AI-driven energy demand: planning infrastructure for an accelerating growth curve

The rapid growth of AI, particularly large Global energy investment language models and compute-intensive workloads, is making data centres one of the fastest-growing sources of electricity demand. The IEA estimates global data centre electricity demand increased by around 17% in 2025, adding approximately 70 TWh of additional electricity consumption in a single year, prompting revisions to forecasts, grid connection queues and generation plans across multiple regions. For energy planners, the challenge is threefold: ensuring sufficient capacity, locating new loads without overwhelming local networks and aligning demand growth with decarbonisation commitments. Core to this how the energy sector will fuel the future of AI and develop the next generation tools critical to improving efficiency and supporting data centre expansion. New commercial models and partnerships between the energy sector, hyperscalers, investors and policymakers are emerging, but the pace of grid development has not yet matched the pace of AI investment. Can energy infrastructure scale fast enough to meet AI's ambitions, or will energy demand become the ultimate constraint on digital growth?