Intelligent computing architectures play a vital role in enabling sustainable, large-scale IoT ecosystems that support AI-driven industrial transformation. By integrating advanced computation, communication, and intelligence, such architectures empower industries to achieve higher efficiency, automation, and data-driven decision-making while addressing sustainability goals. Despite these advancements, existing IoT–AI integration methods face critical challenges, including high energy consumption, centralized data dependency, scalability limitations, increased latency, and privacy risks due to excessive cloud-based data transmission and monolithic learning models. These issues hinder real-time intelligence and sustainable industrial deployment. To address these limitations, this paper proposes the Green Federated Edge–Cloud Intelligent Computing Architecture (GFEC-ICA). The framework combines energy-aware edge computing, federated learning, and cloud-based global coordination to enable distributed AI intelligence while minimizing energy usage and network overhead. Carbon-aware task scheduling, adaptive workload migration, and lightweight model optimization are incorporated to enhance sustainability and scalability. The proposed architecture is applied to AI-driven industrial environments, enabling real-time analytics, predictive maintenance, and intelligent process optimization across geographically distributed IoT systems. Experimental analysis demonstrates that GFEC-ICA significantly reduces energy consumption, communication cost, and latency while improving learning accuracy, system resilience, and operational efficiency, thereby enabling sustainable and scalable AI-driven industrial transformation. The proposed method achieves the energy consumption of 540 to 770 kWh, carbon footprint of 620 gCO₂, learning accuracy of 96%, scalability attains 680 units at 500 nodes, and system resilience of 92%. Reduces energy use and carbon footprint while supporting large-scale AI-enabled IoT systems. Enables faster industrial decision-making through distributed edge–cloud intelligence. Improves reliability and scalability for sustainable AI-driven industrial operations.
Jiaqi Cao (Thu,) studied this question.