Electricity bills

A data center is one of the largest paying customers a utility can sign, and utilities are passing that revenue on. Indiana Michigan Power has proposed cutting Indiana bills by about $59 million in 2027—about $100 a year for a home using 1,000 kilowatt-hours a month—and holding residential rates steady for three years. The utility says the plan is “supported by load growth and increased revenue from large customers, including data centers” (T&D World). The Indiana commission expects to decide in June 2027.

In Kansas, Evergy—the utility serving Lawrence—now has approved rules for customers of 75 megawatts or more: up to five years to ramp up, then at least 12 years of service, with a minimum monthly bill of 80% of the power they contract for. Evergy, the commission's staff, the Citizens' Utility Ratepayer Board, Google, the Sierra Club and others signed the settlement the Kansas Corporation Commission approved on November 6, 2025 (Utility Dive). Evergy says it uses that revenue “to help keep prices and rates lower for our existing customers” (KSHB).

The computers keep getting more efficient, too. Chip performance per watt “has improved annually for decades,” and large operators now report fleetwide efficiency ratings (PUE) of 1.1 to 1.4, where 1.9 or above was common a decade ago (JLARC). A PUE of 1.1 means 10% extra power for cooling and the building beyond the computers themselves.

A data center is a paying customer of the electricity system. Its purchases help cover the wires, substations and other costs shared with homes and businesses. A durable customer gives a utility revenue to plan around and spreads costs across more electricity sales.

How a new customer helps pay for the grid

Some utility costs remain whether the system sells a little electricity or a lot. Existing infrastructure needs maintenance and financing. When a new customer contributes enough revenue to those shared costs, there can be less for other customers to pay.

Serving it may also require new substations, transmission lines or generation. The utility's rate agreements determine how those investments and ongoing costs are allocated. The useful comparison is between the cost of serving the customer and the revenue it brings—not electricity use alone.

PG&E estimates that each additional gigawatt of data-center demand could reduce monthly bills by 1–2% over the long term.

An Amazon-commissioned E3 study compares selected sites' utility payments with their modeled costs to serve. It finds that the payments meet or exceed those costs. Surplus revenue can help offset costs for other customers; rate allocation determines how that value reaches them.

What the historical research found

In their 2026 preprint, Asa Watten, John Bistline and Geoffrey Blanford estimate that data-center growth modestly reduced average U.S. retail electricity rates over 2015–2024. The result fits the shared-cost mechanism: more durable demand can help pay for a system with substantial existing fixed costs.

A Rutgers State Policy Lab analysis studied household bills and data-center openings through 2024. At the utility-territory level, it found no statistically significant effect. The household-weighted estimate was about $2 lower per month, but it was not statistically distinguishable from zero. The analysis did not establish a clear increase or a demonstrated saving.

These studies examine different measures. A rate is the price charged for electricity; a household bill also depends on how much electricity the household uses. Dates and the utility service area help identify which comparison answers the reader's question.

Nic Carter's state-price analysis shows that the states with the most data-center activity have the lowest residential rates, and, adjusted for inflation, the smallest rate increases since 2022. Developers also have a reason to choose cheap power, so the comparison describes a pattern rather than establishing its cause.

Virginia: connect the finding to its date

The CloudHQ-sponsored E3 backcast models Dominion Virginia's 2023 residential rates with and without incremental data-center growth. It finds that the added customers helped cover shared costs and put downward pressure on modeled residential rates.

Virginia's December 2024 JLARC report found that the rates it examined appropriately allocated costs to data centers. Its forward-looking scenarios projected roughly $14–$37 more per month in residential generation and transmission charges by 2040. That estimate uses constant 2024 dollars and monthly consumption of 1,000 kWh; it excludes distribution costs. E3 supplied modeling for both reports, addressing different questions.

A historical contribution to shared costs and a forecast of future investment can coexist. The amount of infrastructure built, its cost and the way customers pay for it determine the next result.

The EIA price observation adds another date to the record. In its September 24, 2026 release, Virginia's preliminary residential average was 17.55 cents per kWh in July 2026, compared with 15.78 a year earlier—about 11.2% higher. Those are statewide prices per kilowatt-hour. A utility's rate-case documents explain which costs enter its rates; a household's usage helps explain the bill it receives.

Put the customer's commitment in the agreement

Long-term contracts make expected demand a firmer basis for investment. Minimum payments can assign the cost of reserved capacity to the customer, alongside charges for electricity actually used. Exit terms specify the customer's obligations if it leaves before the contract ends.

The Virginia SCC's November 25, 2025 order approved Dominion's GS-5 class, effective January 1, 2027. For qualifying customers taking initial service and signing an agreement from that date, it provides 14-year contracts, minimum demand charges and early-exit obligations. The class has demand and load-factor thresholds, and the order includes exemptions.

An approved contract structure is a concrete part of how the opportunity is organized. Its effective date, eligible customers and payment terms belong alongside any forecast of household savings.

Understanding a local proposal

Start with the utility serving the site and the public record for the proposed connection. The county's planning staff can identify local approvals; the utility and its regulator can explain service agreements and rates.

Five questions help connect the project to the household bill:

• What existing infrastructure will serve the customer, and what new capacity is planned?
• Which costs does the company pay directly, and which are allocated through utility rates?
• What demand and minimum-payment commitments are binding, including any early-exit terms?
• When do the approved terms take effect, and which customers qualify?
• Where can residents compare the projected contributions with actual rates and payments once the project operates?

The documents to look for are the applicable large-load tariff, the regulatory docket, upgrade cost allocations and the disclosed contract terms. Together, they show how a new customer's investment and electricity payments contribute to the system used by the wider community.