Yushu Technology IPO: The prospectus mentions "full-stack" self-development 22 times; R&D expenses are only one-ninth of UBTECH's.

Investing in stocks? Look to the Golden Unicorn analysts’ research reports for authoritative, professional, timely, and comprehensive insights to help you uncover potential theme opportunities!

Produced by: Sina Finance Listed Company Research Institute

Author: IPO Refinance/Turing

Recently, Hangzhou Yushu Technology Co., Ltd. (hereinafter referred to as “Yushu Technology”) disclosed its A-share prospectus, making a strong push for an IPO on the Sci-Tech Innovation Board.

Compared to many robotic companies, Yushu Technology’s performance, profit margins, and market shares in niche segments are quite impressive. However, Yushu Technology’s R&D expenses for the complete fiscal year during the reporting period were all under 100 million yuan, lower than comparable peer companies; the R&D expense ratio has significantly decreased overall in the last two years and is notably lower than comparable peer companies. Especially when compared to UBTECH, Yushu Technology’s R&D expenses for three complete fiscal years are only one-ninth of UBTECH’s during the same period.

Yushu Technology mentioned “full-stack” self-research 22 times in its prospectus, claiming to have fully developed core model algorithms for robots, intelligent systems, as well as high-performance motors, reducers, dexterous hands, laser radars, and other core components. Interestingly, despite Yushu Technology developing so many systems and components in a full-stack manner, it has only 262 patents; among the 169 domestic patents, there are only 20 invention patents. In contrast, UBTECH had an impressive 2,790 authorized patents by the first half of 2025, with over 55% being invention patents, all superior to Yushu Technology.

**　　Data Highlights**

The prospectus shows that Yushu Technology focuses on the R&D, production, and sales of high-performance general humanoid robots, quadruped robots, robot components, and embodied intelligent models. By 2025, the company had shipped over 5,500 humanoid robots (purely humanoid, excluding wheeled dual-arm robots), ranking first in global shipments.

From 2022 to 2025, Yushu Technology achieved revenues of 123 million yuan, 159 million yuan, 392 million yuan, and 1.708 billion yuan, respectively, with net profits attributable to the parent company of -22 million yuan, -11 million yuan, 95 million yuan, and 288 million yuan, rapidly turning losses into significant profits. In 2025, the company’s net profit after deducting non-recurring items reached 600 million yuan.

Compared to comparable peer companies (the prospectus considers UBTECH and Yujian as comparable peers), Yushu Technology quickly achieved massive profits. From 2022 to 2024 and in the first half of 2025, UBTECH reported net profits of -975 million yuan, -1.234 billion yuan, -1.124 billion yuan, and -414 million yuan, resulting in a cumulative loss of 3.75 billion yuan over three and a half years.

Yujian reported net profits of -52 million yuan, -103 million yuan, -95 million yuan, and -41 million yuan from 2022 to 2024 and in the first half of 2025, leading to a cumulative loss of 290 million yuan over three and a half years.

Yushu Technology’s gross margin is also significantly higher than its peers. From 2022 to 2024 and in the first three quarters of 2025, Yushu Technology’s gross margins were 44.94%, 44.75%, 56.98%, and 59.83%, showing a significant upward trend overall.

Source: Prospectus

In comparison, UBTECH’s gross margins from 2022 to 2024 were 29.16%, 31.53%, and 28.65%; Yujian’s gross margins during the same period were 40.75%, 43.54%, and 46.56%, all significantly lower than Yushu Technology.

Yushu Technology stated that during the reporting period, its comprehensive gross margin was higher than that of its peer listed companies. On one hand, there are differences in product composition and application scenarios between the company and peer listed companies. UBTECH’s robots mainly consist of educational intelligent robots, logistics intelligent robots, and consumer-grade products; Yujian’s robots mainly consist of collaborative robots, primarily applied in manufacturing, retail, healthcare, STEAM education, and research scenarios. On the other hand, the company’s cost control capabilities are industry-leading. Since its establishment, the company has adhered to a full-stack self-research technology path for complete robots and core components, forming a vertical integration capability over the supply chain by self-developing and self-manufacturing core components, which significantly reduces material procurement and manufacturing costs. Furthermore, large-scale production has further strengthened the company’s bargaining power with upstream suppliers, creating a continuous cost advantage.

**　　R&D Expenses are One-Ninth of UBTECH’s**

In just a few years, Yushu Technology has achieved significant revenue growth, realizing large profits amid massive losses from comparable peer companies, and has adopted a full-stack self-research technology path for complete robots and core components. As the second company to be pre-reviewed for an IPO on the Sci-Tech Innovation Board and a well-known robotics company, many investors believed that Yushu Technology would have R&D expenses and expense ratios far exceeding those of peers, and that intangible assets such as invention patents should be significantly ahead.

However, the fact is that the R&D expenses for the complete fiscal year during the reporting period were all below 100 million yuan, only one-ninth of UBTECH’s; the R&D expense ratio has significantly decreased and is notably lower than the average of comparable peer companies.

From 2022 to 2024 and in the first three quarters of 2025, Yushu Technology’s R&D expenses were 30 million yuan, 50 million yuan, 70 million yuan, and 90 million yuan, with R&D expense ratios of 24.39%, 31.39%, 17.84%, and 7.73%, respectively. For the three complete fiscal years from 2022 to 2024, Yushu Technology’s total R&D expenses were only 150 million yuan, averaging 50 million yuan per year, with no year exceeding 100 million yuan.

