Green AI in IoT-Enabled Supply Chains: A Serverless Cloud Architecture for Circular Economy Optimization and Dynamic Capabilities

The integration of the Internet of Things and Artificial Intelligence into global supply chains presents significant opportunities for operational optimization. However, the computational intensity of traditional machine learning models frequently undermines corporate sustainability goals, introducing a paradox where digital transformation increases environmental degradation. This paper proposes a cloud-native architectural framework utilizing Amazon Web Services to construct a Green AI pipeline for IoT-enabled circular economies. By deploying distributed Python middleware integrated with highly optimized, serverless inference engines, the proposed system ingests high-frequency logistics telemetry while dynamically minimizing computational energy consumption. Preliminary architectural evaluations demonstrate that decoupling data ingestion from energy-intensive continuous training phases allows organizations to build adaptive dynamic capabilities without sacrificing environmental sustainability. This methodology provides digital business researchers with a deterministic, highly scalable technological foundation for evaluating the intersection of disruptive innovation and Green AI.