Inside a Shanghai robot "carnival" where ordinary families meet China's humanoids

An 11-year-old boy walked the floor of a public robotics showcase on the outskirts of Shanghai this week holding a robotic dog the size of a hardcover book. He showed it to a cluster of families crowding around him. The dog rolled forward, tilted its head, and made a sound the boy had clearly rehearsed. A staffer nearby declared a water-bead robot duel, two small wheeled machines facing off in a plastic arena, a draw. A pair of industrial arms folded a basket of T-shirts with metronomic precision. A four-legged robot climbed a short flight of stairs.

The showcase took place at an R&D center on Shanghai's outskirts that houses more than 100 companies working on robotic technologies and equipment. It ran during a public holiday, the kind of event a domestic industry uses when it wants the product to feel ordinary rather than announced.

For roughly nine of every ten two-armed, two-legged robots shipped globally last year, the country of origin is China. The MIT Technology Review dispatch from the showcase puts the figure at nearly 90% of more than 13,000 units delivered worldwide in 2025. Forbes's 2026 shipment tracker for top humanoid companies shows a Chinese-dominated top ten, consistent with that share. A domestic industry that already leads on shipments now has a captive audience for the next layer of the product: the household.

"Embodied AI" means artificial intelligence built into physical machines, in this case robots, and it is a named priority in China's latest five-year plan, the country's central industrial-policy blueprint for the next phase of economic development. A MERICS analysis of the sector frames it as one of the most ambitious bets in that plan, with the public sector shaping supply, demand, and the venues where the public meets the technology. A showcase where an 11-year-old can demo a product to other families is not an accident. It is a venue.

The carnival is also a study in how far the technology still has to travel. Most of the products on the floor were toys and arms, not the bipedal humanoids that dominate the strategic narrative. The DexForce humanoid making coffee, slow and careful, its product manager Yang Duan on hand to explain that the machines "can free people from physical work and mundane chores," was one of the few fully bipedal units on display.

Humanoid progress elsewhere has been slow, and the reasons are the ones the industry already names. Safety challenges from heft and bipedal balance. High prices. Short battery life. Plenty of roboticists remain unconvinced that a human shape is the best form factor for a robot at all, that a robot needs two arms and two legs to do useful work. A 2026 MIT Technology Review piece on humanoid training data lays out how much of the bottleneck is data: humanoids learn from human motion capture, and the supply of that data is limited and expensive. A showcase does not solve that.

What the showcase does buy is a different kind of asset: comfort. The 11-year-old with the robotic dog is the product working as intended at the consumer layer. The four-legged robot on the stairs is a halfway step toward a market that does not yet exist. The T-shirt-folding arms and the water-bead duel are spectacles, but they are spectacles with a job. They let families handle, watch, and argue about machines that, until recently, only appeared in corporate launch videos.

The contrast with markets where humanoids remain exotic is the point. In the United States and Europe, a humanoid in a public space is still an event. In Shanghai's outer districts during a public holiday, it is a thing you walk past on the way to the next booth. That ordinariness is what the supply-share number is buying, not just more units shipped, but a population that has already had the first awkward encounter with the form factor.

Whether the showcase model can keep doing that work is the open question. Battery life still caps a humanoid's usefulness outside a controlled demo. Price still puts the most capable units out of reach of the families in the room. The skeptics are right that the form factor may not be the one that finally ships at scale. For now, the strategy is not waiting for those answers. It is putting the product in the hands of an 11-year-old and letting him show the next family what it does.