Powering the digital age: data centres, AI, and the future of grid stability

Grid Meets AI Demand

The surge in artificial intelligence has transformed data centres from steady background loads into one of the fastest-growing drivers of electricity demand. In the UK these facilities already consume roughly 2% of total supply, concentrated in latency-sensitive clusters around London and Slough, but projections show demand rising fivefold to 26 terawatt-hours by 2030 according to Oxford Economics — more than the annual usage of 9 million households.

This growth collides with a transmission and distribution system ill-prepared for the pace. The National Energy System Operator's connection queue ballooned from 41 gigawatts in November 2024 to 125 gigawatts by June 2025, half attributable to data centres; peak GB demand is only 45 gigawatts. Resulting waits stretch to a decade in some cases, prompting Ofgem to launch a major demand-connections overhaul in recent weeks that explicitly prioritises credible projects and weeds out speculative ones.

Concrete consequences are already visible. Developers are delaying or relocating builds, threatening the economic value tied to the UK's digital economy and government-designated AI Growth Zones. Grid upgrades will require tens of billions in new transmission investment, costs that could flow through to household and business bills if not offset by flexibility from the new loads themselves. In extreme scenarios, concentrated demand risks local voltage issues or reliability shortfalls, echoing isolated incidents seen elsewhere.

Less discussed are the countervailing opportunities and trade-offs. Many AI training tasks can pause, shift hours or curtail without loss, allowing data centres to act as demand-side resources that help balance wind and solar variability — particularly if sited near curtailed Scottish wind farms as NESO's chief recently suggested. Operators are also exploring behind-the-meter generation and nuclear co-location to bypass queues, but this pits the need for rapid deployment against net-zero commitments and raises questions over interim fossil reliance. High British industrial electricity prices, the OECD's highest, further tilt decisions toward power-rich but remote sites over traditional hubs.