Recently, humanoid robots have frequently appeared at major exhibitions and events, garnering significant attention with their impressive technological performances. Amidst the excitement, the core question people are concerned about is: when will these rapidly evolving humanoid robots step off the exhibition stage and into real-world scenarios to perform “serious tasks”?

Currently, there is a consensus among many companies regarding the path to deploying humanoid robots, which is to first penetrate industrial scenarios, refine technology and optimize costs while addressing actual production needs, and then gradually extend into commercial services and household applications, ultimately achieving the long-term goal of “robots becoming household assistants”.

Lang Xulin, Chief Scientist at Yujian Technology, pointed out in an exclusive interview with the Securities Times that in the next decade, every household will have a robot, just like every household has a refrigerator and a washing machine.

Commercialization Path Becoming Clearer

After the Beijing Winter Olympics, public attention once again focused on the National Speed Skating Oval, where humanoid robots became the stars of the event. Compared to the humanoid robot half-marathon held four months ago, the robots’ athletic capabilities have rapidly improved.

“During the half-marathon, many robots stumbled while walking, and some even lost their heads and had to be dragged by staff. I didn’t expect the evolution of humanoid robots to progress so quickly in just a few months,” remarked Wen Zhe from Zhengzhou after watching the recent 2025 World Humanoid Robot Games, expressing that the development speed of robots far exceeds expectations.

Competitive events have tested the stability of high-speed movement and adaptability to complex environments, allowing technical teams to validate technology in real-world settings and drive product upgrades. Beyond the competition arena, as robots’ capabilities continue to improve, transitioning from “performing” to “working” remains a challenge that humanoid robot companies need to overcome.

Zhong Xinlong, Director of the Artificial Intelligence Research Center at CCID Research Institute, pointed out to reporters that the success or failure of humanoid robot commercialization depends on the ability to accurately select and penetrate high-value application scenarios that can create quantifiable returns on investment. “Only by anchoring on essential scenarios and addressing actual pain points can we bridge the technological gap and achieve commercial retention and continuous iteration. Just as facial recognition opened the door for artificial intelligence, the robotics industry is undergoing a critical phase from technology validation to value closure,” said Tang Jian, Chief Technology Officer of the Beijing Humanoid Robot Innovation Center.

The strategic path of transitioning from industrial to commercial and household applications is a common choice among many companies promoting the commercial deployment of humanoid robots. Currently, there are numerous cases of humanoid robots being applied in industrial scenarios. Yujian’s embodied intelligent robots have secured orders from international industrial and commercial giants such as Mazak, ASKA, and Aurotek.

The Moja robot, incubated by Chery, has already achieved commercial applications in Indonesia, the UAE, South Africa, and Hong Kong, where these humanoid robots are used for welcoming guests and explaining vehicle models, thereby enhancing store operational efficiency and user experience to some extent.

A representative from Moja robot stated to the Securities Times that after achieving application in 4S scenarios, Moja robots will further enter supermarkets and government service halls.

Generalization Dilemma and Cost Challenges

Although the application of humanoid robots has begun to emerge in certain scenarios, large-scale deployment still faces numerous challenges.

The lack of generalization has caused humanoid robots to be “stuck” in single scenarios. “They can open a can but cannot peel an apple,” Tang Jian pointed out. The general model of humanoid robots is not yet mature, and their ability to generalize tasks across different scenarios is weak, making it difficult for them to respond flexibly to unexpected situations in unstructured environments, such as homes and complex factories. The “ChatGPT moment” for humanoid robots has not yet arrived.

Tang Jian emphasized that breaking through the generalization bottleneck hinges on algorithm innovation, which involves reconstructing algorithmic logic that adapts to the physical world to enhance robots’ ability to integrate into real-world scenarios. This technological breakthrough will also bring new opportunities to the industry; if small teams can develop breakthrough algorithms, they may achieve a leapfrog advantage, pushing the entire industry’s technological level to new heights.

Price is another key variable determining the scope of humanoid robot adoption. Currently, the overall price of humanoid robots generally reaches three to four hundred thousand yuan, which is still far from being widely adopted in households.

“Without mass production, discussions about cost reduction are futile.” Many industry insiders emphasized during interviews that cost control is the core prerequisite for the large-scale deployment of humanoid robots, but the current industry is still caught in the dual dilemma of insufficient mass production scale and the hard costs of core components being difficult to compress.

