i-UG i-Power Conference 2026

This presentation explores the accelerating impact of generative AI on data centres and the profound implications for energy consumption, infrastructure design, sustainability, and policy. As Gen AI drives unprecedented growth in compute density, power demand, and cooling requirements, data centres are becoming increasingly constrained by grid capacity, environmental limits, and public scrutiny. The session examines the full AI lifecycle—model training, tuning, and inference—and how each stage reshapes power, cooling, networking, and physical design assumptions within modern facilities.
Drawing on global industry data and real-world examples, the presentation highlights the widening gap between traditional enterprise workloads and AI-driven infrastructure, including the shift toward hyperscale facilities, ultra-high-density cabinets, and non-air-based cooling technologies. It challenges conventional efficiency metrics and net-zero claims, arguing that offsets and accounting mechanisms cannot keep pace with absolute growth in energy demand.
Using a political, economic, social, and technological (PEST) lens, the talk considers regulatory pressure, public perception, skills shortages, and the societal consequences of AI-driven automation. It evaluates future options for powering data centres—ranging from grid expansion and renewables to nuclear and on-site generation—and discusses the growing importance of heat reuse and energy system integration.
The presentation concludes by positioning generative AI as both a disruptive force and a catalyst for innovation in data centre design, operations, and workforce development, emphasising the need for more transparent measurement, integrated planning, and long-term thinking across IT, facilities, industry, and government.