When people imagine humanoid robots, the picture is usually dancing on a stage or performing martial arts moves in a demonstration. But the most valuable deployment scenario may in fact be one that people are least willing to talk about — cleaning toilets. In 2026, this is no longer a hypothetical: multiple companies have already deployed robots into restroom cleaning scenarios, and behind this lies a labor gap worth tens of billions.
Why Toilet Cleaning? Because Nobody Wants to Do It
The core logic of commercial cleaning is simple: the harder a job is to hire for, the greater the value of automating it.
Restroom cleaning sits precisely at the intersection of “high frequency, high difficulty, low willingness.” High-end shopping malls, airports, and transportation hubs require restrooms to be cleaned dozens of times per day; the work environment is unpleasant, labor intensity is high, and turnover rates are extremely high. In first-tier cities, the monthly cost of a single cleaner has already risen to RMB 6,000 to 8,000 (approximately USD 833 to 1,111), and even at that level, hiring remains difficult — particularly for night shifts and holiday shifts, where labor shortages are a persistent norm.
Even more critical is the aging of the cleaning workforce. Data shows that the average age of cleaning workers in China’s first-tier cities has already exceeded 55, with people over 60 accounting for a considerable proportion. As the older generation of cleaners gradually withdraws from the labor market, the younger generation is largely unwilling to take up these jobs, and the structural gap in the coming five to ten years will be difficult to fill.
This is precisely the market opportunity for robots. Unlike scenarios such as manufacturing and warehousing logistics — where automation competition is already fierce — the commercial cleaning of restrooms is a “hard-bone” segment that traditional wheeled cleaning robots cannot handle. Toilet bowl cleaning, sink wiping, and floor-corner disinfection all require flexible operational capability, which happens to be the strength of the new generation of humanoid and specialized robots.
Who Is Already Doing This? Real-World Deployment Cases in 2026
Currently, players entering the restroom cleaning robot arena fall into three main categories.
The first category comprises humanoid robot companies. Represented by firms such as Agibot and Unitree, these companies leverage the flexible operational capability of humanoid robots to attempt to accomplish complex tasks such as toilet cleaning and trash removal. Agibot’s Jingling series, spun off through a dedicated subsidiary, has explicitly positioned commercial cleaning as one of its important deployment directions. The advantage of humanoid robots lies in their strong general-purpose capability — theoretically, a single machine could complete the entire cleaning process — but the challenges are high costs and stability issues that still require validation.
The second category comprises specialized cleaning robot companies. Rather than pursuing humanoid forms, these companies target the cleaning scenario itself, developing specialized robots equipped with robotic arms. Gausium (高仙機器人) and Pudu Robotics (普渡機器人), among others, already hold mature positions in commercial cleaning; their equipment has been deployed at scale in shopping malls, office buildings, airports, and other venues, and is now extending from floor cleaning toward more refined scenarios. This category’s advantage is engineering maturity and controllable costs, though its general-purpose capability is comparatively limited.
The third category comprises cross-sector sanitation and property service enterprises. These traditional cleaning service providers, faced with rising labor costs, have taken the initiative to introduce robotic solutions, becoming important buyers driving deployment. According to BCC Research, some leading property management enterprises have already begun testing robots in high-end office buildings and shopping malls to undertake restroom cleaning tasks, achieving human-machine collaboration in which robots handle standardized processes while human workers are responsible for detail touch-ups and quality inspection.
Notably, the current mainstream deployment model is not “fully replacing human labor” but rather “robots plus human collaboration.” Robots handle high-frequency, standardized, repetitive tasks, while human workers focus on complex, personalized detail work. This model both reduces labor costs and ensures cleaning quality, making it the most viable commercial path at the present stage.
Technical Difficulties: Why Cleaning a Toilet Is Harder Than Playing Go
It may sound simple, but having a robot clean a restroom well is technically anything but easy.
The first difficulty is flexible operation in unstructured environments. The layout of every restroom is different — the height of toilets, the position of sinks, and the placement of trash bins all vary. Robots must possess strong environmental perception and adaptive capability, and cannot rely on pre-programmed fixed routines, as is possible in industrial settings. This requires visual recognition, path planning, and robotic-arm control to work in coordination.
The second difficulty is fine-grained force-control operation. Cleaning a toilet bowl requires controlling the strength and angle of the cleaning brush; wiping a sink requires adjusting motion trajectories according to different materials. This kind of “just-right” force control is the core challenge in embodied intelligence, requiring high-precision force feedback systems and control algorithms.
The third difficulty is waterproofing, corrosion resistance, and reliability. The restroom environment is humid and involves contact with cleaning agents and disinfectants; robots must have adequate protection ratings to ensure long-term stable operation. This places extremely high demands on hardware design, and is precisely a common shortcoming of many current humanoid robots.
The fourth difficulty is the balance between cost and efficiency. Even if a robot can technically accomplish the task, if it moves too slowly or costs too much, its commercial viability is limited. Currently the price of a humanoid robot ranges from several hundred thousand to over a million RMB; only when the cost can be recouped through multi-shift operation and reduced labor expenditure will large-scale deployment become possible.
Market Scale: A Hundred-Billion-Level Blue Ocean
Although the current deployment scale remains limited, the market potential of commercial cleaning robots should not be underestimated.
According to industry estimates, the market size of China’s commercial cleaning robots surpassed RMB 10 billion (approximately USD 1.39 billion) in 2025, and is expected to maintain a compound annual growth rate exceeding 30% over the next five years. If restroom cleaning — a segment scenario — can be successfully broken open, considering the vast number of application venues including shopping malls, airports, transportation hubs, hospitals, and schools nationwide, the market for the restroom cleaning segment alone could reach tens of billions in scale.
From a more macro perspective, the deployment value of commercial cleaning robots lies not only in the cleaning scenario itself, but in their role as an important entry point for embodied intelligence technology to move toward real-world commercialization. Cleaning scenarios provide robots with an ideal training ground: sufficiently high frequency to generate large volumes of real-world data; sufficiently standardized to allow gradual technical iteration; and with genuine paying demand to support commercial closed-loop formation. Robots trained and validated in cleaning scenarios can subsequently extend into more service-oriented scenarios such as elderly care, healthcare, and hospitality.
Restroom cleaning may not be a glamorous scenario, but it is precisely the kind of “dirty, tiring, difficult” work that is most likely to be first transformed by robots. In 2026, this transformation has already quietly begun.
[Disclaimer]: The above content reflects analysis of publicly available information, expert insights, and BCC research. It does not constitute investment advice. BCC is not responsible for any losses resulting from reliance on the views expressed herein. Investors should exercise caution.