but also those currently done by human workers. Beyond employment, the physical nature of embodied AI introduces risks of accidents, espionage, malfunctions, or even hijacking. The tendency for software-based AI to exhibit unpredictable or deceptive behavior raises questions about the control of physical AI systems.
As humanoid robots become increasingly prevalent, democratic societies face the challenge of establishing effective control mechanisms to mitigate potential dangers before the technology surpasses human oversight.