Electricity prices are climbing. Fast. In PJM Interconnection, the largest power grid operator in the United States, capacity prices just hit the cap: $329 per megawatt-day. That’s a 22% jump from last year and over seven times what the same market cleared two years ago. This isn’t just some abstract number. It adds up to $16.1 billion in capacity procurement—costs that utilities pass on to homes and businesses. A big part of the spike? Data centers. Their energy demand is growing fast, and unlike past industrial loads, they’re not always plugged in with flexibility in mind. **What I’m Thinking About** 1. **Data centers alone caused a 174% jump in PJM capacity prices.** PJM’s Independent Market Monitor said it outright: this isn’t organic demand. The recent surge is largely from massive new data center loads, which are stretching supply and triggering expensive new generation and transmission needs. 2. **Electric bills are already under pressure.** U.S. utilities requested $29 billion in rate hikes in just the first half of 2025. Bain & Co. says serving future data center growth will require utilities to bring in 10% to 19% more revenue each year. That kind of money doesn’t come out of thin air. It comes from ratepayers. 3. **Flexibility is the lever.** The whole system hinges on when—not just how much—power is drawn. Peak demand drives grid costs. If data centers can shift or shave loads during just 35 key hours a year, they can help avoid massive buildouts. Duke University found PJM could add 13 GW of flexible new load without needing more generation. That’s a big deal. **So What?** This is a planning problem, not a tech problem. The tech exists. Google and others already shift loads between regions in real time. What’s missing is how grid planners treat these loads. Today, most assume all new demand is inflexible. That forces the grid to build for worst-case peaks. Instead, regulators should make flexibility a condition of interconnection for large loads. Hold operators to it. Programs like PG&E’s FlexConnect and SPP’s non-firm transmission pilots are steps in the right direction. If we get this right, data centers won’t just strain the grid. They’ll help pay for it. And they might even help bring bills back down 📉. This should also be a wake-up call for Europe, the UK, and countries like Kenya. Data centers are expanding fast everywhere. If grid operators don’t build flexibility into load planning now, they’ll end up with the same problem: rising prices and expensive new infrastructure. European grids already face tight margins in winter. The UK is seeing rapid data center growth in London and the Midlands. Kenya is positioning itself as a regional digital hub. All three need to shift from reactive to proactive planning. Flexible load rules shouldn’t be optional. They should be baked into how new demand connects to the grid from day one.