Decoding the Differentiation Strategies of China’s Humanoid Robot Industry through the Financial Reports of UBTech and Yushutech

Recently, UBTech (09880.HK), recognized as the “first humanoid robot stock” on the Hong Kong Stock Exchange, released its performance report for the fiscal year 2025, drawing significant attention from the capital market. The following trading day saw UBTech’s stock open sharply higher, with an intraday increase nearing 20%, marking the largest single-day gain in nearly a year. Subsequently, Citi significantly raised UBTech’s target price by 23% to HKD 190 and increased its annual revenue forecasts.

Previously, UBTech announced a comprehensive strategic partnership with global industrial automation giant Siemens. The collaboration aims to achieve full-chain synergy in the integration of industrial digitalization, smart manufacturing, and humanoid robots, jointly developing global industrial humanoid robot solutions.

While many companies are still in the stage of technological validation and scenario exploration, UBTech has experienced a notable increase in business scale. According to the financial report, UBTech achieved revenue of CNY 2 billion in 2025, a year-on-year increase of 53.3%, with overall gross profit reaching CNY 750 million, up 101.5%. The full-sized embodied intelligent humanoid robot business saw significant growth, generating revenue of CNY 820 million for the year, a more than twenty-fold increase, with sales reaching 1,079 units and a gross profit margin of 54.6%. This segment has become the company’s largest revenue source, accounting for 41.1%.

In terms of delivery scale, UBTech has become the world’s first company to achieve the delivery of over 1,000 industrial humanoid robots. This advancement marks an important milestone in UBTech’s development and reflects a broader shift of humanoid robots from technological validation to application exploration. Recently, UBTech has collaborated with high-end manufacturing giants from Europe and the U.S., including Siemens, Airbus, and Texas Instruments, demonstrating the growing attention Chinese manufacturers are gaining within the international industrial system through their capabilities in product engineering and delivery rhythm.

From an industrial perspective, China has established a relatively complete manufacturing and supply chain system in the humanoid robot sector, possessing certain advantages in engineering capabilities, cost control, and scene advancement rhythms. In this context, different companies have formed differentiated development paths. UBTech primarily focuses on industrial scenarios, driving the practical application of humanoid robots in production processes, while Utree Technology leverages its strengths in motion control and body design to create a universal platform aimed at research, education, and developers.

The two approaches correspond to different stages and focuses of industrial development, collectively creating a diversified exploratory landscape in the current humanoid robot industry.

Two Approaches

In the current development phase of the humanoid robot industry, companies exhibit significant differentiation in terms of technical paths and resource allocation. Taking UBTech and Utree Technology as representatives, these two types of companies are laying out their strategies around different technical focuses and application directions, gradually forming a differentiated product system and business structure.

UBTech continues to increase its R&D investment, focusing on advancing the embodied intelligence technology system. In 2025, the company’s R&D investment reached CNY 507 million, accounting for approximately 25.4% of its revenue, with nearly CNY 1.9 billion invested over the past four years. The R&D team comprises 942 members, with nearly half holding master’s or doctoral degrees.

Building on this foundation, UBTech is developing embodied intelligent large models, visual language action models (VLA), and world models, enhancing the autonomous decision-making and collaborative operational capabilities of humanoid robots in complex scenarios. The company has independently developed the Thinker embodied large model, the Thinker-VLA visual language action model, and the Thinker-WM world model specifically for industrial humanoid robots, establishing a complete technology stack. Their industrial robots have completed upgrades to the Co-Agent industrial collaborative intelligence system, transitioning humanoid robots from individual autonomy to collective intelligence, achieving unified scheduling and collaborative operations.

To date, UBTech has received over 2,900 authorized patents, including more than 1,700 invention patents.

In contrast, Utree Technology adopts a more focused R&D strategy, concentrating primarily on core areas such as motion control, body design, and hardware systems. In the first three quarters of 2025, the company’s R&D expenses were approximately CNY 90.2 million, accounting for about 7.7% of its revenue, with total annual R&D investment estimated between CNY 90 million and CNY 120 million and an R&D team size of around 175 members.

