On March 20, 2026, the Shanghai Stock Exchange officially accepted the IPO application of Yushu Technology for the STAR Market. The company plans to raise 4.202 billion yuan, officially launching its bid to become the first "embodied intelligence" stock on the A-share market. As part of the second batch of companies applying under the STAR Market's "pre-review mechanism," Yushu Technology had already completed two rounds of inquiries and responses prior to its formal application. This process is expected to significantly improve listing review efficiency compared to conventional IPO procedures.
The Yushu Technology IPO is considered a milestone event for the humanoid robot industry. On one hand, it signifies that the industry is beginning to establish the foundational conditions for "technology validation - scaled implementation - capital empowerment." On the other hand, the company's high profitability challenges the conventional market perception that robotics companies require heavy early-stage investment with difficulty achieving profits, potentially leading to a revaluation of the entire sector.
Analysis of key points from the company's prospectus reveals several notable aspects. First, the company's counter-intuitive high profitability stands out, with 2025 non-GAAP net profit attributable to shareholders reaching 600 million yuan and a gross margin on its main business exceeding 60%. This high profitability stems from the company's core competitiveness built on technology, cost control, and branding, challenging the typical "high investment, low profit" narrative for early-stage robotics firms.
Second, the company is significantly increasing its investment in AI models, planning to allocate nearly half of the raised funds to intelligent robot model R&D projects. This shift indicates a strategic pivot from "hardware leadership" to a balanced "hardware and software" approach. As hardware matures, competition in the robotics industry is expected to gradually shift towards general intelligence capabilities, and this move is likely to enhance the company's long-term competitiveness.
Third, a successful listing would provide a clear capital market pricing benchmark for the industry, potentially reshaping the valuation framework for the entire robotics sector. Companies within Yushu's supply chain are also expected to benefit. Based on comparable company analysis, the company's market capitalization is projected to exceed 100 billion yuan in the early listing phase, with medium to long-term expectations surpassing 300 billion yuan.
The unexpectedly high profitability is underpinned by a trinity of core competencies in technology, cost, and branding. Yushu's fully self-developed hardware system capability, encompassing high-performance motors, reducers, dexterous hands, LiDAR, and various sensors, enables system-level协同 optimization while reducing external procurement costs. Compared to manufacturers reliant on externally sourced core components, Yushu holds advantages in BOM costs and possesses stronger product iteration capabilities, supporting higher gross margins.
Scaled manufacturing has led to rapid cost reductions. The company's strategy of proactively lowering product prices has expanded application scenarios and customer bases through competitive pricing. Mass production has subsequently diluted manufacturing costs, creating a virtuous cycle of price reductions leading to volume increases, further cost reductions, and additional volume growth.
Strong brand效应 has enabled high market share with low sales expenses. In the first three quarters of 2025, sales expenses were only 76.01 million yuan, with a sales expense ratio of 6.5%, of which advertising and promotion expenses accounted for merely 22.57 million yuan. Yushu has built brand influence through technology demonstrations and open-source ecosystems, allowing market expansion without significant sales investment through organic traffic and word-of-mouth.
The substantial investment in AI models, allocating nearly half of the IPO proceeds to model R&D, aims to bolster the company's software capabilities. This strategic decision reflects founder Wang Xingxing's evolved understanding of the robotics industry. Wang has acknowledged underestimating AI's technological progress over the past decade and stated at the 2026 GTC conference that the industry remains significantly distant from its "ChatGPT moment," with the primary challenge being the absence of a truly capable embodied intelligence model with strong generalization for stable task execution in unfamiliar scenarios.
The company has initially established an AI large model layout, having open-sourced its "World Model-Action" (WMA) and "Vision-Language-Action" (VLA) models, which have gained widespread use and positive feedback in the global developer community. The open-source release of the general WMA model "UnifoLM-WMA-0" in September 2025 has demonstrated operational capability across various robot platforms and tasks in real-machine tests. The open-source VLA model "UnifoLM-VLA-0," released in January 2026, showed strong generalization, completing 12 different manipulation tasks in real physical environments using a single policy, indicating its potential as a general robot multi-modal foundation model.
