Americans are noticing a hit to their wallets this summer as electricity bills climb, with independent analysts forecasting an increase of around 30-60% by 2030, largely driven by the demand from artificial intelligence data centers. It’s time to address this issue.
Recent capacity auctions have shown a dramatic surge in liquidation prices—over ten times the norm in just two years—due to data-driven demand, which now represents more than 90% of the load growth in grid systems. Utilities face the daunting challenge of meeting this historic demand, which will require billions in new infrastructure. Unfortunately, consumers are likely to bear these increased costs, resulting in higher electricity rates.
While AI itself isn’t necessarily a problem, the 2009 federal climate policy poses significant hurdles. The EPA’s finding from that year determined that greenhouse gases, particularly carbon dioxide, threaten public health and welfare, which has led to strict regulations concerning older coal units and many new gas plants. This has delayed the adoption of affordable and reliable energy sources needed to match rising demand. Reversing this finding could be a crucial step for policymakers looking to avoid a crisis in electricity costs tied to the AI boom.
Hidden extra charges: EPA findings
Why can’t utilities tap into cheaper energy sources to meet the surging demand? The 2009 danger finding provides the legal backdrop for many EPA regulations. These include mandates for Carbon Capture and Storage (CCS) and set retirement timelines for fossil fuel power plants. Legal experts characterize this “danger finding” as a significant barrier to federal climate regulations. If dismantled, the complex networks governing emissions could unravel.
In fact, the EPA estimates that repealing current standards could save the electricity sector about $19 billion over two decades, beginning in 2026—roughly $1.2 billion per year. Recently, the EPA proposed overturning its danger findings, anticipating it could save Americans over $54 billion annually by uncoupling from the chain of climate regulations it established. The removal of only the CCS requirements could cut costs significantly as well.
How that discovery affects your bill
According to the EPA, the mandatory CCS infrastructure can drive up capital costs for combined cycle gas plants to between $800 and $1,600 per kilowatt. This translates to an added expense of around 1.5 cents per kilowatt-hour at retail prices. Additionally, the compliance deadlines tied to these rules could force 60 gigawatts of active coal and aging gas capacity offline before 2035, which is when AI demand is projected to peak, impacting household budgets even before any actual plant closures occur.
Dispatchable fuel remains essential in America.
Natural gas is a pivotal player in the US energy sector, supplying 43% of the nation’s electricity last year, supported by a 57% capacity utilization rate. This highlights the importance of gas plants, especially under grid stress. Meanwhile, coal still contributes 16% of electricity generation and serves as a dependable on-site fuel source during extreme weather. However, federal policies are advancing faster than coal’s replacement, causing grid operators to depend on favorable weather while juggling unprecedented demand.
Even in Silicon Valley, there’s acknowledgment of the approaching crisis. Google recently made the decision to halt non-essential AI workloads in Indiana and Tennessee to reduce grid stress, reflecting concerns that current infrastructure is inadequate for ambitious energy goals. It raises a question: why should everyday families shoulder the costs of upgrades when even leading tech companies can’t guarantee the power supply for their AI projects?
Abolishing the danger finding is a relief for consumers
Eliminating the danger finding could offer much-needed relief for consumers. This isn’t about deregulating AI; data centers would still pay market rates for electricity. The goal here is to allow utilities to opt for the most efficient and cost-effective technologies without the burden of additional costs. This would lead to significant reductions in electricity production costs, enabling utilities to run more efficiently and pass those savings on to consumers.
While proposals to remove the EPA’s findings are likely to face legal and political challenges, lawmakers don’t have to be passive. They could take action to formalize the repeal, prevent the reimplementation of CCS requirements through indirect means, and restore essential principles. Energy policy should prioritize the interests of those who pay the bills, not the industries that create them.
The US can support the AI revolution while also keeping household costs in check. It all begins with removing a regulatory obstacle—a danger finding—that will make way for more affordable electricity.
