Green Information Systems Architecture: Aligning AI-Driven Infrastructure with Financial Risk and Sustainable Consumer Trends

Abstract

The surge in AI has led to a growing need for computational infrastructure, power, cooling, and specialized hardware, resulting in complex environmental, financial, and governance issues. This study proposes a Green Information Systems Architecture framework for the alignment of AI-driven infrastructure with sustainability goals, financial risk management, and sustainable consumers' expectations. The framework brings together elements of energy-aware architectural design, AI workload optimisation, life cycle assessment, financial materiality, environmental, social and governance (ESG) governance, scenario analysis, and standardised sustainability measures. It takes into account environmental considerations throughout the value chain, including hardware, software, data centres, energy systems, procurement and beyond, as well as Scope 3 emissions, carbon intensity, resource consumption and operational efficiency. The proposed architecture also features specialized AI accelerators, dynamic workload scheduling, efficient model design, energy-aware infrastructure, and ongoing monitoring and evaluation to optimize a range of performance, cost, and environmental metrics. Formal AI risk management and sustainability practices are embedded into governance mechanisms in response to regulatory exposure, climate related financial risks, infrastructure dependencies and accountability. Potential applications for case studies in scenarios include energy-efficient AI training, sustainable infrastructure design, emissions monitoring, financial risk assessment, and predictive resource management. Rebound, data quality, infrastructure focus, reporting quality and financial returns remain major hurdles, however. Development of the future should focus on carbon aware computing, real time energy and emissions data, efficient AI models, heterogeneous infrastructure, edge computing and integrated climate risk analysis