In contrast, comparable peer UBTECH’s R&D expenses from 2022 to 2024 were 428 million yuan, 491 million yuan, and 478 million yuan, totaling 1.397 billion yuan, which is 931% of Yushu Technology’s total R&D expenses during the same period.

Another comparable peer company, Yujian, had R&D expenses of 52 million yuan, 71 million yuan, and 72 million yuan from 2022 to 2024, totaling 194 million yuan, which is also higher than Yushu Technology’s 150 million yuan during the same period.

According to Wind data, A-share listed companies involved in “robotic body” businesses include Boshi Co., Estun, New Era, and Tuosida (for rights protection); the combined R&D expenses of these four companies from 2022 to 2024 exceeded 370 million yuan, with annual R&D expenses exceeding 100 million yuan.

UBTECH and Yujian’s average R&D expense ratios from 2022 to 2024 were 32.04%, 35.53%, and 27.92%, higher than Yushu Technology’s data during the same period.

In the first half of 2025 (with data for the first three quarters not yet published), the R&D expenses of UBTECH and Yujian were 218 million yuan and 41 million yuan, respectively, accounting for 35.1% and 26.7% of their respective revenues, both higher than Yushu Technology’s 7.73% in the first three quarters of 2025 (with no data for the first half of 2025).

Yushu Technology stated that in 2022 and 2023, its R&D investment as a proportion of operating revenue was within the range of peer listed companies; in 2024 and from January to September 2025, the proportion of R&D investment to operating revenue declined, mainly due to rapid growth in the company’s operating revenue in the most recent year and period, with revenue growth far exceeding the reasonable growth of R&D expenses, resulting in a relative decrease in the R&D expense ratio under the scale effect.

The Prospectus Mentions “Full-Stack” Self-Research 22 Times; Patent Count is One-Tenth of UBTECH’s

In the prospectus, Yushu Technology mentions “full-stack” self-research 22 times, such as “fully developing core model algorithms for embodied intelligence, reinforcement learning, motion control, etc., intelligent systems such as heat management, energy management, motor drive, as well as high-performance motors, reducers, dexterous hands, laser radars, and various sensors as core robot components.”

Source: Yushu Technology Prospectus

According to investors’ simplistic understanding, Yushu Technology’s full-stack self-research of so many core model algorithms, intelligent systems, and core components like reducers, dexterous hands, and laser radars should incur high R&D expenses, and the number of intangible assets such as patents should also be high.

However, the fact is that Yushu Technology’s R&D expenses from 2022 to 2024 were only one-ninth of UBTECH’s and did not exceed Yujian, with the R&D expense ratios in the last two years significantly lower than comparable peer companies.

Interestingly, UBTECH also adheres to full-stack technology and continuously strives to advance key technology groups such as “humanoid brain,” “humanoid cerebellum,” and “high-performance limbs,” including core AI technologies such as high-performance servo drive technology, large model technology, semantic VSLAM technology, learning-based motion control technology, visual perception technology, and multimodal interaction technology, as well as group brain networks 2.0 and Co-Agent technology.

Compared to UBTECH, Yushu Technology has fewer invention patents. As of January 31, 2026, Yushu Technology owned 262 patent rights, with 169 authorized domestic patents and 93 foreign patents, of which there are 20 domestic invention patents, 76 domestic utility model patents, and 73 domestic design patents.

As of June 30, 2025, UBTECH had authorized patents totaling 2,790, more than ten times Yushu Technology’s 262. According to research reports, public information, and industrial and commercial information, over 55% of UBTECH’s 2,790 authorized patents are invention patents.

In contrast, Yushu Technology has only 20 invention patents, which is only one-seventieth of UBTECH’s; Yushu’s invention patents account for only 11.83% of its domestic patent total, far lower than UBTECH’s approximately 55%.

Wind data shows that Yujian has a total of 709 valid patents globally, which also far exceeds Yushu Technology.

Source: Prospectus

Moreover, of Yushu Technology’s 20 invention patents, 11 were applied for after March 21, 2025, indicating that the recently applied invention patents account for 55% of the domestic invention patent count. In contrast, for the nine years leading up to March 2025, Yushu Technology had only 9 invention patents. Yushu Technology was established in August 2016.

In terms of technological accumulation, UBTECH leads with 2,790 authorized patents, with over 55% being invention patents. In terms of R&D investment, UBTECH’s total R&D expenses from 2022 to 2024 amounted to 1.397 billion yuan, far exceeding Yushu Technology’s 150 million yuan. In terms of R&D expense ratios, UBTECH has already far surpassed Yushu Technology in the most recent year.

In terms of technology routes, UBTECH belongs to a full-stack self-research “brain + cerebellum” collaboration, which can be summarized as “full-stack self-research, integrated hardware and software, and end-to-end.” While Yushu Technology also claims a full-stack self-research technology route, it prioritizes cost leadership and small brain development. The prospectus shows that given that embodied large model technology is still in the R&D testing phase globally, Yushu Technology has not yet scaled its self-developed general embodied large model for application in robot products during the reporting period.

Looking ahead, both Yushu Technology and UBTECH face their respective challenges. UBTECH needs to find a balance between technological accumulation and commercial returns to accelerate large-scale profitability; Yushu Technology needs to utilize the funds raised after going public to develop its “brain,” evolving from a “hardware platform provider” to an “embodied intelligent ecosystem builder.”