AI Data Center Demand Strains Power Grids in US and Denmark

The rapid expansion of artificial intelligence infrastructure is creating severe electrical capacity bottlenecks in both Europe and the United States. In Denmark, the state-owned grid operator Energinet implemented a three-month moratorium in March 2026 on all new large-scale grid connection agreements. This pause follows a surge in requests totaling 60 gigawatts—nearly nine times the country's peak electricity demand of 7 gigawatts—driven by hyperscalers like Microsoft, Google, and Apple.

Microsoft had committed $3 billion to Danish data center construction between 2023 and 2027, drawn by the nation's cool climate and renewable energy grid. Energinet is now using the moratorium to develop new prioritization criteria for large energy users, making Denmark the first Nordic nation to restrict grid access due to AI demand.

Simultaneously, the United States is facing similar strains on its power grid and equipment supply chains. Lead times for substation transformers have extended beyond 160 weeks. To address these shortages, the federal government signed four executive orders to accelerate the deployment of nuclear energy to provide reliable baseload power. Private sector responses include Schneider Electric increasing domestic manufacturing and shifting toward prefabricated systems to keep pace with a market projected by Wood Mackenzie to reach $65 billion by 2030.