Electricity prices are rising quickly, and data centers are a big part of the reason why. PJM Interconnection, which manages the power grid for 65 million Americans, just hit maximum capacity prices of $329 per megawatt-day. That's seven times higher than two years ago and adds up to $16 billion in costs that utilities pass on to customers. #### **What's Happening** Data centers are the main driver behind a 174% jump in capacity prices. PJM's market monitor has pointed out that this isn't typical demand growth. It's largely from new data centers that need massive amounts of power and aren't designed to be flexible about when they use it. At the same time, utilities requested $29 billion in rate increases in just the first half of 2025. Research from Bain suggests that meeting future data center demand will require utilities to increase revenue by 10-19% annually. Those costs get passed to customers through higher bills. #### **The Flexibility Solution** The key issue isn't how much power data centers use, but when they use it. Peak demand drives most grid costs, and the system gets built to handle maybe 35 critical hours per year. If data centers could shift some of their operations during those peak times, it would help avoid expensive new infrastructure. Duke University found that PJM could add 13 GW of flexible demand without needing new generation capacity. Companies like Google already do this by moving computing tasks between regions based on grid conditions and power prices. The technology exists. #### **What Needs to Change** The problem is in how grid planners currently treat new demand. Most assume all new loads are inflexible, which forces the system to build for worst-case scenarios. A better approach would require large new loads to demonstrate flexibility as a condition of connecting to the grid. Some utilities are testing this approach. PG&E has FlexConnect, and the Southwest Power Pool is piloting non-firm transmission options. #### **Beyond the US** Other regions should pay attention to this issue. Europe faces tight grid margins in winter. The UK is seeing rapid data center growth. Kenya is developing as a regional digital hub. If these areas don't build flexibility requirements into their grid planning now, they'll face the same problem: rising prices and expensive new infrastructure needs. Done right, data centers could help pay for grid improvements rather than just driving up costs. But that requires proactive planning rather than reactive responses. Ref : [[Data Centre Energy Demand]] [[How AI Will Impact Energy Demand]] [[The infrastructure layer and AI capex]]