China - US: The battle for leadership in humanoid robots

At the Consumer Electronics Show (CES) 2025 in Las Vegas (USA) on January 7, Jensen Huang, CEO of semiconductor chip manufacturer Nvidia (USA), gave an important speech, declaring that the robotics industry is approaching a historic turning point.

Standing beside him were 14 humanoid robots, representing the most advanced achievements in the field, waving to the audience in the bright spotlight.

Just a few weeks later, at the Spring Festival Gala, China's Lunar New Year celebration watched by more than 1 billion people, the Chinese company Unitree Robotics performed in a different way.

Their H1 humanoid robots, billed as the “first human-compatible robot,” performed a perfectly synchronized folk dance, thanks to cloud computing and advanced AI motion control technology.

These two milestones set the tone for 2025, a year that will see fierce competition among robot manufacturers around the world. Companies from many countries are accelerating mass production and commercialization, with ambitions to gain market share in the booming robotics industry.

While no country has an absolute advantage, China and the US are clearly leading the race. Of the 14 robots on display at CES, six are from China and four are from the US, highlighting the fierce competition between the two powers.

China has the advantage of scale, speed and a mature supply chain, allowing it to produce faster and cheaper. Meanwhile, the US has a clear advantage in technological innovation, with the potential to create new tech giants, similar to Tesla or OpenAI, that can reshape entire industries.

The Future of Robotics: Who Will Prevail?

The year 2025 is predicted to be a turning point for the robotics industry, where leaders in technological innovation and mass production will determine the global market landscape.

China and the US both have their own strategies, but which country will come out on top? The answer will gradually be revealed as robots become an increasingly important part of human life.

Xu Xuecheng, chief scientist at the Zhejiang Humanoid Robot Innovation Center (China), commented: "In the US, the goal is to maintain a technological leadership position, exploring unsolved technical challenges. Meanwhile, China focuses on optimizing and integrating existing technologies to apply them in practice most effectively."

He said Chinese companies prioritize developing humanoid robots that can operate in real-world environments, while American companies focus on building general artificial intelligence, sometimes far beyond what the current industry actually needs.

Just before Unitree Robotics launched its new product line, the tech industry continued to be abuzz with the event of DeepSeek, a Chinese AI startup, shocking the tech world with the introduction of the large language model R1. This model achieved performance on par with OpenAI's advanced products but at a fraction of the development and operating costs.

The launch of DeepSeek’s AI model has shaken the long-held view that training advanced AI systems requires massive capital. The event has triggered a wave of sell-offs in competitors’ stocks and forced investors to reassess their investment strategies in the AI ​​field.

“DeepSeek’s algorithmic innovations are a stark reminder that the technological competition between China and the US is becoming more intense than ever. Our technological advantage is not self-evident, and the AI ​​industry needs to constantly optimize to become more effective,” Eric Schmidt, former CEO of Google, said in an interview with the Financial Times.

The US still holds significant advantages in the robotics race. Its universities are constantly leading the way with new breakthroughs that open up new horizons in research, while US companies are pioneering the development of the “soft” technologies that underpin the industry.

Each humanoid robot combines a highly advanced “brain” that includes AI, machine learning, and sensors to perceive and respond to its environment, along with a sophisticated “body” made of lightweight materials that incorporates flexible joints for natural movement. This tightly integrated system enables the robot to mimic gestures, speech, and interact with its environment flexibly.

During his CES keynote, Nvidia CEO Jensen Huang announced Cosmos, an advanced AI platform that is expected to power the next generation of humanoid robots.

According to a report by Goldman Sachs Research, a research division of the international investment bank Goldman Sachs (USA), the hardware for humanoid robots is almost complete, with important components such as cameras, motors, force sensors, actuators and batteries ready for commercialization.

The bank predicts that with the development of AI, robotics will improve more quickly, especially in capabilities such as manipulation and interaction.

Many companies aim to mass produce humanoid robots by 2025, marking a key milestone in optimizing costs and scaling up production.

Meanwhile, China continues to consolidate its position as a manufacturing powerhouse, not only in traditional sectors but also in robotics. With tens of thousands of companies involved in the supply chain, from component manufacturing, assembly, to prototype design, China is gradually expanding its “world factory” into the field of robotics.

According to the International Federation of Robotics, more than half of all robots installed globally by 2023 will have been deployed in China, giving domestic companies a significant cost advantage as they expand into the humanoid robot market.

The race to cut prices and gain market share

“If companies cannot keep prices below 200,000 yuan (about $27,825) this year, they will have difficulty promoting their products,” said He Liang, a professor at Northwestern Polytechnical University in Xi’an.

