HONG KONG, Feb 6, 2026 - (ACN Newswire) – Currently, digital transformation has become the core proposition of global industrial upgrading, and the manufacturing industry is also undergoing a disruptive transformation from being “experience-driven” to “data intelligence-driven”. In this process, the deep integration of artificial intelligence and industrial scenarios has become the key to resolving manufacturing complexity and unleashing productivity. The introduction of the Action Plan for Integrating and Empowering Industrial Internet and Artificial Intelligence by MIIT further marks the upgrade of “AI + Manufacturing” from an optional pathway to a mandatory one, giving rise to a blue ocean of value worth tens of trillions.
As a deep participant and leader in the digital and intelligent transformation of the manufacturing industry, COSMOPlat IoT Technology Co., Ltd. (“COSMOPlat”) has, based on its profound understanding of massive data and complex scenarios in the industrial field, built a full-stack AI technology system that includes training and inference platform, large models, intelligent agent development and application platforms. This has not only redefined the implementation path of industrial digitalization and intelligentization, but also continuously strengthened its technical barriers in industrial practice, building a solid and difficult-to-replicate moat, and is expected to benefit deeply from the dividends of industrial digital and intelligent transformation and development. On January 30, COSMOPlat submitted a listing application to the HKEX, embarking on a new chapter of value realization.
AI Restructuring Industrial Digitalization: Full-stack Architecture and Flywheel Effect Build Core Barriers
In traditional perception, the industrial internet is a platform for data connection and circulation, and its technical boundaries remain more at the levels of equipment access, data collection, and process visualization. However, through its original “space-air-ground integrated” full-stack AI technology system, COSMOPlat has deeply embedded “AI+” into its technology architecture, achieving an intelligent closed-loop across the entire chain from perception and decision-making to execution, transforming data into intelligent decision-making capabilities, completely overturning the technology logic of the traditional industrial internet, and building an inimitable technology moat.
At the “space” level, COSMOPlat utilizes “COSMOMind”, the first domestic vertical large model based on an industrial internet platform, as its decision-making hub. “COSMOMind” deeply integrates Haier Group's over 40 years of manufacturing knowledge, as well as expert models and high-quality multimodal datasets accumulated on the COSMOMind Industrial Big Data. It possesses significant advantages in industrial scenario analysis, knowledge Q&A, and logical reasoning, with a retrieval accuracy rate of up to 93% in industrial knowledge fusion and retrieval. Through the "one industry, one model" strategy, COSMOPlat has developed industry-specific large models for enterprises in home appliances, energy and chemicals, energy and carbon industries, providing intelligent decision-making support for major industrial scenarios including research and development design, production manufacturing, operation services and dual-carbon management.
Industry intelligent agents and industrial software serve as the “air link”, transforming the decision-making instructions of large models into more flexible human-machine interaction methods and more efficient execution capabilities. As of the Latest Practicable Date, COSMOPlat has developed 57 industrial intelligent agents covering over 40 core scenarios across 3 major industries, which can handle key industrial functions such as process optimization, equipment management, and smart energy and carbon, helping customers achieve intelligent control of industrial processes, shorten process parameter adjustment cycles, improve equipment operation and maintenance efficiency, and reduce production energy consumption and energy costs.
At the “ground” level, intelligent terminals and automation equipment, acting as the executing limbs, transform AI capabilities into actual production actions and integrate them into the entire manufacturing process, achieving efficient synergy among data, equipment and scenarios, driving industrial AI from isolated intelligent applications to domain-wide intelligence, creating value-added throughout the entire process for customers across various industries, and maximizing the release of data value.
Relying on the “space-air-ground integrated” full-stack AI technology system, COSMOPlat has achieved full-link intelligence across “perception-analysis-decision-execution-feedback”, forming a unique flywheel effect of “data accumulation—model refinement—intelligent agent deployment—data reinforcement”, achieving self-reinforcement and a virtuous cycle where “more data is generated through use, models become more precise through use, and intelligent agent scenarios become broader through use”, enabling COSMOPlat's AI capabilities to continuously evolve in practice and building a solid technology barrier.
