Concerns are rising that the explosive growth of data centers, fueled by AI and cloud computing, will drive up electricity prices for homes and businesses. Headlines warn of blackouts, higher bills, and strained grids. Yet two recent, rigorous analyses show the opposite: data centers have not raised national average electricity prices — and may even have contributed to modestly lower rates.
The Economic Logic Behind Stable or Lower Prices
Electricity is unlike most commodities. Power systems have enormous fixed costs — power plants, transmission lines, substations, and distribution networks — that must be paid regardless of how much electricity flows through them. When durable new demand arrives (like from data centers that run 24/7), these fixed costs get spread across more kilowatt-hours. This naturally lowers the average cost per unit of electricity.
Utilities and regulators typically set retail prices based on average costs, not short-term marginal costs. Adding load also unlocks economies of scale in transmission, distribution, and generation. New capacity built to serve growing demand is often cheaper per kilowatt-hour than the embedded costs of older infrastructure.
This isn’t theoretical. Historical U.S. periods of strong electricity demand growth (such as the 1980s–2000s) coincided with declining real (inflation-adjusted) household electricity prices. Recent demand growth since 2020 has been slower than mid-20th-century rates, giving supply time to respond.
Real-World Evidence: No National Price Increase
As of 2025, data centers had not raised national average household electricity prices. Nominal price increases since 2019 were driven primarily by inflation, surging natural gas prices (exacerbated by Russia’s 2022 invasion of Ukraine), and other supply-side factors — not data center load.
A Lawrence Berkeley National Laboratory analysis (covering 2019–2024) found that states experiencing the highest load growth — including from data centers, manufacturing, and electrification — actually saw lower average retail electricity prices after inflation adjustment. States with shrinking loads saw prices rise. Data centers accounted for roughly 40% of U.S. electricity demand growth from 2020–2025, yet national prices remained largely flat in real terms until external shocks hit.
Causal Evidence: Data Centers Actually Lowered Prices (2015–2024)
The most rigorous study to date uses an instrumental variables approach to establish causality. Researchers Asa Watten, John Bistline, and Geoffrey Blanford analyzed U.S. data from 2015 to 2024 and concluded that data centers caused average retail electricity rates to fall modestly.
Key finding: A 10% increase in data center capacity was associated with approximately a 0.4% reduction in average residential retail electricity prices (in their preferred specification). Over the period, data centers added 21.3 GW of capacity. Despite data centers consuming 4.5% of U.S. electricity by 2024 (double the share from five years earlier), prices did not rise — they edged lower.
The study documents clear economies of scale across transmission & distribution costs, generation costs, and within customer classes. It rules out widespread cross-subsidization between customer types. The causal identification uses the 1947 Eisenhower Interstate Highway System as an instrument (exogenous to modern electricity prices but predictive of data center locations via fiber infrastructure).
The result aligns with basic economics: when average costs exceed marginal costs of new supply (common in the power sector with slack or expandable capacity), durable demand growth lowers average prices.
Why Local Concerns Don’t Translate to National Price Hikes
Some localities (e.g., parts of Northern Virginia) have seen faster rate increases alongside heavy data center concentration. However, even there, increases were often below or in line with national trends when accounting for other factors. Data centers tend to locate in areas with existing or expandable grid capacity and low-cost power, minimizing their system-wide impact.
Nationally, 11 of 15 states with significant data center buildouts experienced below-average rate increases compared to states with minimal data center activity.
Important Caveats — And Why They Don’t Change the Picture
The studies note that future supply constraints (transformer shortages, permitting delays, gas turbine backlogs, etc.) could temporarily reverse the effect if new generation and transmission can’t keep pace. Overbuilding capacity ahead of confirmed load also carries risks. Local distribution upgrades may still raise costs in specific areas.
However, these are supply-side challenges, not inherent flaws in demand growth from data centers. Markets and regulators have strong incentives to expand supply efficiently, especially with durable, high-value load like data centers.
The Bottom Line
Data centers represent productive, high-utilization demand that helps amortize the power system’s large fixed costs. Rigorous evidence from 2015–2025 shows they have not raised national electricity prices — and causal analysis indicates they modestly lowered them through economies of scale.
This dynamic supports broader electrification (EVs, heat pumps, industry) as well. With continued investment in generation and transmission, data center growth can remain compatible with stable or declining real electricity prices for everyone.
The prevailing narrative that data centers must drive up bills lacks support in the data. Productive demand growth, when met with responsive supply, strengthens the economics of the grid rather than undermining it.
