A shocking number is reverberating throughout the PJM Interconnection territory: $4.3 billion in added home energy costs last year alone, driven to a great extent by the voracious electricity appetite of data centers powering artificial intelligence and cloud computing. According to PJM’s Independent Market Monitor, this surge is almost entirely from existing and projected large-scale data center loads, with Northern Virginia’s “Data Center Alley” at the epicenter.
Goldman Sachs projects 165% growth in energy demand from data centers by 2030, driven by the unprecedented proliferation of AI workloads demanding high-density racks and liquid cooling systems, all of which require uninterruptible power. Global data center power use is currently estimated to be approximately 55 gigawatts, and AI applications are known to already make up 14% of that, while by 2027, AI will account for 27%, further constraining supply and pushing up prices.
PJM predicts peak load growth of 32,000 megawatts between 2024 and 2030 within its 13-state footprint – almost entirely due to data centers. To support this additional capacity, widespread transmission upgrades are needed: more than 150 local projects that launched in states such as Virginia and Ohio between 2022 and 2024 were constructed solely to connect data centers. The Union of Concerned Scientists estimated these transmission expansions added $4.3 billion to consumer bills last year, as regulatory cost reviews remain “practically non-existent” for such facilities.
The financial impact on households is stark. The Natural Resources Defense Council estimates that, by 2028, the average family in the region will pay about $70 more per month for electricity. Cumulative capacity costs could reach $163 billion through 2033 once temporary price caps expire. Average utility bills have already increased by $20 to $30 due to the price increase that we saw in the last auction, and this problem will just get worse through the coming decade if nothing is done, said Claire Lang-Ree of NRDC. She warned that projected demand is outpacing supply, raising the risk of blackouts and scarcity pricing.
To counter this, some tech giants have been taking unconventional steps to secure dedicated power. Microsoft will reopen part of Pennsylvania’s Three Mile Island nuclear site as part of a 20-year agreement with Constellation Energy; Google intends to restart Iowa’s Duane Arnold Energy Center by the end of 2029 in partnership with NextEra Energy. Nuclear projects, touted as carbon-free and reliable, demonstrate what companies are willing to pursue to keep their AI running nonstop. Meta pledged last December that all of its U.S. data centers would be powered by 100 percent renewable energy sourced from a $900 million, 600 MW solar plant in Texas.
For homeowners facing increasing bills, distributed solar has become a hedge against grid volatility. Leasing programs-most notably Palmetto’s LightReach-eliminate upfront costs and let residents pay a fixed monthly fee while using all the electricity generated on site. Rooftop arrays paired with battery backups, such as Tesla’s Powerwall, are able to store excess generation for use during peak pricing or outages, reducing reliance on the grid. Advances in solar panel efficiency and installation have lowered costs and increased output. Meanwhile, grid-scale battery storage has emerged as the fastest-build option for integrating renewable energy and meeting sudden spikes in demand.
Consumers for a Better Grid experts believe that data centers “should bring their own capacity” through on-site generation and storage, adopt energy efficiency measures, and better load forecasting in order to avoid overpaying. “Unless something is done, everyday people will be left holding the check for some of the wealthiest companies in the world, and that’s unacceptable,” said Clara Summers of the Citizens Utility Board of Illinois.
The engineering problem is thus twofold: one of expanding transmission infrastructure to meet unprecedented loads, while accelerating deployment of clean, fast-to-build resources. As AI’s energy footprint grows and utility-scale projects face multi-year permitting delays, the balance between technological progress and grid sustainability will define both household budgets and the pace of the renewable transition in the decade ahead.
