Views: 0
Humanoid robots have become one of the biggest stars of artificial intelligence. They appear in factories, warehouses, technology fairs, and increasingly spectacular demonstrations. The idea is fascinating: human-shaped machines capable of understanding instructions, moving around our environment, and performing tasks that until now depended on people.
But there's a question worth asking: do we really need robots to be human-shaped?
Economic interest explains a good part of this frenzy. Some estimates place the general-purpose robotics market at around $370 billion by 2040. The figure may be very optimistic, but even if it falls short, we're talking about a huge market with extraordinary growth potential. And when a market of this size emerges, nobody wants to be left out.
The main advantage of the humanoid robot is quite simple: our world is designed for people. Stairs, doors, tools, shelves, vehicles, and workstations have all been designed with our dimensions and movements in mind. A robot with two legs, two arms, and hands could navigate these spaces without requiring any complete modification.
This is where artificial intelligence comes in. So-called physical AI, or embedded artificial intelligence, aims to enable robots to do more than simply repeat programmed movements. Instead, they can perceive their environment, understand instructions, make decisions, and learn.
The theory is incredibly appealing. But it's also worth remembering an old engineering rule: "Keep it simple, stupid." If a task can be performed by a much simpler robot, why use a humanoid one?
.A mobile robot doesn't need to mimic our gait to carry a box. Humanoid robots make sense when we need a high degree of versatility.
Complexity comes at a price. A humanoid must coordinate a vast number of elements: motors, actuators, sensors, joints, balance systems, batteries, cameras, and software. Walking alone is a complex problem. Doing so while manipulating an object, avoiding an obstacle, and responding to a real-time instruction is even more challenging.
However, there is something that could radically change this equation: mass production. The more robots that are manufactured, the lower the cost of their components can be and the more efficient their production will be. And this seems to be what is beginning to happen.
And China has a distinct advantage. The most recent data shows just how far ahead China is going. Smart Analytics Global estimates that around 19,100 humanoid robots were shipped worldwide during the first half of 2026, a 272% increase compared to the previous year. Chinese manufacturers accounted for more than 97% of those shipments.
Two companies stand out in particular: AgiBot and Unitree. Between them, they shipped around 14,300 units during those first six months. AgiBot reached approximately 8,400 units and Unitree around 5,900. In contrast, Western companies manufactured only a few hundred robots.
The Chinese advantage is not solely explained by labor costs. There is a vast local supply chain for motors, batteries, sensors, electronics, actuators, and mechanical components. This allows for cheaper manufacturing, but also for rapid design modifications and scaling of production. And in a technology that is still evolving, that flexibility can be decisive.
But of course, the robotics race also has a geopolitical dimension, so the United States has made its move. In July 2026, the US Federal Communications Commission announced restrictions on new imports of humanoid and quadruped robots manufactured abroad, citing risks related to national security and cybersecurity. The measure particularly affects Chinese manufacturers.
Although this measure is a setback, it will be very difficult to halt Chinese development. China does not depend solely on exports: it has a huge domestic market, specialized manufacturers, private investment, and strong institutional backing. In fact, Unitree is currently preparing to become the first Chinese humanoid robot company to go public, a clear sign of the enormous financial interest this sector is generating.
I suppose the real battle will be for profitability. The future will probably not belong exclusively to humanoid robots. There will be robotic arms, mobile robots, specialized systems, quadruped robots, and, of course, humanoids.
Each solution will have to answer a very simple question: Does it provide more value than it costs? That will be the real test.
The current obsession with humanoid robots may have a trendy element, but it also stems from a very sound technological strategy: to create a machine flexible enough to perform many different tasks without having to redesign the world around it.
We are still in the early stages. Intelligent robotics is entering a phase of industrial production, experimentation, and the emergence of new applications. Perhaps in a few years we will discover that humanoids were the solution for some tasks, but not for all.
And that is precisely what is interesting. The robotic revolution will not necessarily consist of filling factories with human-like robots. It will consist of discovering where an intelligent machine can do things better, faster, cheaper, or more safely.
In the world of robotics, the coming years are undoubtedly going to be very exciting.
My Best Regards and have a good day
EN 
ES