The rapid proliferation of AI workloads, particularly those involving large-scale model training and inference, has significantly altered energy consumption patterns in data centres. This paper critically examines how the growth of AI is driving these transformations, explores the environmental consequences of escalating energy demands, and evaluates the technological and policy-based strategies being developed to mitigate their impact. Through a comprehensive review of hardware, infrastructure, algorithmic, and governance innovations, we highlight the need for interdisciplinary collaboration to align AI development with sustainability goals.