July 18, Pudong, Shanghai. An AI innovation event in the steel industry that resonated with the 2026 World Artificial Intelligence Conference (WAIC) — the “AI Transformation · Intelligent Innovation for the Future” Artificial Intelligence Collaborative Innovation Forum — officially opened here. Hosted by Baosight Software, a subsidiary of Baowu Group, the forum brought together nearly 200 experts from steel, non-ferrous metals, mining, telecommunications, and other industries. Government, industry, academia, research, and application sectors joined forces to explore new pathways for artificial intelligence empowerment in steel and process industries.
The most anticipated moment of the forum was the joint release of the “AI + Steel” White Paper — Redefining Steel with AI (2026) (hereinafter referred to as the “White Paper”) and the industrial vision foundation model platform “Zhijian Workshop xVue.” With the joint launch by Hu Wangming, Secretary of the Party Committee and Chairman of China Baowu; Jiang Wei, Secretary of the Party Committee, Vice Chairman, and Secretary-General of the China Iron and Steel Industry Association; Liu Baojun, Secretary of the Party Committee and General Manager of Baosteel Co., Ltd.; and Wang Jianhu, President and Deputy Secretary of the Party Committee of Baosight Software, this strategic document, regarded as a “navigation system” for the steel industry’s transition into the artificial intelligence era, was officially unveiled. This was not merely the release of a technical guide, but also a milestone indicating that global steel industry AI applications are moving beyond scattered “pilot experiments” and advancing toward full-process, full-domain intelligent transformation.
As a core foundational industry supporting the national economy, the steel industry has entered a critical period of deep adjustment and transformation upgrading. Against the backdrop of reshaped global market structures and increasingly strict “dual carbon” constraints, the industry has long faced challenges from traditional reliance on human experience, high energy consumption, and low efficiency in long-process production. The “black box” nature of metallurgical processes and technological bottlenecks urgently require breakthroughs. As AI technology advances from discriminative perception toward generative cognitive decision-making, a new technological paradigm provides opportunities to address these long-standing challenges.
The White Paper, jointly compiled by Baosight Software and Baosteel Co., Ltd. under the guidance of the China Iron and Steel Industry Association, represents the first systematic AI transformation methodology covering the entire value chain of the steel industry. Moving beyond the limitations of individual enterprise practices, it provides a complete and practical standardized transformation framework for global steel companies to overcome challenges in existing competition and achieve digital and intelligent upgrading.For the first time, Baosight Software systematically proposed four core concepts in the White Paper: “data-driven development, mechanism integration, closed-loop control, and continuous evolution,” while introducing the new AI 2.0 technological paradigm of “generative cognitive decision-making.” The White Paper builds a comprehensive technological foundation covering underlying support platforms, full-process data governance systems, computer vision, predictive decision-making algorithms, industrial large models, intelligent agents, embodied robots, and other core capabilities.
Its core application analysis section demonstrates how AI capabilities can penetrate every stage from product design to final delivery, including new material research and development, optimization of ironmaking, steelmaking, and rolling processes, production scheduling, quality control, integrated energy-carbon management, and full-chain supply chain decision-making. Even implementation pathways for specialized production lines such as bar and wire products and steel pipes have been analyzed in detail.
The most innovative aspect of the White Paper lies in moving beyond purely theoretical concepts and introducing a four-stage intelligent transformation roadmap applicable across the entire industry. It provides clear advancement pathways for enterprises of different scales and with different digital foundations:
Stage One: Strengthening the Digital Foundation.
Address data silos by establishing unified data governance systems and interconnected industrial internet platforms, enabling dormant data assets to become active and valuable.
Stage Two: Achieving Breakthroughs in Intelligent Single-Point Business Scenarios.
Apply AI in high-frequency, high-value scenarios such as surface inspection, predictive equipment maintenance, and blast furnace operations to achieve precise perception and decision support.
Stage Three: Driving Collaborative Optimization Across Entire Production Lines.
Break down barriers between processes and achieve coordinated optimization of material flow, information flow, and energy flow across ironmaking, steelmaking, and rolling processes, moving from “single-point intelligence” toward “line-level intelligence.”
Stage Four: Achieving Full-Process and Full-Domain Intelligent Upgrading.
