---
title: "Air conditioning, data centers, and why your power bill is rising faster than inflation"
description: "US residential electricity prices rose 11.5% in 2025, against headline inflation of 2.9%. The cause is not the total amount of power consumed but when it is consumed: cooling and data centers both peak on the same hot afternoons, and peak capacity is the most expensive thing a grid buys."
category: "Economy"
category_url: https://boursel.com/category/economy
author: "Sofia Marchetti"
published: 2026-07-20T13:18:00.000Z
updated: 2026-07-20T13:18:00.000Z
canonical: https://boursel.com/article/air-conditioning-data-centers-and-why-your-power-bill-is-rising-faster-than-infl
tags: ["electricity", "utilities", "data-centers", "inflation"]
---
# Air conditioning, data centers, and why your power bill is rising faster than inflation

US residential electricity prices rose 11.5% in 2025, against headline inflation of 2.9%. The cause is not the total amount of power consumed but when it is consumed: cooling and data centers both peak on the same hot afternoons, and peak capacity is the most expensive thing a grid buys.

Electricity is behaving unlike the rest of the consumer price basket. Understanding why requires setting aside the question of how much power gets used and looking instead at the hour it gets used in.

## Cooling is a large load, and a badly timed one

Air conditioning accounts for about [19% of electricity used in US homes](https://www.eia.gov/tools/faqs/faq.php?id=1174&t=1), roughly 254 billion kilowatt-hours in 2020, according to the Energy Information Administration. Commercial building cooling took about 14%, or 170 billion kWh, in 2018.

Those are large shares but not alarming ones. The problem is the shape of the demand rather than its size. Cooling load is **peak-coincident**: it arrives when it is hot, which is when it is hot for everyone at once, in the late afternoon and early evening. California illustrates the swing, running near 25,000 megawatts at the 4am minimum and around 42,000 megawatts at the 6pm peak, a rise of about 68% within a single day, driven substantially by air conditioning. That peak also lands as solar output falls away.

This is the crux of the economics. A grid must be built to serve its highest hour, not its average one. Generation, transmission and distribution capacity sized for a handful of extreme afternoons sits underused for most of the year, and someone pays for it in every month's bill.

The International Energy Agency's [Future of Cooling](https://www.iea.org/reports/the-future-of-cooling) analysis projects that, absent stronger efficiency measures, energy demand for space cooling would more than triple by 2050.

## Two peak loads arriving together

Cooling alone would be a manageable, well-understood problem. What has changed is that data-center load is growing in the same regions, and it does not switch off in the evening.

The North American Electric Reliability Corporation's [2026 summer assessment](https://www.nerc.com/globalassets/our-work/assessments/2026-summer-reliability-assessment-snapshot.pdf) records more than 58 GW of new generating capacity added since summer 2025, comprising 16.4 GW of solar, 14.7 GW of battery storage, 6.7 GW of natural gas and 1.6 GW of wind. That is a substantial build.

Over the same period summer peak demand rose by 11 GW, following a 10 GW increase the year before. Supply is being added quickly and demand is keeping pace.

Data centers accounted for about 4.4% of US electricity in 2023, and estimates put their share at roughly 6.7% to 12% by 2028.

## What the capacity market reveals

The cleanest evidence sits in PJM Interconnection's capacity auctions, which pay generators to be available at peak. The clearing prices are stark:

| Delivery year | Clearing price ($/MW-day) |
|---|---|
| 2024/25 | $28.92 |
| 2025/26 | $269.92 |
| 2026/27 | $329.17 (at the cap) |
| 2027/28 | $333.44 (at the cap) |

Two features deserve attention. The first is the near tenfold jump between 2024/25 and 2025/26. The second, and more telling, is that the last two auctions cleared **at their administrative price cap**. A market clearing at its ceiling is one where the ceiling, not supply and demand, set the price. Absent the cap, it would have gone higher.

The independent market monitor has quantified the data-center contribution directly. Monitoring Analytics found that including 11,993 MW of existing and forecast data-center load in the peak forecast produced capacity market revenues of about $19.3 billion. Excluding the forecast new data-center load and the forecast increase in existing load, some 7,892 MW, would have produced about $8.9 billion. The difference is roughly **$7.3 billion, or 82.1%** of the total.

Looking ahead, forecast peak load for 2027/28 is about 5,250 MW higher than the forecast used for the previous auction, and nearly 5,100 MW of that increase is attributable to data centers.

Capacity costs are recovered from all customers in the zone. A household that has added no load and changed no behavior pays a share of a capacity price set largely by someone else's data center.

## Where it lands

Residential electricity prices rose 11.5% in 2025, against headline inflation of 2.9%. Utilities requested more than $29 billion in rate increases in the first half of 2025, roughly double the first half of 2024.

The forward spending is larger still. S&P Global Market Intelligence puts planned capital expenditure by US investor-owned utilities at [approximately $1.3 trillion across 2026 to 2030](https://www.spglobal.com/market-intelligence/en/news-insights/research/2026/04/surging-energy-demand-puts-us-utility-capex-forecast-near-1-3t-in-2026-30). Utility capex is recovered from ratepayers over time with an allowed return, so that figure is a reasonable forward indicator of the direction of bills.

## How to read this as an investor or a household

Three observations follow, and none is a recommendation.

**Electricity is becoming a rate-of-inflation story in its own right.** A component rising at four times the headline rate, driven by infrastructure commitments spanning decades, is not a transitory price move.

**Regulated utilities are in an unusual position.** Large capital programs, recovered through rates with an allowed return, are the core of the regulated utility model. Whether that translates into returns depends on regulators permitting full recovery, which is a political question as bills rise.

**The concentration matters more than the total.** These costs are not spread evenly. They fall hardest in the zones where data centers cluster, which is why some regions are seeing bill increases far above national averages while others see very little.

The underlying mechanism is a single sentence: two large loads that peak at the same time force the construction of capacity used for only a few hours a year, and everyone connected to the wire pays for it.