In terms of technical paths, Utree’s products emphasize high-performance motion capabilities and modular design, promoting algorithm iteration and application expansion through an open-source ecosystem and developer community.

From an R&D structure perspective, the two types of companies are accumulating strengths in different dimensions, namely “system capabilities” and “body capabilities.”

Differing technical paths correspond to distinct product forms. UBTech focuses on full-sized industrial humanoid robots, emphasizing autonomous operation and scene adaptability, while Utree Technology primarily offers small to medium-sized humanoid robots, stressing flexibility, cost control, and platform universality.

In terms of applications, UBTech mainly targets industrial scenarios, with related businesses accounting for over 80% of its revenue. Its products have begun exploring applications in manufacturing, logistics, and other fields. In 2025, UBTech’s full-sized embodied intelligent humanoid robots achieved mass production, delivery, and confirmation of over 1,000 units, applied in various industrial areas including automotive manufacturing, smart logistics, 3C electronics manufacturing, semiconductor manufacturing, aerospace, and industrial data collection. On January 18, UBTech signed a service agreement with European aviation giant Airbus, marking the first deployment of humanoid robots in the aviation manufacturing sector. Additionally, the company has partnered with global Fortune 500 companies such as Texas Instruments, FAW-Volkswagen, BYD, and Foxconn to scale up the application of its robots in handling, sorting, and quality inspection processes.

Utree Technology, on the other hand, primarily serves the research, education, and developer markets, with its products widely utilized in universities, research institutions, and various exhibition scenarios. In 2025, its non-full-sized humanoid robots shipped over 5,500 units, with approximately 70% sold to major universities and research institutions globally for algorithm research and teaching demonstrations. The remaining products have found broad applications in cultural tourism performances, live streaming sales, and family entertainment scenarios. Utree’s robots have even graced the stage of the CCTV Spring Festival Gala, enhancing brand recognition through performances such as dance and martial arts.

Overall, the two types of enterprises are driving the development of the humanoid robot industry from different dimensions: one type emphasizes scenario implementation and system integration capabilities, while the other promotes technological diffusion and ecosystem construction.

What Type of Humanoid Robot Does the Industry Need?

By early 2026, operational data disclosed by multiple humanoid robot companies reflected that the industry is advancing its commercialization exploration along different paths. In the early stages, products with strong motion capabilities and impressive demonstrations help raise public awareness and market attention while providing a foundational platform for researchers and developers to drive technology iteration and application validation.

Data shows that by 2025, the number of humanoid robot manufacturers in China exceeded 140, with over 330 products released, and the total industry financing reached USD 2.65 billion, surpassing the sum of previous years, accelerating the influx of capital and industrial resources.

As the industry progresses, the focus of attention is also shifting. Rather than merely asking, “Can complex actions be completed?” the market is increasingly interested in the robots’ sustained operational capabilities in specific scenarios, or “Can they reliably complete tasks?”

According to IDC, the competitive focus of the industry is shifting from hardware performance to technical depth, service capabilities, and ecosystem construction. By the end of 2025, the Ministry of Industry and Information Technology will establish a standardization technical committee for humanoid robots and embodied intelligence, releasing China’s first comprehensive standard system covering the entire industry chain and lifecycle for humanoid robots and embodied intelligence (2026 edition), marking the industry’s entry into a new phase of standardized and systematic development. In this phase, the depth of technology, the ability to implement in the industry, and the localization of core components will be key to determining who can complete the entire journey.

As a property of productivity, humanoid robots are increasingly becoming a core issue in the technological competition among major nations. IDC data indicates that by 2025, global humanoid robot shipments are expected to reach approximately 18,000 units, a 508% year-on-year increase, with Chinese manufacturers dominating the market. Research comparing the innovation capabilities of Chinese and American humanoid robot companies across four dimensions—product development, technological innovation, financing, and market applications—has found that Chinese companies excel in the breadth of product application scenarios, while the U.S. holds an advantage in the breadth of patent distribution.

This structure suggests that the current industry is still in a parallel phase of “application promotion” and “technological accumulation.” Overemphasis on front-end demonstrations while neglecting foundational capability building could constrain long-term competitiveness.

