I have written before that winning the race for artificial intelligence leadership requires building power faster than we have in a generation. That challenge remains. The more important question is whether we can keep electricity prices affordable for American households and globally competitive for the businesses that create jobs.
The good news is that we can. With the right mix of innovation, operational reform, and smart policy, rapid growth in data centers does not have to raise electricity bills for everyday consumers. Done right, it can help lower them.
Demand Growth Meets Affordability Pressure
After two decades of flat demand, U.S. electricity use is now growing nearly 3% annually, with data centers projected to drive about 70% of that growth over the next decade. By 2030, they could consume more than 10% of all U.S. electricity.
Meanwhile, the grid itself is aging rapidly. More than half of U.S. distribution transformers are nearing the end of their useful lives, and over 70% of large power transformers exceed their design age. Replacing this infrastructure is essential but costly, and those costs ultimately flow to all customers.
Building new generation has also become more expensive. Since 2020, capital costs for gas-fired power plants have nearly doubled due to labor shortages, supply-chain disruptions, and material inflation. Renewables remain competitive, but they too face higher equipment costs. Without a new approach, rising demand combined with infrastructure renewal risks pushing electricity costs higher.
Smarter Use of Existing Assets
New generation and transmission will be necessary, but relying solely on greenfield construction is too slow and too expensive. The electric grid’s utilization has been in a 25 year decline recently falling below 70% because many power plants and transmission lines run below technical limits due to outdated limitations. Unlocking this stranded capacity can deliver cost relief far faster than new builds.
Battery storage is central to this effort. Batteries smooth peaks and valleys in demand allowing power plants to operate closer to optimal conditions, reducing wear, lowering costs, and increasing power output. On transmission systems, strategically placed batteries allow existing lines to run closer to full utilization, helping avoid costly new corridors. Using the grid more efficiently should be the first line of defense for affordability.
Data Centers as Part of the Solution
Contrary to popular belief, large data-center loads do not automatically raise household rates. Much of the cost of running the electric grid is fixed and does not change much whether demand is high or low.
When large customers join, and pay for the grid upgrades they require, those existing costs are spread across more electricity consumption. In other words, the same grid costs are shared more broadly, easing pressure on smaller customers. This is often the simplest and fastest way to lower costs for everyone.
We are already seeing this effect. At AES utilities in Indiana and Ohio, growing large-load demand is increasing revenue to help pay for the shared system. AES Indiana’s latest Integrated Resource Plan shows that higher data-center demand lowers the levelized supply cost per kilowatt-hour for customers over both 10- and 25-year planning horizons.
Data centers can do more to support the grid by deploying behind-the-meter batteries to smooth highly variable AI workloads and by shaving their demand during grid emergencies, reducing system stress and avoiding costly upgrades that would otherwise burden all customers.
Policy Choices Matter
Operational innovation alone will not solve affordability. Policy choices matter, especially when it comes to equipment and materials costs. Roughly 50–60% of critical power-system equipment is imported, much of it from allies such as Mexico, Canada, and South Korea, and essential inputs like copper mostly from Chile. Higher tariffs on this equipment raise modernization costs and, ultimately, customer bills.
Targeted tariff relief for trusted allies on electric equipment and critical minerals like copper could materially reduce electricity costs without new federal spending. Similar exemptions already support AI chip and semi-conductor imports. Power infrastructure deserves the same pragmatic approach.
Smart policy should also give energy companies the confidence to invest for the long term to lower costs. Power plants are 30-year investments, so when markets frequently change pricing rules or cap payments, as happens in some parts of the country, it discourages exactly the infrastructure we need. Stable, long-term market design reduces risk, lowers financing costs, and accelerates the build-out required to keep electricity reliable and affordable.
Innovation, Not Trade-Offs
America does not have to choose between AI leadership and affordable electricity. That is a false trade-off.
If we operate the grid more intelligently, fully deploy battery storage, ensure large customers pay their share, and reduce unnecessary cost adders, we can meet rising demand while protecting and even improving affordability.
Getting this right will shape where factories and data centers are built and determine whether American businesses can compete globally while creating good-paying jobs at home.