AI large models are crucial for transitioning robots from flexible execution to intelligent decision-making. While consolidating its hardware position, the company is shifting resources towards enhancing general intelligence software capabilities to support long-term development. The industry's evolution suggests initial competition focuses on hardware and engineering capabilities, building early barriers through performance and cost advantages, while later stages will shift towards AI models and algorithms, defining robots' intelligence ceiling and scenario boundaries. The increased investment aligns with this industry evolution, signaling a strategic intent to transition from "hardware leadership" to an "AI + hardware" ecosystem builder. Companies with both hardware engineering积累 and AI model R&D capabilities are expected to dominate future competition and benefit fully from the embodied intelligence era.
The listing provides a valuation anchor for the industry chain and is expected to reshape the sector's valuation system. Yushu's initial issuance market capitalization is projected to exceed 42 billion yuan. Given the lack of profitability among comparable listed robotics companies, a PS valuation metric is applied. Compared to domestic peers like Ubtech (HKEX:9880) and Dobot (HKEX:2432), which traded at around 36x PS upon listing, and considering the premium for Yushu's status as the "first embodied intelligence stock" on the A-share market, a 60x PS multiple is applied. Based on the company's 2026 humanoid robot shipment target of 10,000-20,000 units, the early market capitalization could surpass 100 billion yuan.
Globally, leading humanoid robot companies like Tesla and Figure are valued based on the long-term potential of "AI + robotics platforms," often commanding high PS multiples. At scale, a fair valuation is estimated around 20x PS, which raises the valuation ceiling for robot manufacturers. Figure's post-money valuation reached $39 billion after its Series C round in September 2025, implying a ~20x PS based on 100,000 units at $20,000 each. Tesla's 10-year compensation plan for Elon Musk targets cumulative deliveries of 1 million Optimus robots, implying a $500 billion valuation for the business, or ~20x PS at a远期 price of $25,000 per unit.
Comparatively, Yushu leads in hardware systems and motion control capabilities but requires further development in software and model capabilities, with global布局 and ecosystem building still in early stages. However, the company has率先 achieved key milestones: establishing a virtuous cycle of cost control and volume growth through full-stack self-development and scaled manufacturing, achieving profitability – a rarity among global robotics firms – and strategically pivoting towards an integrated hardware-software platform through its IPO-funded AI model investments. Currently, Yushu is better characterized as a "profitable hardware leader transitioning towards a platform company."
Considering the gap in software capabilities and global presence, but balancing this with its profitability and volume certainty, a 15x PS multiple is considered appropriate, referencing the ~20x PS of leading overseas firms. Assuming the domestic humanoid robot market reaches 1 million units at an average price of 100,000 yuan, with Yushu capturing a 20% market share, and applying a 15x PS to the humanoid business plus other segments like quadruped robots, the medium to long-term market capitalization expectation could exceed 300 billion yuan.
The potential revaluation of整机厂 manufacturers could create a Matthew effect within the humanoid robot industry chain, with core component companies also enjoying significant valuation premiums. An upward revision in the valuation space for robot本体 companies would elevate the overall valuation level of the industry chain. Particularly for core component companies, substantial early-stage investment in整机 companies will generate significant demand for core components, offering considerable room for valuation expansion during the industry's initial phase.
Recent corrections in the robotics sector, with pullbacks普遍 reaching 25%-30%, have充分释放 risk, with sentiment and attention at a low point, presenting strategic allocation opportunities. The Yushu IPO is not merely a single company's capital event but a significant milestone for the robotics industry. Its core significance lies in validating the commercial viability of robot本体厂 and providing a clear valuation anchor for the industry chain, pushing the sector from a "technology narrative" phase towards an "industrial validation" stage.
Beneficiaries are expected to include three categories within the Yushu ecosystem. First, deeply integrated core suppliers: While Yushu initially focused on in-house R&D and production of core components, its disclosed procurement includes structural parts, plastic parts, general electronic components, and certain materials. As scale increases, the company is likely to further build a controllable supply chain, benefiting companies already in or likely to enter its supply system and participate in joint development. Second, model and software ecosystem partners: Given the emphasis on embodied intelligence large models, companies cooperating or adapting at the model, algorithm, and software levels will benefit from enhanced ecosystem value through deeper hardware-software协同. Third, scenario implementation partners: Collaborators working with Yushu to deploy solutions in industrial, security, and service applications will accelerate market penetration and benefit from downstream demand释放.