Calculating the "economic" for industrial physical AI: Interpreting the pragmatic philosophy of Weiyi Zhizao (03275).

On October 20, Changzhou Weiyi Zhizao Technology Co., Ltd. (03275) officially launched its Hong Kong public offering and is expected to list on the Main Board of the Stock Exchange on October 28, with the potential to become the "first industrial embodied intelligence stock" in Hong Kong. The company plans to globally offer 37.2758 million H shares, including 3.7276 million shares under the Hong Kong public offering and 33.5482 million shares under the international offering, at an offer price range of HK$16.63 to HK$21.46 per share, with an entry fee of approximately HK$2,168 per board lot. Based on the midpoint offer price of HK$19.05, the net proceeds from the offering are estimated at approximately HK$642 million. As the listing process advances, this AI company deeply engaged in industrial scenarios provides the capital market with a sample of real revenue and delivery experience for observing the commercialization progress of embodied intelligence. From 2023 to 2025, Weiyi Zhizao's revenue increased from RMB434 million to RMB796 million, representing a compound annual growth rate of 35.4%. Among this, revenue from embodied intelligent industrial Siasun Robot&Automation (EIIR) increased from RMB114 million to RMB453 million, a cumulative increase of approximately 297.1% over two years, with its share of total revenue rising from 26.3% to 57.0%. In the first five months of 2026, revenue from this business further increased by 39.7% year-on-year to RMB260 million, with its revenue share rising to 61.4%. According to CIC data, based on 2025 revenue, Weiyi Zhizao ranked first among Chinese industrial embodied intelligence Siasun Robot&Automation (EIIR) product providers with a market share of 20.0%. Its position in the segmented market, core product revenue growth and changes in business structure together constitute important highlights of the company's move toward the capital market. For investors, understanding the value of Weiyi Zhizao requires placing technology back into the operating scenarios of factories: whether Siasun Robot&Automation can strike a balance among stability, production takt time and changeover efficiency, whether suppliers can transform process experience into products that can be repeatedly delivered, and whether these capabilities can continuously revenue, gross profit and order growth. The factory's economic calculus opens application space for industrial embodied intelligence Manufacturing is one of the application areas where embodied intelligence has clear commercial demand. Production tasks have specific objectives, quality and efficiency can be measured, and equipment investment also has clear return evaluation criteria. With the increase in small-batch, multi-variety production, systems operating with fixed programs and preset conditions often require additional programming, debugging and tooling investment when facing changes in workpiece position, complex appearance defects or frequent changeovers. Siasun Robot&Automation that can perceive the environment, adjust actions and execute specific processes therefore have room for further penetration. Industrial customers measure the overall cost of use. Recognition accuracy, yield, equipment utilization, downtime and delivery cycles all affect investment returns. Suppliers that can integrate algorithms, mechanical structures and process requirements into production lines are more likely to translate technological progress into efficiency improvements that customers are willing to pay for. Weiyi Zhizao's business path corresponds to this demand. The company accumulated experience from industrial applications such as AI visual inspection and gradually expanded into EIIR products with perception, planning and physical execution capabilities, covering tasks such as inspection, assembly, handling, loading and unloading, grinding and repair. Its products include both single-process Siasun Robot&Automation and composite workstations integrating multiple tasks. For example, in die-casting production, inspection and grinding need to be connected; in the production of new energy vehicle electric drive systems, problems found in inspection need to be further translated into actions such as repair welding. Connecting recognition results with subsequent process execution can increase the tasks undertaken by a single system and also give suppliers the opportunity to extend from one workstation to more production links. The commercial significance of this expansion lies in the simultaneous increase in customer value and supplier revenue space. New processes, new production lines and product changeovers in the same factory may all bring subsequent procurement opportunities. For enterprises with existing delivery experience, understanding customer processes, meeting on-site standards and building trust are important foundations for pursuing these opportunities. Process experience and model reuse build engineering competitiveness Differences in lighting, materials, equipment interfaces, workspace and process standards all affect the performance of industrial AI in different projects. Based on self-developed visual perception, process models and motion control capabilities, Weiyi Zhizao has built a