.jpg?w=1024)
Humanoid robots in China are trading flashy demonstrations for hotel shifts, household chores and other decidedly ordinary work.
The second World Humanoid Robot Games (WHRG) is putting robots through real-world scenarios that include making beds, moving luggage, shelving books and collecting deliveries, testing whether today’s humanoids can handle the messy, multi-step jobs they may eventually be deployed to perform.
At a hotel competition in Beijing, robots have 30 minutes to move luggage, restock rooms with towels, slippers and bottled water, clear used linen and make beds. The tasks test everything from navigation and grasping to balance and the ability to manage several steps in sequence.
The hotel environment adds complications that are difficult to reproduce in a laboratory. Robots must navigate narrow corridors, push loaded carts and handle soft items such as towels and bedding that can change shape.
“Judges will deliver comprehensive evaluations based on the number of successfully finished tasks, any collisions during operation and the smoothness of robot movements,” said Li Ziye, the hotel contest's person in charge, according to Bastillepost.
Finishing quickly is not enough. Neatness and errors affect scores, while autonomous operation receives twice the scoring weight of remote control. A remotely operated run is weighted at 0.5 compared with 1 for full autonomy.
From hotels to homes
The same approach is appearing elsewhere. In a working library, robots have 30 minutes to collect returned books, shelve them and fix misplaced volumes, testing precision, visual recognition and reasoning.
The household event, being held for the first time, goes even further. Robots must tidy a living room, fold clothes, wash and hang laundry, then interrupt their chores when they receive a voice command to bring a parcel inside before continuing.
The household competition carries 300 possible points, with laundry receiving the largest share at 125 points, followed by bedroom tasks at 100 and living-room tidying at 75.
The commercial test
The competitions are effectively forcing humanoid makers to confront problems that controlled demonstrations can hide.
Hotels, libraries and homes contain narrow spaces, irregular objects and long chains of tasks. A robot that performs one movement reliably may still stumble when it must navigate to another room, identify the right object, manipulate it and continue working without assistance.
That makes autonomy one of the most important measures emerging from the games. Teleoperation can make a robot look capable, but it also keeps a person in the loop. For businesses evaluating humanoids, that distinction could determine whether the technology actually reduces labor requirements or simply shifts workers from performing tasks to supervising machines.
The WHRG includes six real-world scenario events and eight dexterous-hand contests. The broader competition has attracted 666 teams and 2,056 robots from 16 countries, eWeek previously examined.
Why this matters for businesses
For companies watching the humanoid robotics race, the value of these competitions lies less in spectacle and more in reliability. Hotels, warehouses, retailers, and other labor-intensive businesses need robots that can handle repetitive work in unpredictable environments without constant human intervention.
That is why autonomy, consistency and recovery from mistakes matter as much as speed. A robot that can make a bed once in a controlled demo is interesting. A robot that can navigate a hallway, restock a room, handle soft objects and complete the job repeatedly with minimal supervision starts to look more like a deployable workforce tool.
For business and technology leaders, the next phase of humanoid robotics will likely be judged on practical measures: how often the machines succeed, how much oversight they require, and whether their productivity justifies the cost of deployment. The WHRG offers an early glimpse at which platforms may be moving closer to that threshold.
Related reading: For another look at how humanoids are being pushed beyond controlled demos, see how Figure 03 completed an autonomous ladder climb in one of its latest real-world tests.