Relying on industrial large models and multi-agent collaboration, enterprises can achieve autonomous decision-making and dynamic evolution across the entire value chain, from research and development design to after-sales services, ultimately moving beyond human-dependent production models.
Supporting measures include organizational transformation strategies tailored to the steel industry, as well as data security and privacy protection systems, ensuring that technology develops along safe, fair, and beneficial pathways.
During the achievement release session on the same day, Fei Peng, Member of the Party Committee and Deputy General Manager of Baowu Group; Ai Bing Yu, Foreign Member of the Chinese Academy of Engineering and Fellow of the Australian Academy of Science; Wang Jianhu; and Li Haigang, Vice President of Baosight Software, jointly launched Baosight Software’s latest industrial vision achievement — “Zhijian Workshop xVue.”With the vision of “giving industrial equipment intelligent eyes,” the platform builds an industrial vision “model factory” covering the entire lifecycle of data, models, and applications. It integrates more than 700 AI scenarios across production lines, including blast furnace inspection, converter sampling and transportation, and rolling material handling.
Through the core chain of “data samples — algorithm models — engineering applications,” the platform upgrades traditional project-based visual development models into a standardized and reusable system. This not only solves challenges such as difficult industrial vision data governance and weak model reusability but also represents a significant advancement in Baosight Software’s industrial vision capabilities, providing stable perception support for harsh steelmaking environments characterized by high temperatures, strong interference, and heavy dust.
Hu Wangming, Secretary of the Party Committee and Chairman of China Baowu, stated in his speech that as a global steel leader, Baowu has always regarded artificial intelligence as a key driver for cultivating new quality productive forces and building a world-class enterprise. He emphasized four commitments:
First, uphold industrial responsibility and focus on solving practical challenges to overcome the “last mile” of technology implementation;Second, pursue independent innovation and secure strategic positions in key industrial AI technologies;Third, promote openness and win-win cooperation to build a diverse and collaborative industrial innovation ecosystem;Fourth, prioritize people-oriented development by improving digital skill cultivation systems for industrial employees, enabling mutual empowerment between technological progress and workforce development.
Jiang Wei, Secretary of the Party Committee of the China Iron and Steel Industry Association, noted that the industry’s digital and intelligent transformation still faces systemic challenges, including “daring not to transform, not knowing how to transform, being unable to transform, and transforming too slowly.” Additional challenges include insufficient full-chain integration and the difficulty of adapting general-purpose large models to complex metallurgical processes.
The association is implementing the “Digital and Intelligent Transformation” initiative and accelerating the preparation of the AI + Steel Implementation Guidelines. Jiang Wei emphasized the need to focus on scenario-driven applications, deepen integration between digital intelligence and green development, and jointly build industrial ecosystems, helping small and medium-sized steel enterprises overcome transformation barriers with lower costs and higher efficiency, while promoting the transition from “single-point intelligence” to “global intelligent control.”
The forum invited Ai Bing Yu, Foreign Member of the Chinese Academy of Engineering and Fellow of the Australian Academy of Science, to deliver a keynote speech. As an internationally recognized expert in particle science and process engineering, he systematically reviewed the development of AI applications in process industries and pointed out that mathematical models can study variable impacts at low cost, support process control, and assist scale-up design, leading the industry from “experience-driven” approaches toward “data + mechanism-driven” development.
During the subsequent ecosystem initiative signing ceremony, representatives from Baosteel Co., Ltd., Zhongkuang Zhoushan, Fudan University, CCID Research Institute, Huawei Technologies, Kupass, Baosight Software, and other organizations jointly proposed building a new industrial ecosystem featuring collaborative innovation among industry, academia, research, and application sectors, with shared development and mutual benefits.Huawei discussed pathways for building industrial intelligence supported by domestic computing power. Shandong Iron & Steel Rizhao shared practical experience in AI vision-driven hot rolling shape control. Baosight Software introduced the integrated application of its enterprise intelligent agent framework xAgent and visual intelligence factory.
Moving alone may bring speed, but moving together brings distance. With the release of the White Paper and the construction of the ecosystem, the steel industry is approaching a critical transition point from “steelmaking” to “AI-driven steelmaking.” In the future, as human-machine collaborative intelligent paradigms become increasingly widespread, a global industrial revolution featuring fully autonomous and controllable processes will reshape the steel industry, injecting continuous core momentum toward a smarter, greener, and more efficient future.
Source: Steel Network