cloud-edge-device collaborative technical architecture that connects model training, edge inference and equipment execution, enabling process requirements to be implemented into Siasun Robot&Automation actions through models and control systems. Among these, the design of "fast and slow thinking" and "human-in-the-loop" reflects a pragmatic approach in industrial implementation. In the initial deployment stage, the human-machine interaction system supports engineers in reviewing and correcting model outputs, bringing expert experience into the workflow; in the continuous optimization stage, the "Yixiu Cloud" and algorithm development platform support data governance, model training and iteration. The combination of real-time adaptation and long-term optimization helps reduce reliance on high-level on-site programming and promotes the reuse of process experience and models in similar projects. Data provides the foundation for this process. According to company disclosure, as of the latest practicable date of the post-hearing information pack, its unstructured precision-labeled database exceeds 23TB, containing more than 1.6 billion precision-labeled records. CIC data shows that in 2025, most industrial embodied intelligence Siasun Robot&Automation enterprises in the industry owned less than 5TB of comparable real industrial scenario data. For industrial models, data corresponding to process standards, sensor information and motion trajectories can support training and scenario adaptation. The company discloses that its labeled data comes from its own facilities; data such as operation logs and yield rates at customer sites remain local to customers and are used to optimize the corresponding customers' local models. Within this boundary, proprietary data, process methods and model development experience constitute the company's technical foundation for sustainable accumulation and reuse. Engineering capabilities still need to be tested through complete project delivery. The information pack discloses that traditional AI projects need to go through data collection, data labeling, model training and model deployment, taking at least 2 to 3 months on average; the company's EIIR projects, with the help of the human-machine interaction system, can provide real-time feedback and integrate human experience and judgment, fundamentally solving the "cold start" problem and taking only 21 days on average. The value of model reuse and remote collaboration lies precisely in gradually reducing repetitive work in these links and improving delivery efficiency and predictability. On the basis of existing scenario-specific models, Weiyi Zhizao is developing an industrial vision-language-action (VLA) model, aiming to further align multimodal information and improve task adaptation capabilities. Existing revenue reflects the commercial value of specialized models and the engineering delivery system, while VLA research and development provides direction for the next stage of product upgrading. The connection between the two is expected to further expand the application scope of the company's existing industrial experience. Core products become the main revenue axis, and commercial value is reflected in financial performance The growth of EIIR revenue is the most noteworthy clue in Weiyi Zhizao's business changes in recent years. From 2023 to 2025, revenue from this business had a compound annual growth rate of 99.3%, significantly higher than the group's overall growth rate; in the first five months of 2026, revenue from this business was RMB260 million, continuing to grow by 39.7% year-on-year and accounting for more than 60% of revenue in the period, with core products having become the main source of the company's growth. This change also affects revenue quality. In 2025, the gross profit margin of the EIIR business was 53.5%, higher than the group's overall gross profit margin of 48.4%; this business contributed approximately 62.9% of the group's gross profit. In the first five months of 2026, the gross profit margin of the EIIR business was 54.8%. The rising share of core product revenue provides support for the growth of the company's gross profit scale. Changes in the product mix reflect the company's progress in expanding into higher-integration applications. In the first five months of 2026, the overall average selling price of EIIR rose from approximately RMB651,000 in the same period last year to RMB1.089 million, mainly related to the integration of more equipment into larger composite workstation sets. Under this counting basis, the change in average selling price reflects adjustments in system integration scale and product mix. A single system undertaking more processes provides a path to expand project value. At the group level, revenue in the first five months of 2026 was RMB424 million, up 22.6% year-on-year; gross profit was RMB205 million, up approximately 20.6% year-on-year. The overall gross profit margin in the period was 48.4%, slightly down from 49.2% in the same period last year, reflecting mix and pricing changes in other businesses. The gross profit advantage of EIIR and the increase in its revenue share are important factors supporting the group's gross profit growth. In terms of profitability, the company recorded net profit of approximately RMB15.739 million and RMB5.066 million in 2024 