Zhong Xinlong analyzed that by referencing the cost reduction paradigm of the new energy vehicle industry, humanoid robots can achieve cost reduction within 3-5 years through modular design and localized supply chains.

Yushu Technology CEO Wang Xingxing stated that from a practicality perspective, the simpler the hardware design, the easier it is to significantly reduce design and manufacturing difficulties, thereby lowering costs and improving production efficiency.

More challenging are the non-compressible “hard costs.” Zhong Xinlong pointed out that high-precision harmonic reducers used for the core joints of robots, six-dimensional force torque sensors for precise force control, and high-performance system-level chips like NVIDIA Jetson that support complex AI models for robot operation are the three major “hard costs” for robots.

“These components cannot achieve cost reduction simply by scaling; they must achieve price reductions through breakthroughs in upstream foundational technologies,” admitted a research and development leader from a robotics company.

Fortunately, humanoid robot products on the market have shown a trend of price reduction: in June of this year, Youliqi launched a “Steel Nanny” priced at 88,000 yuan; in July, Yushu Technology reduced its product price to below 40,000 yuan, accelerating the approach of humanoid robots’ “price range” to the consumer market.

Policies and Funding Need Precision Relay

“Robot performances can attract audiences, but they are hard to generate revenue.” The industry consensus shows that although high-value scenarios for embodied intelligence have emerged, such as automotive assembly in industrial manufacturing, warehouse sorting in commercial logistics, and inspection and rescue in hazardous environments, many companies still have doubts about the return on investment for “replacing human labor with robots,” which means that the development of humanoid robots still requires long-term support from policies and capital.

Zhong Xinlong believes that the current policy and financial support for the humanoid robot industry from major global economies has reached unprecedented levels, sufficient to support leading companies in completing the critical phase from cutting-edge research and development to commercial scenario validation. However, to achieve true large-scale deployment, a more long-term and systematic support system still needs to be established.

However, the leap from “commercial validation” to “large-scale deployment” requires capital intensity comparable to that of the new energy vehicle industry.

Yan Shiye, a partner at Global Law Firm, believes that humanoid robots belong to technology-intensive and capital-intensive industries, requiring substantial upfront investment. Founders’ self-financing is insufficient to support the long-term development and technological breakthroughs of enterprises, while debt financing is less suitable for early-stage companies in the research and development phase. Although government subsidies can provide some financial support, they are limited by subsidy amounts and application conditions, making them unsuitable as a long-term stable funding source for enterprises. In contrast, equity financing can provide humanoid robot companies with ample funds and comprehensive resource support, aligning with the funding needs of enterprises in the technology research and development and market expansion stages, and is an important driving force for rapid enterprise development.

Zhong Xinlong believes that in the future, it is necessary to guide more capital forces with a long-term perspective, such as government-led industrial funds and large industrial capital, to enter and form a capital relay covering the entire lifecycle of “research and development – mass production – application.”

At the same time, policy support also needs to shift from universal research and development subsidies to more targeted application-driven approaches, such as establishing large-scale application demonstration projects in key industries like elderly care services and special industries, providing equipment procurement subsidies, and leading the establishment of industry technical standards and safety certification systems, to create real and sustainable market demand for large-scale deployment, thereby completing the “last mile” of industrial development.

Editor: Wan Jianyi

Proofreader: Zhao Yan

Copyright Statement

All original content from Securities Times platforms is not to be reproduced without written authorization. Our company reserves the right to pursue legal responsibility for related actions.For reprints and cooperation, please contact the Securities Times assistant, WeChat ID: SecuritiesTimes

END

Click on keywords to view

In-depth reports on the Periscope series丨Stock Conference Column丨Investment Xiaohongshu丨eCompany Investigation丨 Times Salon丨Top Ten Star Private Equity Interviews丨Wu Qing’s latest statement! Continuing to consolidate the positive momentum of the capital market丨Big surge! A-shares, August ends! 丨Industrial Fulian, breaking 1 trillion! “Ning Wang”, big surge!丨Related to interest rate cuts! Federal Reserve Chairman “preferred person”, latest statement!丨Late-night breaking! EU: Plans to cancel tariffs on some US goods丨Cambricon, the new “stock king”!丨Chip stocks surge! SMIC hits new high丨The list is here! 26 foreign heads of state and government will attend丨The world’s most valuable company releases financial report, after-hours plunge!丨Late-night, Chinese concept stocks fell! Half-year reports “submitted” gradually! Tonight, over 30 companies’ profits doubled year-on-year!