Artificial intelligence (AI) computing is expanding at an unprecedented rate, and AI-centric data centers are quickly becoming some of the largest energy consumers in the tech industry. Projections indicate that these specialized facilities—fueled by intensive training and inference processes on advanced algorithms like Large Language Models (the backbone of ChatGPT and alike)—will demand gigawatts of power rather than merely megawatts in the near future. This surge in demand presents both challenges and opportunities for innovative power solutions, particularly in Alberta, where abundant natural gas reserves, a cold climate, and a supportive regulatory environment create a compelling backdrop for AI data center growth.

Building on the foundational work established in the Future Smart Grids Structures (T06-P01) project under FES’s Grids and Storage (T06) theme, this project—The AI Data Center Power Consumption and Microgrid Operation—treats the data center as a self-contained microgrid with dispatchable potentials. Our goal is to model the high-density AI computing power consumption and integrate diverse energy sources and storage with advanced control strategies into data-center microgrids, ensuring reliability, efficiency, and dynamic load handling. This approach aligns directly with Alberta’s AI Data Centres Strategy and its pillars of Power Capacity, Sustainable Cooling, and Economic Growth by providing scalable, flexible energy solutions that preserve grid stability and affordability.