and 2025, respectively. In 2025, research and development expenses increased to RMB235 million, accounting for 29.6% of revenue, providing continuous investment for product and model iteration and also affecting profit in the period. In the first five months of 2026, the company's adjusted net profit was RMB33.835 million, up 41.5% year-on-year; the adjusted net profit margin rose from approximately 6.9% in the same period last year to approximately 8.0%. This non-IFRS measure adjusts for share-based payment, interest on redeemable liabilities and listing expenses. Net profit in the first five months of 2026 was RMB22.528 million, up 219.1% year-on-year, with a notable contribution from the decline in net financing costs; research and development investment in the period exceeded RMB100 million, up 6.17% year-on-year. Revenue growth, the gross profit advantage of core products and continuous investment together constitute the company's current operating characteristics. Orders provide another support for subsequent operations. In the first five months of 2026, the company newly obtained tax-inclusive orders of approximately RMB530 million, up 49.3% year-on-year; as of the end of May, tax-inclusive contracts in hand amounted to approximately RMB666 million. Order growth reflects customer demand, but its conversion into revenue still depends on project implementation and acceptance, and tax-inclusive contract amounts also differ in basis from financial revenue. Taking these indicators together, Weiyi Zhizao has formed a business foundation in which core product revenue, gross profit contribution and new orders support each other. The focus of the next stage is to transform this foundation into more stable delivery, cash collection and profit accumulation. From existing customers to overseas markets, the subsequent growth path gradually unfolds Industry expansion provides the backdrop for this growth path. According to CIC forecasts, the global market for industrial AI-enabled equipment is expected to increase from RMB465.7 billion in 2025 to RMB1,310.6 billion in 2030, with a compound annual growth rate of approximately 23.0%; the Chinese market is expected to increase from RMB189.1 billion to RMB641.8 billion over the same period, with a compound annual growth rate of approximately 27.7%. Industrial embodied intelligence Siasun Robot&Automation is still in the early stage of penetration within this market. According to the same institution, the proportion of industrial embodied intelligence Siasun Robot&Automation in China's industrial AI-enabled equipment market is expected to rise from 1.2% in 2025 to 14.2% in 2030. This forecast reflects industry expectations for the expanded application of flexible intelligent equipment and also means that enterprises with product and delivery accumulation have opportunities to expand their business as customer demand upgrades. In this market, Weiyi Zhizao has secured a commercialization position in the industrial embodied intelligence Siasun Robot&Automation (EIIR) segment. Facing industrial automation manufacturers, Siasun Robot&Automation enterprises and newly entering embodied intelligence companies, its existing process accumulation, product revenue and customer base provide a starting point for participating in subsequent competition. For the company, scenario expansion among existing customers is a growth direction relatively close to its existing capabilities. According to disclosure, the company served more than 350 industrial customers during the track record period; as of the latest practicable date, it has served more than 25 Fortune Global 500 companies and industry-leading enterprises. Expanding applications around validated process links helps leverage the value of existing customer relationships and industry experience. Automotive and new energy are currently important sources of demand. In the first five months of 2026, the two industries together contributed approximately 89.1% of the group's revenue. They provide the company with opportunities for higher-integration applications and also closely link business performance to the capital expenditure, production line construction and acceptance pace of relevant customers. Expanding mature applications into more industries is a subsequent highlight for the company to expand growth space and enhance business resilience. Overseas business provides a new market entry point. The company has advanced overseas expansion since 2025; as of the latest practicable date, it has obtained 19 overseas orders in hand, with a total contract amount of approximately RMB35.7 million. The existing scale is still early-stage, but it has provided actual projects for observing the applicability of products in overseas markets. As project delivery advances, local service capabilities, customer repurchases and delivery gross profit will become important indicators for evaluating this direction. Core product revenue nearly tripling in two years, the increased gross profit contribution of EIIR and the growth of new orders have already outlined the path of industrial embodied intelligence from application to commercial scale. As process model reuse, composite workstation delivery and overseas projects gradually advance, how this company will expand its existing advantages deserves continued investor attention.