In this context, different types of companies play distinct roles in the industry: one type promotes technological diffusion and application exploration through general platforms and developer ecosystems, while the other focuses on the practical application of robots in production systems around specific scenarios.

Thus, Chinese humanoid robots can genuinely lead change in the new round of global industrial revolution, transforming mass production advantages into technological advantages, scale advantages into standard advantages, and scenario advantages into ecosystem advantages, ultimately gaining a competitive edge in the global technological landscape.

The Year of Mass Production for Embodied Intelligence

2025 is globally recognized in the tech community as the “year of mass production for humanoid robots.” According to the latest data from international research firm Omdia, global humanoid robot shipments in that year are estimated to be around 13,000 units, with Chinese companies capturing 90% of the market share and taking the top six spots in the shipment rankings. This figure challenges the previous perception of “U.S. technological leadership with China following and imitating,” as both countries have embarked on entirely different development paths in embodied intelligence.

The U.S. camp, represented by Tesla’s Optimus and Figure AI, tends to focus more on general capabilities and cutting-edge technological exploration in the humanoid robot domain. However, related products are still primarily in the testing and early validation phases. Tesla’s latest disclosure indicates that Optimus’ production in 2025 is expected to be around 150 units for internal testing, significantly below Elon Musk’s earlier target of 5,000 units. Although Figure AI has attracted investments from tech giants like Microsoft, NVIDIA, and Amazon, its current delivery scale remains limited, with fewer than 200 units delivered throughout the year, primarily targeting laboratories and early customers.

Overall, while U.S. companies have accumulated advantages in foundational technologies such as algorithms and large models, they still face challenges in aspects like industrial chain compatibility, cost control, and scenario implementation.

U.S. companies may excel in foundational algorithms and large model technologies, but they also have clear weaknesses: their industrial chains are incomplete, and they heavily rely on external procurement for core components, leading to high production costs. Currently, the average price of similar humanoid robots in the U.S. is around USD 300,000, which is 5-10 times higher than that of Chinese industrial products, making large-scale commercial deployment challenging. More critically, U.S. companies generally lack experience in manufacturing scenarios, with their products remaining in a stage of “being able to move” rather than achieving true industrial application.

In contrast, Chinese companies tend to focus more on specific application scenarios, driving technology validation and iteration in real-world environments. In this process, a number of representative companies have emerged domestically. For instance, UBTech concentrates on industrial scenarios and advances the application of humanoid robots within production systems, while Utree Technology builds a universal platform for research, education, and developers, leveraging its strengths in motion control and body design. These two types of companies explore different dimensions, corresponding to differing emphases on scenario implementation and technology diffusion, collectively fostering the growth of the humanoid robot industry.

China’s leading advantages stem from three irreplicable foundations: first, it possesses the world’s most complete industrial chain support system. China is the only country with all industrial categories classified by the United Nations, achieving full-chain coverage in humanoid robotics, from reducers and servo motors to sensor production and complete assembly, with over 80% localization of core components. Cities like Shenzhen and Dongguan have formed industrial clusters, where UBTech can assemble a humanoid robot from parts to finished product in about one hour, far quicker than the 15-20 day cycle typical in Europe and the U.S.

Second, Chinese companies exhibit exceptional cost control abilities. Leveraging large-scale production experience gained from the consumer electronics and new energy vehicle industries, they have dramatically reduced the costs of core components for humanoid robots. The price of the new generation of joint modules has dropped from thousands of yuan to hundreds, and the overall machine cost has decreased from millions to hundreds of thousands, laying the groundwork for large-scale commercialization.

Finally, China has the richest resource of application scenarios globally. As the largest manufacturing market, it provides ample testing ground for industrial robots. From automotive manufacturing to 3C electronics and aerospace, Chinese companies can swiftly deploy products in real scenarios for validation and iteration, creating a positive cycle of “scenario-data-technology.”

Industry experts generally believe that the “Scaling Law” in the field of embodied intelligence is becoming evident: R&D demands are rising exponentially, with only leading companies able to bear the high R&D costs and risks of trial and error. In the next 3-5 years, the industry landscape is expected to converge rapidly, concentrating funds, technology, and scenario resources in leading companies, ultimately establishing a stable state of “dual dominance with differentiated competition.”