AsianFin -- As 2025 kicks off, the hottest topic in the tech world is undoubtedly humanoid robots.

Wang Xingxing, the founder of Unitree Robotics, was seated in the middle of the front row at a prestigious symposium attended by Chinese President Xi Jinping, even overshadowing DeepSeek’s founder Liang Wenfeng in terms of seating.

The Chinese public is buzzing about the humanoid robots performing traditional Chinese dances at the Spring Festival Gala held on January 19, eagerly anticipating the day when robots will enter households.

Chinese entrepreneurs are seeing the humanoid robots market as the next trillion-dollar one after the markets of smartphones and new energy vehicles.

China’s A-shares market hasn’t missed this explosive hype. Leading stock Changsheng Bearings has seen a tenfold increase in just six months, with numerous stocks doubling in value. A few sectors are on the rise, with every machinery and auto parts factory jumping on the bandwagon to boost their valuations.

A sharp single-day drop at the end of February temporarily cooled down the stock market frenzy. No matter how promising the future looks, it’s unrealistic to exhaust the market potential so quickly. Of course, short-term fluctuations don’t alter the long-term trend—humanoid robots still have a boundless future ahead.

This trillion-dollar incremental market has already stirred up excitement among all of China’s tech and manufacturing companies. After all, China has repeatedly proven its strong competitive edge in large-scale, high-end manufacturing. In the new energy vehicle sector, China has risen to the top, maintaining the world’s highest sales for ten consecutive years. In the drone sector, China leads the pack, accounting for over 70% of global sales.

Currently, the commercialization of humanoid robots is in its infancy, with everyone starting from a similar baseline—much like the early days of the drone industry. So, in this booming humanoid robot race, can China replicate its leading position in the drone sector?

“A Troop of Monkeys” Climbing the Tech Tree

The company that first brought robots into the public eye was Boston Dynamics. Back in the day, its Spot robot dog navigating rugged terrain and the Atlas robot performing backflips were viral internet sensations. Its groundbreaking bionic gait design was a milestone, even more impressive than Unitree’s robots performing traditional Chinese dances at the 2025 Spring Festival Gala.

But by 2025, the “center stage” has shifted. In the quadruped robot sector, Boston Dynamics’ Spot sold only 2,000 units in 2024, while Unitree Robotics holds nearly 70% of the global market share. In the humanoid robot sector, Boston Dynamics has yet to launch a mass-produced product, while companies like UBTECH and Unitree have already implemented their robots in education and logistics scenarios.

Why did Boston Dynamics "get an early start but arrive late to the party"? There are many reasons, and one of them is that the company took a wrong turn in its technological approach.

In an exclusive interview with AsianFin, Wang, who is the CEO of Unitree Robotics, stated, "As early as 2013, I had already concluded that hydraulic drives were not a commercially viable solution. The reason is simple: they rely on precision mechanical components, and whenever precision mechanical components are involved, costs cannot be reduced. Moreover, all hydraulic systems leak oil. That’s why you’ll notice that even household cars rarely use hydraulic systems anymore—they’ve all been replaced by electric drives."

In essence, Boston Dynamics chose the wrong technological path. By relying on high-performance yet high-cost hydraulic drive technology in its early stages and failing to pivot to electric drive solutions in time, the company made its products difficult to scale. In 2024, Boston Dynamics announced that its hydraulic-powered humanoid robots would be retired, and future developments would focus on electric-powered products.

Boston Dynamics’ misstep is not uncommon in the history of the tech industry. Giants like Kodak and Motorola once dominated their respective fields but made critical errors when new technological paradigms emerged, leaving them unable to recover. The reason? They had invested too heavily in their old paths, making it exceedingly difficult to change course.

From a broader perspective, Unitree’s success may involve an element of luck, but the success of Chinese robotics companies carries a certain inevitability.

In the U.S., the main players in the humanoid robotics field include Boston Dynamics, Tesla, Figure AI, and a few others like Agility Robotics and Apptronik.

In contrast, China boasts a diverse array of companies focused on robotics, including Unitree, Zhiyuan, Fourier, Leju, Stardust Epoch, General Galaxy, Pansini, and CloudMinds;

Automotive companies venturing into robotics, such as Xiaomi, Xpeng, NIO, BYD, and GAC Group;

Manufacturing giants crossing into robotics, like Midea and Foxconn;

And tech companies entering the robotics space, including Huawei, Alibaba, Tencent, Baidu, Horizon Robotics, and DeepSeek.

