The race to commercialize embodied artificial intelligence is moving out of the laboratory and into the living room. Chinese robotics firm GigaAI has announced plans to deploy 100 humanoid robots into households in central Wuhan this autumn for free home-service trials, marking one of the largest real-world domestic testing programs for humanoid robots to date.
The initiative is part of a broader push by Chinese robotics companies to gather the real-world data needed to train their AI models in genuinely unstructured environments, the kind of unpredictable, cluttered, and socially complex spaces that controlled laboratory settings cannot replicate.
From Labs to Living Rooms
The push into domestic environments is already underway through smaller-scale commercial pilots that are providing early insights into both the technical challenges and the market opportunity. In a collaboration between local services platform 58.com and robotics startup X Square, a human-robot cleaning service has been rolled out in Beijing and Shenzhen. Since its March launch, approximately 200 households have booked the service, which costs 149 yuan ($22) for three hours.
The robot used in this pilot, the Quanta X1 Pro, features a wheeled base, an array of cameras, and two mechanical claws designed for basic manipulation tasks. It is paired with a human cleaner, such as 56-year-old Lin Meiqiong, who noted that working alongside the robot “has reduced the workload a bit.”
For the engineers developing these systems, the value lies primarily in the data generated by operating in unstructured environments. Hu Bowen, an engineer at X Square, explained that sending robots into a “completely unfamiliar environment” is crucial for training the AI to handle the infinite variability of domestic spaces, the scattered toys, the irregular furniture arrangements, the unexpected obstacles that no simulation can fully anticipate.
GigaAI’s upcoming deployment of 100 humanoid robots in Wuhan represents a significant scaling up of this approach. By offering robots for free trials, the company aims to rapidly build a diverse, real-world dataset necessary to refine its embodied AI models for complex household tasks. The free trial model also serves a secondary purpose: building consumer familiarity and trust with domestic robots, which will be essential for eventual commercialization.
The Investment Boom Fueling the Push
These deployments are fueled by a massive influx of capital into China’s robotics sector. According to data provider ITjuzi, investment in Chinese embodied AI companies reached 57.7 billion yuan ($8.5 billion) in the first half of 2026 alone, already surpassing the total for the entire year of 2025. This investment surge reflects the widespread belief that embodied AI, systems that can perceive, reason, and act in the physical world, represents the next major paradigm shift in technology.
As China’s 15th Five-Year Plan explicitly targets the integration of consumer AI into everyday devices and domestic robots are viewed as a key growth area, state and private capital are flowing into the sector in unprecedented volumes. The government’s nationwide humanoid robot commercialization program, which aims to have 10,000 robots in commercial use by the end of 2026, is providing additional tailwinds for companies like GigaAI.
Technical and Ethical Challenges
Despite the rapid progress and massive investment, significant technical and ethical hurdles remain. Christoforos Mavrogiannis, a robotics expert at the University of Michigan, cautioned that the infrastructure required for widespread deployment is still lacking. “We don’t have a robot internet yet,” he noted, referring to the need for standardized communication protocols and cloud infrastructure to support fleets of autonomous machines operating across diverse environments.
Yang Jianfei of Nanyang Technological University in Singapore echoed this sentiment, stating that the industry is “still at a very elementary stage” in terms of the robots’ ability to reliably perform complex, multi-step tasks in dynamic environments. The gap between impressive laboratory demonstrations and consistent, reliable performance in real homes remains significant.
Privacy is another major concern that the industry has yet to adequately address. The robots rely on continuous video and audio feeds to navigate and learn, raising serious questions about data security in the home. Valeria Alessandra Macalupu Chira of the Queensland University of Technology highlighted the opacity of current data practices: “We don’t know where that data is going, where it’s located… who is looking at that information.”
For early adopters, the experience is a mix of novelty and limitation. Tan Pei, an advertising professional who tested the X Square cleaning service, summarized the current state of the technology: “Even though it’s not that perfect, there are still parts of it that surprised me.”
As GigaAI prepares to release its fleet of 100 humanoids into homes in Wuhan, the trials will provide critical insights into both the technical viability and social acceptance of robots as domestic helpers. The data gathered will likely shape the trajectory of China’s embodied AI industry for years to come, and the lessons learned in Wuhan’s living rooms may ultimately prove more valuable than any amount of laboratory testing.