TL;DR
- 19.49 ¢/kWh: Ohio residential electricity prices in April 2026 (EIA), the highest reading in the three-year window we track and the steepest household move among the PJM mid-Atlantic states in our set.
- Up 19.4% year over year, from 16.32 ¢/kWh in April 2025 (EIA). That’s a change in the rate, not a bill total.
- 8.66 ¢/kWh: the U.S. industrial electricity rate in April 2026 (EIA), the tier relevant to hyperscale facilities.
- 18.7%: the year-over-year rise in private fixed investment in information processing equipment and software, Q1 2026 (BEA). Economy-wide capital deployment, not a data-center-specific measure.
EIA data through April 2026 (EIA); May 2026 (Henry Hub); Q1 2026 (BEA)
What happened to residential electricity prices in Ohio?
Ohio households were at 19.49 ¢/kWh in April 2026, up 19.4% from a year earlier. That is the steepest household move of the mid-Atlantic states we track, and the level itself is the highest reading in our window. Both halves of that sentence matter and they are different kinds of statement: one is where the price sits, the other is how far it traveled.
We’re spending most of this brief on Ohio because it’s the sharpest single case, and because a lot of the coverage of residential electricity prices in this footprint has been written as a regional mood rather than a set of state-level numbers. The rest of the footprint is in the report. What we’ll say here is that no state in the set moved the way Ohio moved, and that every one of them moved.
A double-digit annual increase in residential electricity prices is not a rounding error to the household absorbing it, and it isn’t a rounding error to the commissioner who has to explain it.
That’s the whole reason we pulled this data in the first place. The AI infrastructure conversation is conducted almost entirely in megawatts and capex, and the number that actually determines whether a region stays politically hospitable to that buildout is the one printed on a household bill.
Does the industrial tier look like the household tier?
No. The U.S. industrial electricity rate sat at 8.66 ¢/kWh in April 2026. Ohio’s residential rate sat at 19.49 ¢/kWh in the same month. Those are two levels in the same unit at the same moment, and they are not close.
The industrial tier is the one that maps, loosely, to the class of customer a hyperscale facility falls into, subject to how any given power purchase agreement is actually structured. We are not going to pretend the retail industrial average is a clean read on what a specific campus pays. It isn’t, and the report says so at more length than a brief can carry.
What we will say is that the gap between how the household tier behaved and how the industrial tier behaved over the past year is the single most consequential shape in this dataset, and it’s the reason the report exists rather than a chart tweet. The direction of that divergence, its size, and what it does and does not license you to conclude are all in there.
Who does this actually land on?
It lands, first, on state utility commissions and consumer advocates in five states who are going to spend the next several rate cases explaining a number they did not set. Residential electricity prices are the most politically legible energy statistic in existence. Nobody has an intuition for a capacity auction clearing price. Everybody has an intuition for a bill that went up by a fifth.
It lands on retail suppliers and utility regulatory affairs teams, who will be asked to reconcile a household number with an industrial number in front of an audience that did not come to hear about tariff classes.
It lands on economic development offices. There is a version of the next three years in which a governor’s office is negotiating a large-load interconnection while simultaneously fielding constituent mail about residential electricity prices, and the two conversations are happening in the same building on the same afternoon. Our read is that those offices have not yet built the analytical vocabulary to hold both at once, and the ones that do will have a real advantage in siting negotiations.
It lands on hyperscale site selection teams, though not in the way most people assume. The risk to a campus in this footprint is not primarily the industrial rate. It’s the possibility that residential electricity prices become the organizing grievance of a local politics that then reaches the permitting process. That’s a second-order exposure, and it’s the kind that doesn’t show up in a pro forma until it’s already a delay.
And it lands on households, obviously, including a lot of them in places that have no data center and never will. The mid-Atlantic grid does not sort its cost recovery by county sentiment.
What does it change for the people holding the bag?
It changes what “cheap power” means as a siting argument. For most of the last decade the pitch to a region was that large load brings ratepayer benefit, and that pitch was made in an environment where residential electricity prices were boring. They are no longer boring in this footprint. Any argument made to a commission in 2026 has to survive contact with a household number that moved by double digits, and we think a lot of the standard decks have not been rewritten for that.
It changes the shape of tariff design fights. When the household tier and the industrial tier separate visibly, the policy response tends to be structural rather than incremental: new large-load classes, minimum-take provisions, collateral requirements, ratepayer-protection language written into interconnection agreements. Whether that response is warranted by what the data actually shows is a separate question, and it’s exactly the question the report takes up.
It changes the risk profile of the buildout narrative for anyone underwriting it. The consensus story is that AI infrastructure is constrained by turbines, transformers and interconnection queues. Our read is that the binding constraint in a handful of these states is going to be political tolerance, and political tolerance runs on residential electricity prices, not on queue position.
It also changes what a reporter should ask next. “Is AI raising my bill?” is the wrong shape of question for this data to answer. “What can these series establish, and where do they stop?” is the right one, and the honest answer is narrower than most headlines have been.
Why might the obvious read be wrong?
Because the obvious read is that data center load pushed residential electricity prices up, and this dataset does not establish that.
We want to be precise about what we’re saying, since it would be easy to read that as a dodge. Retail price series and state generation series are real measurements of real things. They are not attribution machines. There are several plausible mechanisms sitting behind a double-digit household move, some of them have nothing to do with AI, and at least one important price signal in this market does not appear in the pulled data at all. The report names which mechanisms are in play, walks through what each series can and cannot carry, and is explicit about the one number that would settle the argument and is not available.
The finding here isn’t that we know why residential electricity prices moved. It’s that the household tier and the industrial tier did not move alike, and that is a fact worth having before anyone builds a causal story on top of it.
The fuel-cost side of this is where the analysis gets genuinely interesting, and it’s also where a brief has no business going. There is a relationship in the data between the gas benchmark and the industrial rate, it does not sit where a naive read would put it, and working out what that implies for the household series took most of the report’s middle section. That reasoning is the product. We’re not going to hand it over in a summary and then ask you to pay for it.
Is the capital cycle telling the same story?
Separately, private fixed investment in information processing equipment and software rose 18.7% year over year in Q1 2026. That’s a national, economy-wide figure covering far more than data centers, and we’re including it as backdrop rather than as evidence about any grid.
Still, it’s the number that keeps this whole subject from being a regional utility story. Capital is going into computing equipment at a pace that is visible in national accounts data, and the electricity to run that equipment has to be delivered somewhere by someone at a price somebody pays. Residential electricity prices in the PJM mid-Atlantic core are one of the places where that abstraction stops being abstract. Whether the two facts are connected in the way everyone assumes is, again, the report’s problem, not the brief’s.
You can pull the retail price series yourself from the EIA Electricity Data Browser if you want to see the shape before you read our read on it. We’d encourage that. The numbers in this piece are all public.
Where the rest of this lives
The brief gave you one state, one national tier and one capital figure. The report gives you the rest: the state-by-state residential electricity prices breakdown across all five mid-Atlantic states, the household-versus-industrial split worked out rather than gestured at, the generation-side picture including which states are producing more and which are flat, Virginia’s position as a demand center that leans on the wider system, the natural gas fuel-cost variable and the lag relationship we found in it, and a plain statement of the caveats, including the price signal that is missing from the data entirely and what its absence rules out.
It also carries the forward view, which is the part we deliberately keep out of the brief, and the investor and policy implications section that follows from it.
Read the full report on PJM residential electricity prices and the household-industrial split