The stage of China’s 2026 Spring Festival Gala became a grand showcase for humanoid robots, which performed a series of high-difficulty moves including consecutive backflips, nunchaku displays and skit acting with a “zero-error” performance, once again bringing the development of embodied intelligence into the spotlight. The Gala’s viral effect has driven a sharp surge in market enthusiasm for related products: search volumes for robot models have skyrocketed by over 300%, many products sold out within minutes of their launch, and even high-end models priced at nearly 630,000 yuan were snapped up. However, the stunning performances on stage cannot mask the practical predicaments in the commercialization of the humanoid robot industry. The sector shows a stark divergence in development between the B2B industrial market and the B2C consumer market, meaning humanoid robots still have a long way to go to move from the stage to real life.
China’s 2026 Spring Festival Gala featured products from four leading manufacturers including Unitree, Magic Atom and Galaxy General Robotics, with robots delivering performances across diverse scenarios such as martial arts, crosstalk and musical ensemble. Yet their flawless stage acts stand in stark contrast to their real-world applications. A netizen observed that robots in practical settings like welfare homes can only perform simple tasks such as slow walking and direction guiding, with severely limited flexibility and intelligence. This has led many to question the possibility of humanoid robots entering households in the short term, and some even argue that many current companion and service robots lack core value and are merely a haphazard stack of simple functions.
The core divergence in the robot industry’s development lies in the large-scale advancement of the B2B industrial market versus the sluggish growth of the B2C consumer market. Specifically, according to a report released by the International Federation of Robotics (IFR), the global new installations of industrial robots had already reached 542,000 units in 2024, exceeding 500,000 units for four consecutive years, and the figure is expected to rise further in 2025. In contrast, data from IDC shows that the global shipment volume of humanoid robots will only reach about 18,000 units in 2025, and not all of these are for ordinary consumer use. Industrial robots have become an important productivity tool in enterprise production thanks to standardized operational scenarios, while consumer-grade humanoid robots are mired in a predicament of “pseudo-intelligence” — all their movements are pre-programmed and trained, lacking the ability to make independent decisions in real scenarios, thus failing to meet the complex demands of consumer scenarios such as households.
Multiple factors contribute to this developmental gap. In terms of pricing, consumer-grade robots cost tens of thousands or even hundreds of thousands of yuan, far exceeding the affordability of ordinary consumers, and market demand for such products is more driven by emotional consumption than rigid practical needs. In terms of scenario requirements, industrial scenarios only demand robots to perform efficient, fixed repetitive tasks, while consumer scenarios require robots to not only execute movements skillfully but also possess empathy and independent judgment — capabilities that current hardware performance and algorithm models are unable to deliver. As for industrial chain supporting facilities, the consumer-grade robot industry has not yet established a mature after-sales service system; problems such as difficulty in repairing components and the lack of offline service outlets have further hindered its popularization. According to latest robotic news, the Ministry of Industry and Information Technology has officially set up the National Technical Committee for Standardization of Humanoid Robots and Embodied Intelligence, which will formulate and revise industry standards across various fields, laying a solid foundation for the sector’s development.
Overall, the development of humanoid robots is not a false demand, and their “zero-error” stage performances are a testament to technological progress. Nevertheless, moving from industrial scenarios to millions of households, and from pre-programmed movements to independent decision-making, humanoid robots still need to break through multiple bottlenecks in technology, scenario adaptation and industrial chains. Only by truly solving the problem of “pseudo-intelligence” and aligning with consumers’ real needs can humanoid robots truly step out of the show stage and into real life.