He is also the head of a global humanoid robotics company and said businesses are operating on razor-thin profit margins, forcing them to quickly gain market share to survive.

“In fact, we have seen a price war going on in China,” he added.

The battle became clear when Unitree Robotics, a Hangzhou-based company, announced plans to mass produce the G1 model at a price of just 99,000 yuan (about $12,520), a shocking figure compared to the standard price of 500,000 yuan (about $69,560) for its competitors.

Founder Wang Xingxing revealed that prices will drop further as production scales up, with ambitions to ship hundreds of thousands of robots a year.

Meanwhile, Shanghai-based startup Agibot announced in December that it would begin mass production of its multi-purpose robots. To date, the company has built a total of 962 wheeled bipedal robots, of which more than 200 are used internally and nearly 700 have been shipped, mainly for service interaction and manufacturing.

Not only are Chinese companies rushing to expand production, but the race is also fierce in the US. Tesla CEO, billionaire Elon Musk, aims to produce thousands of multi-purpose Optimus robots in 2025, before increasing production to millions of robots per year.

He predicts that production costs could fall below $20,000, but the final selling price will depend on market demand.

In addition, Agility Robotics, another leading US company, has raised a $150 million investment from Amazon and built the RoboFab factory in Oregon (USA), with a production capacity of tens of thousands of Digit robots per year.

In an increasingly competitive environment, speed and rapid deployment are becoming the key to success. Professor He Liang said the focus is no longer on overcoming individual technical challenges but on optimizing the entire system.

“If you are behind in 30 of 100 technical areas, it doesn’t matter, as long as you are ahead in the remaining 70,” he stressed. “Right now, I don’t see any challenge that is insurmountable. Ultimately, in this game, the fast fish will eat the slow fish.”

Humanoid Robot Race: China andAmerica shapes the future of technology

Speed ​​of implementation is key, and the difference between US and Chinese companies is evident in production scale, financial resources and talent.

Tesla, with its powerful computing platform and deep pockets, has the ability to shorten the research and development (R&D) phase and quickly bring products to market. It can sell millions of robots a year, while Chinese companies have the advantage of significantly lower production costs.

However, with its huge number of enterprises, China can generate cumulative output equal to, or even surpass, its US rivals.

“This is the problem,” He Liang said. “The more you sell, the more data you collect from real-life customer situations. When combined with a powerful computing platform, this data makes robots smarter.”

Over the long term, this loop of improvement can widen the gap between leaders and laggards, especially as the industry matures.

Wang Lei, chairman of Shanghai Qingbao Engine Robot, a company that has sold dozens of robots by 2024, warned that Chinese businesses should not follow Tesla's path.

“Once Tesla has paved the way in a particular field, we should not copy them,” he stressed. “They are flying a rocket, and we are flying an airplane. If we try to catch up, the gap will only widen.”

Unlike Tesla, which already has a strong foothold in the electric vehicle industry, Chinese companies are looking to explore untapped niche markets.

“If we just chase Tesla, we will always be the follower,” Wang said, emphasizing the importance of creating unique segments to maintain a competitive advantage.

At the same time, as the technology race between China and the US heats up, Washington is tightening control over strategic technologies, especially AI chips, adding pressure to Beijing's high-tech ambitions and raising questions about the future of humanoid robotics.

Last month, former US President Joe Biden announced new regulations to restrict AI chip exports, targeting China, Russia, Iran and North Korea.

The measures, enacted just before Biden leaves office on January 20, would completely cut off exports of this technology to four geopolitical rivals, while ensuring close US allies have unrestricted access to critical components.

However, Lu Hancheng, director of the Gaogong Robot Industry Research Institute in Shenzhen, said that China's humanoid robot companies have not been greatly affected for now, as mainstream chips are still sufficient for training needs. However, this could change as computing needs increase exponentially.

“When the industry shifts from being data-centric to requiring superior computing power, that's when the real challenge begins,” Lu said.

Yet, amid geopolitical tensions and fierce competition, there is one bright spot: academic cooperation between the two countries has remained largely uninterrupted.

"In the exploration stage, China and the US tend to jointly expand the boundaries of artificial intelligence, rather than compete individually," said Xu Xuecheng, an expert at the Zhejiang Humanoid Robot Innovation Center.

He also pointed out that cooperation projects between the two countries take place regularly, especially in the fields of developing robotic intelligence and brain model design.

In addition, the world's top robotics conferences still record the participation of scientists from both powers, with many research papers co-authored by experts from both countries.