The humanoid robot revolution is no longer a futuristic promise: it is now on display to the general public at the Mobile World Congress in Barcelona. Among them, the leader of a still emerging market: AgiBot.
Founded in 2023, the Chinese start-up has gained a head start in terms of sales over its rivals, whether Chinese – Unitree, Fourier Intelligence or Magic Lab – or American, such as Tesla, Boston Dynamics and Agility Robotics. In 2025, it claims to have produced nearly 5,000 robots, or more than a third of the global market. Building on this dynamic, the company intends to accelerate its international expansion, with Europe as a strategic priority. William Shi, former president of Huawei France, joined AgiBot in January with the mission of conquering the European market.
JDN. Can you introduce us to AgiBot?
William Shi. AgiBot is a Chinese start-up created three years ago. We focus on the research and development of humanoid robots. Today we already have three types of products. The first is called Lumi, it’s the X2 series. It weighs approximately 35 kg. Her main ability is to dance and easily learn choreography for shows and events.
The second model is our A2, a “full size” robot, the size of a human. It is designed for interaction with people. It can present products in stores, or direct visitors in airports or large shopping centers. Finally, we have the G2, which is capable of working. Today, he carries out simple tasks in the factory. Our three products are already sold in China, and are ready to be exported to Europe.
Where is AgiBot today in terms of global production and what is your vision for market development in the medium term?
According to our understanding of the market and several industry reports, global production last year was around 13,000 units. In 2025, our production exceeded 5,000 units, approximately 40% of the overall market. We are therefore ranked number one in the world by volume, even if that doesn’t mean much at the moment. In the next ten years, global production could reach two or three million units.
Today, 10,000 or 20,000 units is very little. The most important thing is to find the right ecosystems and work together. It’s a bit like the automobile industry: no one does everything alone. Some make the engines, others the design, others the paint. Everyone brings their own technology. This is why we are looking for partners in Europe.
What differentiates you from your competitors, whether Chinese or American?
“We are also developing a smaller model, around 80 centimeters”
When we say we are number one, it is only in terms of production volume last year. This does not guarantee anything for the next three or five years. Today, humanoid robots face two major challenges: production and artificial intelligence. For this new generation of robots, immense volumes of data are required. We do not yet have enough data to rapidly evolve AI. It’s the same challenge for Tesla, Figure AI or other players. Everyone faces this problem.
There is strong competition between China and the United States in advanced technologies, but I think it is positive for consumers. Without competition, prices remain high. With competition, prices fall and technology advances faster. That’s a good thing.
Will robots soon be accessible to the general public?
Today the price is still high. A robot costs almost the price of a car. The cheapest model we’ve sold is around $20,000. Research and development costs are significant, and production automation is not yet sufficient. But over time, prices will drop.
We are also developing a smaller model, around 80 centimeters, which will be much more affordable and more accessible to consumers. It could be ready in a few months.
Why did AgiBot choose to focus on humanoid robots?
When a robot looks like a human, it can use the tools designed for humans. He can climb stairs, use scissors, a pen, manipulate standard objects. This facilitates its integration into our environment. We recently demonstrated a robot that can write calligraphy with a pen. In summary, a humanoid robot can adapt more easily to the human world.
But not all of our robots are entirely inspired by human morphology: the industrial model has no legs. Because to work in a factory, he needs stability. And with two legs, walking can be less stable.
What is your strategy for the European market?
The European market is strategic for us. It is still in its infancy when it comes to humanoid robots and does not, at this stage, have any large, dominant players. This is precisely why we are coming in with our technology and looking to build local partnerships. A robot cannot exist alone. It can be compared to a chatbot, but with a physical body. It combines big data, artificial intelligence and robotics. For example, for large language models (LLM) technologies, we need partners. Today, our robots speak Chinese and English. But for French, German, Spanish or Italian, we must work with local partners.
Furthermore, we work with cloud solutions in Europe, such as Alibaba, Amazon or Microsoft, which have the necessary big data capabilities. For example, in China, the robot does not speak French. But in Europe, thanks to local infrastructure, he can start speaking the language. To evolve towards industrial uses, we must also collaborate with companies that understand the specific needs of European sectors. This is essential for future development. We are therefore looking for partners. If the robot is to work in different sectors, it is necessary to understand the needs of each industry. This is why we came to MWC: to meet partners.
How long do you think it will be before we see robots becoming an integral part of our daily lives?
Progress is very rapid. Last year, this time, our robots only knew how to walk. Today they dance and perform complex movements. Next year there will still be a lot of progress. We continually train our robots with data and manipulations. They learn very quickly. In three to five years, I think families will be able to find value in these robots, whether for interaction or some simple tasks.
