Real-time Analytics
Xiangtan Steel needed a way to scale industrial AI across many production scenarios because fragmented projects, repeated model-workflow setup, and the lack of a unified AI development platform made steel AI expensive and hard to replicate.
Reported impact
Xiangtan Steel moved industrial AI from isolated experiments to a centrally managed platform, putting the first wave of large-model applications into live steelmaking workflows and giving the group a scalable way to expand AI across the plant.
Xiangtan Steel used Huawei Cloud Stack to build a large-model hybrid cloud and a plant-wide AI training center for steel-production workflows. The foundation combines an AI platform for model training and inference, a resource scheduling platform for compute management, and a data integration platform that connects production data under unified standards, architecture, and data specifications. On top of that base, Xianggang's AI large model supplies vision and prediction workflows for image classification, object detection, instance segmentation, semantic segmentation, and structured-data prediction. An AI operations management center governs datasets, applications, deployment, optimization, cloud-edge coordination, and asset management so teams can run the full lifecycle of industrial AI from training through production operations. The result is a reusable large-model platform that replaces one-off workshop-style development with a scalable steel AI factory.
Production and process data from coking, sintering, ironmaking, and steelmaking are connected through the data integration platform into the hybrid-cloud AI foundation. The AI foundational platform trains and serves vision and prediction workflows, and the AI operations management center manages deployment and optimization as results are pushed into live industrial applications.
"By December 2024, Xiangtan Steel had explored more than 130 application scenarios across 9 specialties and launched 32 intelligent applications. (original: 截止到2024年12月，已在9大专业，探索了130多个应用场景，上线32个智能应用。)"
"Fragmented scenarios made AI hard to land because each scenario repeated model selection, data processing, model optimization, and iteration, which made costs high and replication difficult. (original: 碎片化场景导致技术落地难，钢铁行业存在大量生产场景，每个碎片化场景独立完成模型选择、数据处理、模型优化、模型迭代等一系列开发步骤，导致技术落地成本高、难以规模化复制的问题。)"
"Most AI scenarios previously used workshop-style development, so Xiangtan Steel lacked a unified AI development platform and could not accumulate common knowledge at scale. (original: 缺少统一AI开发平台，目前大部分人工智能场景都是采用传统“作坊式”开发，无法沉淀通用知识，且不同领域的调试方法不同，影响了人工智能在钢铁行业的规模应用。)"
"The AI training center uses Huawei Cloud Stack large-model hybrid cloud to build a one-stop large-model platform on top of unified standards, architecture, and data specifications. (original: 湘钢人工智能训练中心基于华为云Stack大模型混合云提供具备先进技术架构的通用算力和AI算力服务...建设“三统一”，即统一标准、统一架构和统一数据规范的智算云底座，构建一站式大模型平台。)"
"The Xianggang AI large model provides both vision and prediction workflows, while the AI operations management center supports training, deployment, optimization, cloud-edge coordination, and lifecycle management. (original: 湘钢人工智能大模型：提供视觉大模型和预测大模型...AI运营管理中心：支持数据集管理、应用管理、应用优化、服务管理、云边协同、资产管理和系统管理。)"
"Xiangtan Steel has already launched its first batch of large-model applications on Huawei Cloud Stack across coking, sintering, ironmaking, and steelmaking, delivering plant-wide digital empowerment. (original: 湘钢基于华为云Stack大模型混合云，已完成首批人工智能大模型开发上线，落地涵盖焦化、烧结、炼铁、炼钢等钢铁生产环节，实现全流程、全业务数字化赋能。)"
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Huawei Cloud Sample Archived · Apr 21, 2026 Chinalco Group Chinalco needed to apply AI across a dispersed nonferrous-metal value chain where complex process control, safety risks, fragmented data, and plant-specific deployments made traditional optimization and inspection… 18 production scenarios Aluminum Intelligence Cloud Scheduling · Green Aluminum Cloud Wisdom · Huawei Cloud Stack Huawei Cloud Sample Archived · Mar 11, 2026 Nanjing Iron and Steel Nanjing Iron and Steel had rich plant data and high-value optimization opportunities, but data sharing, model adaptation to steel processes, and fragmented scenario development made it hard to scale AI across the… 20 intelligent applications in production AI coal blending system · Huawei Cloud Stack · double-model intelligent Q&A Huawei Cloud Sample Archived · Jan 15, 2026 Conch Group Conch needed to scale AI across cement production, but traditional model development was slow, hard to generalize, and too weak to support real-time quality, energy, equipment, and safety optimization at industrial… 28 equipment-monitoring scenarios AI training center · Huawei Cloud Stack · NLP large model 