In the U.S., manufacturing—an asset-heavy, high-debt, and low-margin industry—does not attract much interest from Wall Street’s financial capital. Entrepreneurs in the tech sector also tend to focus on original, groundbreaking innovations (going from 0 to 1) rather than incremental, iterative improvements (going from 1 to N).

In China, however, the manufacturing sector is comprehensive, fiercely competitive, and constantly seeking to release and export its production capacity. Any promising technological path will find numerous companies willing to explore it.

Years ago, an IT media outlet summarized Huawei's strategy for defeating multinational telecom giants, calling it the "victory of the coyote over the lion." The coyote employed a pack strategy, using a 100-to-1 force ratio to nibble away at the lion's peripheral battlegrounds, gradually closing in on its core.

Today, China's humanoid robotics industry has become a formidable force. On the technological tree of humanoid robotics, it is now swarming with "Chinese monkeys," a phenomenon that could be described as a "troop of monkeys strategy."

In reality, it's hard to determine in advance whether a technological path is right or wrong. Choices that are later deemed correct are often the result of extensive trial and error and market selection. The larger the market scale, the more players there are, and the higher the tolerance for errors.

For instance, in the current debate over technical approaches in robotics, there are two main camps: the "small brain motor control" faction, represented by companies like Unitree and ZongQing, and the "big brain modeling" faction, represented by firms like Fourier Intelligence and Universal Robotics. Regardless of which camp ultimately prevails, Chinese companies are bound to be among the winners.

A Robust Supply Chain for Drones, Electric Vehicles, and Humanoid Robots

When analyzing Boston Dynamics' choice of technical routes, there’s another angle worth considering: supply chain constraints.

Lacking supply chain support, Boston Dynamics' robot dog Spot and humanoid robot Atlas can only target the high-ticket, premium market, making large-scale production unattainable.

Similarly, due to the absence of supply chain backing, Boston Dynamics' shift to an electric-powered technology route has proven far more challenging than anticipated. Without component suppliers, designing and manufacturing everything in-house takes too long; fully self-developing and self-producing components is prohibitively expensive.

The U.S. supply chain has long been heavily reliant on globally distributed, high-precision components. For example, harmonic reducers are imported from Japan, servo motors from Germany, with import costs accounting for over 60% of the total. This reflects the hollowing out of American manufacturing, leaving its ecosystem less rich and complete.

In terms of production difficulty and supply chain complexity, robots fall somewhere between electric vehicles and smartphones. Smartphones require hundreds of components, electric vehicles tens of thousands, and robots thousands.

When a product requires thousands of parts, any single component can become a bottleneck. The compatibility between different components also affects the choice of parts. As is well known, these are precisely the areas where Chinese manufacturing excels.

China's humanoid robotics supply chain has partially benefited from the growth and development of the supply chain that began with smartphones in China.

In earlier years, Chinese companies participated in Apple's supply chain, mastering "intermediate technologies" in consumer electronics. Some of these technologies evolved into "sub-technologies" for electric vehicles at the right time, subsequently serving the drone industry. Now, it’s the robotics industry's turn to reap the benefits.

In principle, drones, autonomous driving, and humanoid robots all follow the "perception—decision—action" model, and can even be considered as part of the same industry.

Some envision a future world with "a robot in the sky, a robot on the ground, and a robot at home."

From an industrial perspective, this can be divided into sensors (vision, hearing, touch)—the brain (chip, operating system, AI large model)—and the power transmission control system.

The supporting industrial chains for drones, electric vehicles, and humanoid robots therefore have significant overlap. On one hand, existing components are reused; on the other hand, talent, experience, and craftsmanship are leveraged to design more suitable hardware.

The leading company in the current stock market speculation on humanoid robots is Changsheng Bearings, which, as the name suggests, is a mechanical factory. Currently, it is a bearing supplier for Unitree, but its main clients are still Caterpillar, Bosch, Sany Heavy Industry, and SAIC.

Reducers, motor servos, batteries, LiDAR, cameras... these are all common to drones, electric vehicles, and robots, with leading companies in these segments supplying all three industries simultaneously.