Practical Evidence: AI Empowerment from Concept to Significant Scenario Implementation Results
The value of technology must ultimately be tested by the market. COSMOPlat has demonstrated through numerous benchmark cases that its AI empowerment has moved beyond the “proof of concept” stage and is capable of delivering tangible economic benefits to enterprises.
In the petrochemical industry, COSMOPlat has developed the world's first petrochemical large model covering the entire industry chain, entire ecosystem chain, and entire value chain, which has been implemented in a large-scale petrochemical enterprise. Based on the model, the Company has developed 38 intelligent agents, such as the dynamometer intelligent assistant and the aldehyde separation tower process optimization agent, covering scenarios such as oilfield extraction, oil and gas refining and chemical processing, and production safety, with an average efficiency increase of 20%. Taking the dynamometer intelligent assistant as an example, based on the professional capabilities of whole-network search and the COSMOMind, this intelligent agent can “read and understand” dynamometer data of industrial equipment such as deep well pumps, accurately identify equipment failures and provide disposal suggestions, improving the efficiency of oil well equipment failure diagnosis and disposal by more than 50%; the aldehyde separation tower process optimization assistant integrates historical data and expert knowledge, enabling dynamic identification and adjustment suggestions for key parameters, providing decision-making assistance for operators, and significantly enhancing quality control precision and the product first-time conversion rate; the intelligent oil blending assistant quickly generates optimal blending schemes through interpolation algorithms and non-linear programming models, balancing product performance with green production requirements, fully validating the practical value of AI in the heavy chemical industry.
In core manufacturing scenarios, COSMOPlat integrates AI technology throughout the entire chain of R&D, production, operation and maintenance. On the R&D side, taking a home appliance interconnected factory selected as a world “lighthouse factory” as a typical example, an AIGC virtual design assistant was introduced. Based on massive user preference data and an expert design asset library, it generated core solutions for three-drum washing machines. By simulating inner drum operation scenarios in advance to construct a dynamic balance optimization model, it solved the industry pain point of “collisions and interference” during multi-drum operation at the source; on the production side, the injection molding industrial intelligent agent, through deep learning of injection molding professional knowledge, equipment operation data, and process data, etc., precipitated the debugging methods that previously existed only in the minds of process technicians into various intelligent algorithms, ultimately forming professional injection molding knowledge Q&A and automated process parameter recommendation capabilities, building the most powerful “industrial brain” in the field of injection molding.
In the energy sector, COSMOPlat’s energy-carbon model (COSMOMind for energy-carbon) is the first energy-carbon model originating from sustainable lighthouse factory practices. Based on this model, COSMOPlat offers enterprises and industrial park users an Energy-Carbon Management Platform and end-to-end smart-energy solutions, helping customers lower energy cost, improve utilization and advance green manufacturing. These cross-industry and multi-scenario success stories strongly demonstrate the practical effectiveness and replicability of COSMOPlat’s AI solutions.
From an industrial cycle perspective, as the manufacturing system grows increasingly complex and cost and efficiency constraints intensify, enterprises' demand for end-to-end digital and intelligent transformation will continue to rise, ensuring long-term certainty in industry development.
COSMOPlat has pioneered the “space-air-ground integrated” technology system centered on large industrial models, with intelligent agents deeply embedded in production processes. This delivers a closed loop of perception, decision-making, and execution. Through multi-industry and multi-scenario validation, it demonstrates that AI possesses the capability to continuously generate industrial value. As application scenarios continue to expand, intelligent agents scale up, and the data flywheel effect intensifies, COSMOPlat is poised to translate its technological leadership into heightened platform stickiness and greater value per customer. Riding the wave of digital transformation, it is unlocking greater potential for value growth.