--- title: "\"First Stock of Visual Embodied Intelligence\" Listed on Hong Kong Stock Exchange: A Commercialization Path Leading with Scenarios" type: "News" locale: "en" url: "https://longbridge.com/en/news/292225257.md" description: "On July 8th, RECONOVA was listed on the Hong Kong Stock Exchange, receiving over 3600 times subscription in Hong Kong stocks. As the \"first stock of visual embodied intelligence,\" its valuation logic reflects the capital market's shift towards focusing on AI commercialization capabilities. The company adopts a \"scenario-first\" approach, addressing machine perception issues from the actual needs of the B-end, and then extending into the field of embodied robotics, distinguishing itself from traditional ontology or general AI routes" datetime: "2026-07-09T02:07:51.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/292225257.md) - [en](https://longbridge.com/en/news/292225257.md) - [zh-HK](https://longbridge.com/zh-HK/news/292225257.md) --- # "First Stock of Visual Embodied Intelligence" Listed on Hong Kong Stock Exchange: A Commercialization Path Leading with Scenarios On July 8th, the Hong Kong stock "first stock of visual embodied intelligence" RECONOVA officially landed on the Hong Kong Stock Exchange. The Hong Kong public offering was subscribed 3,646.06 times, and the international offering was subscribed 3.08 times. This is not an isolated IPO event. Placing it within the coordinates of the AI industry in 2026, it reflects a more noteworthy signal: the capital market's aesthetic for AI companies is changing. In the past two years, AI and embodied companies have grabbed headlines through "parameter scale" and "technology demonstrations." However, by mid-2026, the secondary market began to measure the value of enterprises using a different metric, shifting the valuation logic to "commercialization capability"—not "how large your model is" or "how smoothly it runs," but "can your technology work in real scenarios?" Unlike other companies, **RECONOVA provides a "scenario-first" sample. This company does not "find scenarios with algorithms," but instead starts from actual B-end demands such as airports and commercial spaces, first solving how machines can "understand," and then extending towards embodied robots along the business process, gradually enhancing decision-making and execution capabilities.** The founder of RECONOVA, Zhan Donghui, summarizes it in one sentence: "AI must evolve from being a human intelligence assistant to a human productivity assistant." ## **Starting from "Vertical Scenarios"** RECONOVA's entry into the field of embodied intelligence was not a spur-of-the-moment decision. When the company was founded in 2012, it started with the goal of enabling machines to "see." Over fourteen years, it first focused on perception, then cognition, and finally execution. Zhan Donghui describes this path as "first building the brain, then building the limbs." This sequence determines the difference between RECONOVA and other embodied intelligence companies. Currently, there are roughly two mainstream paths in the embodied intelligence track: one is the robot body company, which aims to open up scenarios with products—starting from motion control, whole machine engineering, and mass production capabilities, attempting to reduce costs through scaled production and thus push robots into more scenarios. The other is the general AI company, which tries to equip robots with a "brain" using more powerful models to enhance generalization capabilities. However, both routes face a common challenge: the evolution of capabilities has not kept pace with the actual demands of general scenarios. **RECONOVA follows a third path—defining products based on specific scenarios.** It first identifies which industrial tasks are suitable for machines to complete, and then works backward to determine the required perception, decision-making, and execution capabilities. This approach is related to its team’s background. Founder Zhan Donghui graduated from Nanjing University with a major in electronics and information systems and worked at Huawei for nearly nine years. The company's CTO, chief scientist, and other technical executives all have doctoral backgrounds. This determines RECONOVA's "engineering" path: starting from specific scenario problems and then expanding the product range based on customer needs. This route has already been validated in its existing visual intelligence business. According to the prospectus, from 2023 to 2025, RECONOVA's revenue is expected to grow from 242 million yuan to 443 million yuan; in 2025, smart civil aviation, smart commerce, and smart safe driving are expected to contribute revenues of 172 million yuan, 154 million yuan, and 116 million yuan, respectively This also indicates that **entering the scene first, understanding the task, refining the product, and then extending to more complex execution capabilities may also become a realistic path for the commercialization of embodied intelligence.** ## **Focusing on productivity, will B-end applications be the main battlefield?** In the judgment of RECONOVA founder Zhan Donghui, in the next three to five years, embodied intelligence will first scale in enterprise-level scenarios. Although the entry threshold for B-end is higher, compared to the C-end with dispersed demand and open scenarios, the task boundaries in fields such as industry, logistics, and transportation are clearer, processes are more standardized, and efficiency improvements are easier to quantify. Airport baggage handling is a typical scenario. In 2025, RECONOVA will combine cognitive decision-making capabilities with robotic execution capabilities to launch the "Xiao Yi" baggage transfer robot. The composite actuator developed by the company integrates functions such as suction cups, grippers, and hooks, allowing it to operate on different types of baggage combinations, which is different from the traditional impression of dexterous hands, reflecting a practical-first approach. **In Zhan Donghui's view, "for B-end customers, whether the robot is human-like enough is not the primary standard; the ability to stably complete tasks, reduce costs, and continuously create production value is more important."