The extreme cost efficiency of the supply chain is a notable advantage of Chinese manufacturing. For instance, the cost of Green Harmonic's reducers is only one-third of international competitors, CATL's solid-state batteries support 18 hours of endurance, and the localization rate exceeds 90%. The price of Unitree Technology's G1 robot has dropped to 99,000 yuan, initially meeting commercial viability.

Looking back, Tesla's decision to build a factory in China not only provided a significant boost to China's new energy vehicle industry but also benefited Tesla greatly. Without China's supply chain, it would be hard to say whether Tesla could have avoided being overtaken.

The Easily Underestimated "Lego-like" Adaptability

The United States has identified the key reason for gradually losing its edge in the drone sector as the hollowing out of its manufacturing industry. American companies either become obsessed with making groundbreaking technological advances or focus solely on mature technologies that can quickly generate profits, neglecting the accumulation of practical production experience.

It's like being able to draw blueprints but not knowing how to build a house—no matter how good the technology is in the lab, if it can't be mass-produced in factories and continuously improve processes, it ultimately won't become a significant force.

The transition from a technology prototype in the lab to a marketable product involves an "intermediate technology" phase. This is akin to the difference between a recipe and actual cooking—you must repeatedly test (manufacturing process), combine various ingredients (sub-technologies), and adjust the heat (verification and optimization) to transform a theoretical delicacy into a real gourmet dish. Many companies that initially invent technology often lose out to those who continuously improve the technology during mass production.

This technology transformation work carries moderate risk, low profit, and long cycles, much like scaffolding on a construction site—everyone acknowledges its importance, but no one is willing to invest in it long-term. As a result, when American companies outsourced these steps, it was like removing the stairs, leaving them unable to control key technologies and hindering the growth of new industries.

DJI drones can iterate rapidly in Shenzhen because they discovered loose screws by the time they produced the 10,000th unit and improved the design by the 100,000th unit. In contrast, American counterparts, no matter how advanced their designs, either fail to produce prototypes or face skyrocketing costs, ultimately being defeated by the "mass production upgrade" model of Chinese companies.

This is evident in our vast auto parts industry, which can quickly respond and produce all the hardware components needed for robots. It may even foster innovations that other supply chains cannot support.

Some say that Chinese manufacturing has entered a "Lego-style" adaptability phase. When you want to create something or achieve a specific function, you can break down the goal into individual parts, and find qualified suppliers for each component, much like assembling Lego pieces.

This ability to blend and integrate supply chains is precisely the advantage of Chinese manufacturing. Numerous experienced engineers provide timely feedback, allowing for small-scale trial production and rapid scaling; technologies and processes from similar or adjacent fields can quickly be transferred to product design improvements.

This advantage is already evident in drones and electric vehicles. "Silicon Valley prophet" Marc Andreessen believes that drones are the first step for China to surpass the U.S., electric vehicles are the second, and robots will be the third. The American industry believes that drones, intelligent driving cars, and humanoid robots will become the "new trio" of exports from China in the smart era.

Currently, the topic of robots is hot, the market is enthusiastic, and the outlook is promising, but caution is still needed.

Despite the supply chain advantages, with China leading the robot industry's "body" and holding a 73% share of the global humanoid robot supply chain, the "brain" is still dominated by Europe and the U.S. The U.S. leads in AI models (such as OpenAI, Figure) and chips (NVIDIA GR00T).

In the aforementioned conversation between Wang and AsianFin he was candid: "I believe that by the end of 2025, at least one company globally will have developed a relatively general-purpose large-scale robotic model. Of course, our company hopes to be the one to achieve this, but realistically speaking, the probability is higher for a company in the United States to succeed."

The year 2025 marks the early stages of mass production for humanoid robots. A recent Goldman Sachs research report pointed out that the technological inflection point for humanoid robots remains unclear, and there is still a long way to go before they can truly "take on jobs." In response, the stock market has declined, signaling that the industry is about to enter a phase of "separating the wheat from the chaff."

In the long run, as the industry continues to heat up, there is a pressing question for all stakeholders to ponder and prepare for: after the initial frenzy, will it repeat the trajectory of the photovoltaic and new energy vehicle sectors, leading to excessive internal competition? This is a matter that requires calm reflection and proactive planning.