en.Wedoany.com Reported - At the 2026 World Artificial Intelligence Conference and High-Level Meeting on Global AI Governance (WAIC 2026), held in Shanghai from July 17 to 20, China's COSMOPlat launched the Industrial World Model and showcased its full-stack industrial AI capabilities, covering underlying industrial AI technologies, core products, and industry applications. The total exhibition area of this year's conference exceeded 100,000 square meters for the first time, attracting over 1,100 exhibitors and featuring more than 300 products making their global debut.
COSMOPlat made its appearance under the theme "Industrial Intelligence Transformation, Ecosystem for the Future." The Industrial World Model unveiled this time is its core technology foundation for moving toward AI-native capabilities. Based on ontology modeling and digital twin technologies, and fueled by high-quality industrial data, this model enables AI to understand equipment operation and process characteristics. The model boasts three core advantages: ontology grounding, mechanism integration, and trustworthiness with executability. Among these, the data-driven industrial ontology graph construction technology provides a trustworthy data foundation for AI, advancing industrial AI from usable to trustworthy deployment; the complex scenario-oriented industrial mechanism embedded fine-tuning technology builds a continuously evolving industrial knowledge management framework through knowledge fusion, mechanism embedding, and model fine-tuning, allowing AI to truly "understand industry." The platform has accumulated over 4,700 industrial mechanism models, empowered 160,000 enterprises, participated in building 20 lighthouse factories, and led the formulation of standards across four major international standards organizations.
Based on the Industrial World Model, COSMOPlat has developed the Tianyu Industrial Flexible Operation Intelligent Agent Development Platform. This platform adopts a "three-step rapid deployment" zero-code development model. After users describe production tasks in natural language, the platform automatically decomposes tasks, generates code, and completes simulation verification, reducing the development cycle by up to 90%. Additionally, the platform introduces three major AI agent products: COSMO-Sphere focuses on the enterprise strategic level, using a CEO digital human to assist in data governance and real-time AI-driven decision-making; COSMO-iMOM focuses on production execution, using a factory manager digital human to coordinate and dispatch production agents for flexible line changes, equipment management, and quality control; COSMO-iEMS focuses on green and low-carbon operations, using an energy station manager digital human to achieve autonomous energy efficiency optimization and dynamic regulation. These three products cover nine agent matrices for core scenarios such as equipment management and quality control, achieving a 75% improvement in equipment maintenance efficiency and a 50% reduction in line changeover time in industries such as automotive and machinery equipment.
The COSMOPlat exhibition area showcased the value flywheel of industrial AI: real-world scenarios generate high-quality industrial data, data enhances the Industrial World Model, the model drives intelligent agents to execute tasks, agents create business value, and new data flows back to the model, promoting continuous capability evolution. In the next step, the platform will iterate the Tianqi COSMO Claw development platform, lowering the barrier for intelligent agent development and integration, enabling small and medium-sized enterprises to conveniently benefit from industrial AI and promoting the large-scale adoption of industrial AI.
During the results release session, Chen Lucheng, Chairman of COSMOPlat IoT Technology Co., Ltd., shared the platform's evolution path from industrial internet to AI+ industrial internet, and then to an industrial AI-native platform. The platform officially launched the ONE-COSMO Rainforest Ecosystem Plan, with initial partners including Haichuanghui, Shanghai Jiao Tong University, Chinese Academy of Sciences Shanghai, Huawei, China Mobile Shanghai, China Machinery Enterprise Management Association, Fenglan Technology, and Yunxi New Materials, focusing on collaboration in technological innovation, data elements, and resource ecosystems.
The platform demonstrated practical achievements in vertical industries. At the automotive connector company Shunke Zhilian, the platform helped establish a full-process data link from material management, equipment maintenance, and production execution to warehousing and logistics, achieving precise global quality traceability with a finished product inspection pass rate of 99%. In the green and low-carbon field, COSMOPlat's Carbon Exploration Energy collaborated with Xiamen Tianma, focusing on the full chain of "carbon calculation, carbon reduction, and carbon trading," to create the first centralized cooling system carbon credit project in China. Through AI intelligent optimization and system integration, the cooling system's energy efficiency ratio was increased to 6.22, reaching industry-leading energy efficiency levels, achieving a carbon reduction of 17,000 tons in the first year, saving 10.5 million yuan in annual electricity costs, and listing the carbon reduction at 65 yuan per ton on the Guizhou Green Finance Low-Carbon Trading Center. In terms of urban empowerment, the platform adopted a 1+N+X model and has been implemented in over 10 cities across China, including Qingdao, Jinan, Jingdezhen, and Deyang. In the Qingdao Jiaozhou SCO Demonstration Area, the platform promoted full digital transformation coverage for 1,000 industrial enterprises above designated size, driving a 9.3% increase in the added value of industrial enterprises above designated size in the city.
