[Kuala Lumpur, Malaysia, September 30, 2025] Shi Mao, CEO of Huawei’s Steel & Non-ferrous Metals Business Unit, announced Huawei’s new top-level architecture for industry digitalization and intelligence called** “One Network, One Cloud, Three Platforms”** at the 2025 ASEAN Iron and Steel Forum.

The architecture aims to streamline IoT devices, networks, cloud infrastructure, and upper-layer applications to help the steel and smelting industries go digital and intelligent. Shi also shared how the architecture was already being implemented in frontline steel production scenarios, helping address customer needs and solve industry challenges.

Steel production involves a number of complex manufacturing processes, from ironmaking and steelmaking to rolling and production management. AI is being explored by many to optimize high-value scenarios within these processes. Shi explained that Huawei has identified 367 specific scenarios across 16 process categories where AI can be used to improve production quality and efficiency. These scenarios include blast furnace temperature prediction, belt anomaly monitoring, steel plate surface quality inspection, and energy consumption analysis and prediction. Intelligent solutions tailored to these scenarios can cut costs, improve quality, increase efficiency, and enhance safety.

Shi Mao, CEO of Huawei’s Steel & Non-ferrous Metals BU, delivering a keynote speech

Huawei’s new “One Network, One Cloud, Three Platforms” architecture supports the implementation of AI in these scenarios in a number of different ways. Its “One Network” refers to a unified physical network that uses technologies like network slicing to support multiple services, such as production control, campus management, and video security, as well as time-sensitive networking (TSN) technology to provide the deterministic low latency needed by core industrial control systems.

The “One Cloud” refers to synergistic cloud infrastructure that connects production facilities with the systems of their parent group. By deploying a group-level cloud, manufacturers can centrally deploy shared resources like data lakes and AI training platforms. The group-level cloud serves as the hub for decision-making and capability development, and is linked to edge clouds in individual factories. These factory clouds act as production execution centers, managing local data collection, AI inference, and closed-loop production more reliably in real time. This cloud-edge synergy model guarantees both centralized management and agility at the edge.

The “Three Platforms” refer to the final layer of the architecture which includes a data enablement platform, an AI enablement platform, and an application enablement platform. The data enablement platform establishes a unified data standard system and governance capabilities that help transform production data into high-quality data assets. The AI and application enablement platforms are built upon a large model tailored specifically for the steel industry and provide foundational vision and prediction models, as well as other agile and efficient application development capabilities. Together, these three platforms significantly lower the threshold for application development, allowing more industry experts to engage in innovation targeted at specific business scenarios.

To date, Huawei has collaborated with more than 300 metallurgy firms around the world. In a partnership with a major steel company, Huawei and the company unveiled the world’s first large model for the steel industry, which has already been applied in 32 AI-enabled production scenarios. In refining optimization scenarios, a prediction AI model has been used to precisely control the composition and temperature of molten steel, helping save approximately CNY2 million (approx. USD 0.28 million) per furnace every year. In intelligent belt monitoring scenarios, AI has been used to increase inspection efficiency by 50% compared to manual inspections.

In collaboration with another steel company, Huawei has applied the prediction AI model to improve width prediction in the hot rolling process. This new application has improved accuracy by over 3% compared to the previous model, improving the yields of a single production line by over CNY10 million (approx. USD 1.4 million) annually. This model has also been used in blast furnace operations where it can analyze hundreds of factors that affect furnace temperature. This helps improve furnace temperature and silicon content prediction accuracy to over 90% while reducing annual costs per furnace by over CNY16 million (approx. USD 2.3 million).

Shi concluded his speech by emphasizing that an open, collaborative ecosystem is crucial for digital and intelligent transformation, and that digital and intelligent tools will play a key role in breaking down barriers between industry partners who understand scenarios, engineers who understand equipment, and digital talent who understands AI.

Moving forward, Huawei plans to continue collaborating with partners from industry and academia to transform expertise and experience into applications that can be rapidly replicated at scale. Together, they can provide reliable digital and intelligent platforms that can unlock the potential of equipment and assets, create value with technology, and contribute to a thriving digital and intelligent ecosystem of the steel industry.

The 2025 ASEAN Iron and Steel Forum was jointly organized by the South East Asia Iron and Steel Institute (SEAISI) and the China Iron and Steel Association (CISA). It brought together industry representatives from ASEAN and over 20 other Asian countries to discuss the future development of the steel industry. The global steel industry is at a critical juncture of evolving from digital to intelligent transformation, where AI’s deep integration into industry scenarios is playing a key role in driving more quality, efficient, and sustainable industry development.