Portuguese 
English 
Spanish 
6 min of reading 
0 comments 
Service being tested in China brings domestic robots closer to cleaning routines, but still depends on human supervision and faces technical, security, and privacy challenges before achieving autonomy in homes.
AI-powered cleaning robots have started operating in apartments in China alongside human cleaners, in a service that also functions as a test in real domestic environments.
The operation is offered by the Chinese platform 58.com in partnership with the robotics company X Square, is available in Beijing and Shenzhen and costs 149 yuan for three hours, about US$ 22, according to a report by AFP published by The Straits Times.
The comparison with Rosie, the robot maid from the Jetsons, arises due to the similarity between the proposal and the idea of a machine capable of assisting with household chores.
-
Lunar meteorite found in the desert bears marks of 3 impacts and reveals a colossal asteroid collision that melted part of the Moon 3.5 billion years ago, at the same time life was beginning to emerge on Earth.
-
A giant parrot that doesn’t fly, weighs up to 4 kg, and smells like honey almost disappeared from the planet, but an extreme program in New Zealand used predator-free islands, genetics, fake eggs, and a controlled diet to save the kakapo from extinction.
-
Flush system promises to save up to 50% of water in the bathroom by replacing traditional parts with a simple tilting tank, without a button, without a seal in the drain, and with an adjustment from 1 to 6 liters per use.
-
Archaeologists descend a forgotten staircase in Aswan, Egypt, and find a huge limestone sarcophagus beneath the Aga Khan necropolis, with hieroglyphs, mummies, and signs of a cemetery used for centuries.
At the current stage, however, experts consulted by AFP state that the technology is still far from the autonomy shown in the animation.
The equipment can collect trash, identify areas of the apartment with cameras, and fold some pieces of clothing, but depends on supervision and still faces limitations given the variety of objects, furniture, and unforeseen events in a residence.
In Beijing, cleaner Lin Meiqiong, 56, started working with an unusual colleague: a white and silver robot, on wheels, equipped with cameras and two mechanical claws.
While she performs tasks that require more adaptation to the environment, the Quanta X1 Pro carries out more structured actions, such as picking up items from the floor and attempting to organize clothes left on the sofa.
“It’s definitely different,” Lin told AFP during the service. “Before, I did everything by myself. This has reduced the workload a bit.”
The statement shows how the model is being applied at this moment: the robot does not replace the worker but starts to share simple tasks in a cleaning team formed by people and machines.
How the AI cleaning robot works in China
The Quanta X1 Pro enters the apartment accompanied by an engineer from X Square.
The human presence serves to monitor the operation, intervene when necessary, and observe how the machine reacts to an environment that was not previously organized for it.
Upon arrival at the location, the system uses cameras to map the space and recognize points where it can operate.
In a demonstration described by AFP, the robot collected trash and folded a pair of dark pants, stretching the fabric with its claws before leaving it in aligned parts.
The task took several minutes, which highlights, according to experts cited in the report, the gap between current execution and the agility of a person trained for domestic work.
Engineer Hu Bowen, from X Square, stated that future versions should respond to voice commands and even converse with residents.
The forecast indicates the companies’ goal to bring these robots closer to more complete domestic assistants, although the limitations observed in the tests still place this application in its initial phase.
Since the launch of the service in March, around 200 households have hired the experience, according to the AFP report.
One of the clients, Tan Pei, who works in advertising, stated that she decided to test the robot out of curiosity, to “see what it was capable of.”
Despite the limitations, she said she was surprised to see it fold a pair of pants “very well.”
Real homes become testing grounds for domestic robots
The interest of companies is not limited to cleaning.
For companies like X Square, placing robots in real apartments allows collecting data that is difficult to reproduce in the laboratory.
Homes have varied conditions, such as furniture in different positions, clothes wrinkled in unpredictable ways, fragile objects, irregular lighting, and narrow spaces.
This type of challenge is central to the so-called embodied artificial intelligence, or embodied AI.
Unlike language models trained with large volumes of text, robots need to learn to act in the physical world, calculating distance, force, weight, position, friction, and movement.
Christoforos Mavrogiannis, a researcher at the University of Michigan, explained to AFP that there is still no kind of “internet of robots.”
According to him, placing the machine in the field and studying what happens provides more useful information than keeping it only in the laboratory.
Reuters also reported, in April 2026, that Chinese robots attracted attention in races, dances, and public demonstrations, but still face difficulty in simple domestic tasks for humans, such as tidying up a messy room, folding clothes, or dealing with objects out of position.
Wang Qian, CEO of X Square Robot, summarized the problem during an event in Beijing.
“The hardware is largely ready. But the brain has not yet caught up,” he stated, according to Reuters.
For the executive, when a machine manipulates objects with its hands, a minimal error can compromise the entire action.
Embedded AI gains ground in Chinese robotics
The race for domestic robots is part of a broader movement in the Chinese robotics industry.
Companies in the country have been trying to take humanoids and service robots beyond factories and public demonstrations, with tests in situations closer to everyday life.
In Shenzhen, the partnership between 58.com and X Square was presented as a service where the human cleaner performs more complex tasks, while the robot handles repetitive actions, such as cleaning tables, collecting waste, and helping with fabric organization.
The state agency Xinhua described the model as a dual team, consisting of a cleaning professional and a machine.
The city also concentrates a significant part of the country’s technology sector.
According to Xinhua, Shenzhen brings together more than 2,600 artificial intelligence companies and over 70,000 companies linked to the robotics chain.
This business environment helps to contextualize why tests with service robots are being conducted in homes and commercial spaces in the region.
The sector has also received significant investments.
According to the business database ITjuzi, cited by AFP, investors allocated more than 57.7 billion yuan to the Chinese embedded artificial intelligence industry in 2026, an amount already exceeding the total recorded in 2025.
Another Chinese company, GigaAI, presented in May 2026 the SeeLight S1, a general-purpose domestic robot aimed at tasks such as doing laundry, making beds, cooking simple items, and opening curtains.
According to the South China Morning Post, the company plans to test a fleet of 100 units in homes linked to high-tech sector workers before expanding the pilots to residences in Wuhan in 2027.
Cleaning robots still face technical limits
Despite advances, the adoption of cleaning robots in homes still faces technical, regulatory, and social barriers.
Human dexterity remains one of the most challenging aspects to replicate, especially in tasks that require hand adaptation, material perception, and response to non-standard objects.
Mavrogiannis stated that many companies are working to develop better robotic hands and give autonomy to these systems, but the technology has not yet reached that point.
The performance of the Quanta X1 Pro in folding clothes exemplifies this limit: the machine can perform the action, but at a slow pace and with little room for improvisation.
Privacy also appears among the main concerns.
Domestic robots need to move through bedrooms, living rooms, kitchens, and other areas of the house, recording visual and environmental information from private spaces.
Valeria Alessandra Macalupu Chira, from the Queensland University of Technology, told AFP that it is still unclear where this data goes, where it is stored, and who can access it.
Physical security is another point under discussion.
Yang Jianfei, from Nanyang Technological University in Singapore, said that the technology remains in a “very elementary” phase.
According to him, current robots still require human supervision to activate emergency stop functions, and the sector does not have widely recognized safety standards.
At the moment, the Jetsons’ Rosie remains a fictional reference for the imagination about home automation.
Tests in Chinese homes, however, show that robotics companies have already begun to bring AI machines for common everyday tasks, albeit with supervision, limited pace, and significant use for data collection.
Be the first to react!