** **In April 2026, RECONOVA will launch the VTFLA multimodal embodied large model, adding tactile and force feedback to the traditional VLA framework**, enabling robots to determine whether the grasp is stable, whether the force is appropriate, and to adjust actions in real-time. RECONOVA will focus its self-research efforts on end-effectors, whole machine engineering, and hardware-software collaboration. Zhan Donghui particularly emphasizes a metric that is rarely mentioned in the field of embodied intelligence—robustness. "To really complete 7×24 hours of work in the production environment of enterprise customers, the engineering design, robustness, and safety reliability of the robot itself will become very core factors." **Compared to companies that start from models or robot bodies and then look for application scenarios, RECONOVA's advantage lies in having already entered real production processes,** accumulating a large amount of industry data, customer needs, and engineering delivery experience in long-term services in civil aviation, commercial spaces, and safe driving scenarios. ## **Cross-scenario replication, from "vertical" deepening to "horizontal" expansion** If the first step in the commercialization of embodied intelligence is to enter the scene, then the second step is to extract replicable capabilities from a single scenario. Currently, many robotics companies have been able to complete pilots in scenarios such as factories, warehouses, and airports, but there is still a significant distance from "completing a project" to "replicating a type of product"—each time entering a new industry often requires re-adapting models and hardware, making it difficult to reduce R&D and delivery costs, and the business ultimately remains at a highly customized project stage **In other words, the key to scaling embodied intelligence is not to move the same robot unchanged into all industries, but to clarify which capabilities can be reused and which aspects must be re-adapted.** RECONOVA's strategic difference provides a new observation sample. The company's founder, Zhan Donghui, summarizes his development approach over the past decade as a **"market unchanged, product changed"** vertical deepening strategy: continuously increasing product varieties around markets such as civil aviation, commercial space, and safe driving. After entering the embodied intelligence stage, the company proposed to shift to a **"product unchanged, market changed"** horizontal expansion strategy, entering more application scenarios with relatively focused robotic products. Whether horizontal expansion can be established depends on the ability to reuse across scenarios. Visual, tactile, and force perception, multimodal models, end effectors, and whole machine engineering technologies have certain transferability, but the customer needs, business processes, and safety constraints in different industries are difficult to replicate directly. In this regard, **RECONOVA attempts to reuse not only specific technologies and products but also the methodologies accumulated in identifying needs, building customer trust, and in vertical scenarios and real business processes**. This also explains why the company started by extending embodied intelligence products from airports, as RECONOVA has deep experience in the civil aviation field, understanding that scenarios and trust can be transferred—deploying robots is shorter than anyone else. According to the prospectus, industrial logistics and warehouse automation are already planned in the product roadmap, sharing the same technology platform with airport scenarios—the extension of visual intelligence from perception to execution is a natural expansion of the same technical foundation to adjacent scenarios. From this planning perspective, what RECONOVA attempts to replicate is not a specific form of a robot, but a productization path of "reusing underlying capabilities and adapting to specific scenarios." Additionally, unlike general cloud-based large models, RECONOVA has clearly defined its positioning from the beginning—aiming to achieve **edge-side intelligence**, striving to realize the performance of nearly 100 billion parameter models under limited computing power and cost constraints. ## **Increased investment in R&D, supply chain, and overseas expansion: a new journey after going public** Frost & Sullivan estimates that by 2030, **the potential market size for embodied intelligence products in airports alone is expected to reach 30 billion yuan, which is far** higher than the market size of approximately 6.3 billion yuan for civil aviation visual intelligence products. **New businesses also mean higher investments, and** this IPO has secured RECONOVA a longer window for R&D and commercialization. The prospectus shows that the raised funds will be used for R&D, channel, and production base construction, with an expected annual production capacity of 600 smart boarding gates, 120 security gates, and 200 baggage handling robots after reaching production capacity. At the same time, the company plans to use part of the raised funds to expand overseas sales channels. In Zhan Donghui's view, overseas will become RECONOVA's next market focus in the next ten years. Moving forward, the capital market will focus on whether the various embodied intelligence products developed by the company can transition from pilot projects to large-scale delivery, whether the related capabilities can move out of airports and into logistics, manufacturing, and special operations scenarios, and whether the technological investments can ultimately be converted into revenue, profit, and cash flow For RECONOVA, going public is a summary of more than a decade of visual intelligence business; moving from "seeing and understanding" to "being able to act" is a new battle that has just begun. Whether RECONOVA can successfully navigate this scenario-first path will also provide an important sample for the industry to assess the commercialization direction of embodied intelligence ### Related Stocks - [07656.HK](https://longbridge.com/en/quote/07656.HK.md) ## Related News & Research - [AI glasses have a new hater: Lorde](https://longbridge.com/en/news/292517914.md) - [The whirlwind 72 hours of rival AI announcements](https://longbridge.com/en/news/292338769.md) - [As AI Platforms Move Away From Unlimited, Sogni AI Launches a $20 Fair-Use Unlimited Plan on Community GPUs](https://longbridge.com/en/news/292632827.md) - [9Yards X announces acquisition of Labib AI](https://longbridge.com/en/news/292491763.md) - [The AI Ponzi Machine Is Getting More Sophisticated](https://longbridge.com/en/news/292409228.md)