Electricity Prices 2026: 18.17 Cents, a Capped Auction Still Short, and the Crossover Year

The average US home pays 18.17 cents a kilowatt-hour in 2026 against 17.24 last year, about $100 more on a typical $1,925 annual bill, with EIA forecasting 18.68 cents for 2027, and the year’s defining question is how much of that is the AI buildout. The measurable answers, from PJM’s own market monitor: data-center load added $9.95 per megawatt-hour, 9.8 percent, to wholesale costs that rose 40.9 percent January-September, and $29.4 billion to capacity-market revenues across the last four auctions, the most recent of which cleared at its $325/MW-day price cap and still came up 6,831 megawatts short of the reliability target. The next auction prices in December with a higher cap, and two milestones frame the era: commercial electricity sales are set to pass residential for the first time, and Virginia’s data centers already draw 26 percent of their state’s supply.

⚡Electricity Price Tracker
Updated: 15-21 November 2026
US residential average, 2025 → 2026 → 2027f17.24 → 18.17 → 18.68¢/kWh
Grade / ItemLowHighUnit
Typical annual bill (10,597 kWh)1,9251,925USD, ~+100 YoY
PJM wholesale Jan-Sep79.19111.61USD/MWh, +40.9%
Data-center share (monitor)9.959.95USD/MWh = 9.8%
2028-29 capacity auction325325USD/MW-day, at cap, 6,831 MW short
EIA and Monitoring Analytics via compilations of 15 Nov 2026. Full scrap price index

The tape, meter to market

Series Level Context
Residential price (EIA) 18.17¢/kWh 2026 17.24¢ 2025; 18.68¢ forecast 2027
Typical annual bill ~$1,925 (~$160/mo) ~+$100/yr at 10,597 kWh use (our arithmetic)
PJM wholesale cost $111.61/MWh Jan-Sep $79.19 year earlier: +40.9%
Data-center share of that $9.95/MWh (9.8%) market monitor attribution, capacity market alone
Capacity revenue added by DC load $29.4B across last four PJM auctions
2028-29 auction (Jul 2026) cleared at $325/MW-day cap; $16.4B 6,831 MW short of reliability standard
Next auction (Dec 2026) cap $337.83 +$5-10B projected; peak-load forecast +3.6%
US data-center use 183 TWh 2024 (>4% of US) 426 TWh projected 2030; 3,000 built, 1,500 in development
Commercial vs residential sales 2026 1,557 vs 1,546 billion kWh commercial passes residential

EIA price and sales data and PJM/Monitoring Analytics figures as compiled 15 November 2026; attribution study of 385 data centers and 42 utilities as cited therein. A second compilation using the January-July window puts wholesale at $116.53 (+46%) with a $10.48 data-center share, the same tape at a different cutoff.

What the attribution math actually says

The honest center of this story is a gap between two numbers. In PJM, the grid hosting the buildout’s densest corridor, the monitor assigns data centers 9.8 percent of this year’s wholesale cost through the capacity market alone, real money, $29.4 billion of it, socialized across every meter. Nationally, the broadest attribution study, 385 data centers against 42 utilities, credits them with about 4 percent of residential price increases from 2020 to 2025 (nearly 16 percent of industrial), because most of the bill’s rise still comes from the unglamorous ledger: grid hardening, fuel, wages, transmission. Both numbers are true at once, and they bracket the debate: data centers are not yet the main reason the national bill rose $100 this year, and they are precisely the reason PJM’s capacity market cleared at its cap while still short nearly seven gigawatts, administered scarcity, priced at the legal maximum, with December’s auction raising the maximum.

The crossover year

The quiet milestone in EIA’s table deserves its own marker: 2026 is the first year commercial electricity sales (1,557 billion kWh) exceed residential (1,546), households losing their seventy-year status as the grid’s biggest customer class to the building category that contains the server halls. Virginia previews the end state, 26 percent of a state’s supply into one industry, and the 183-to-426 TWh trajectory says the crossover widens through 2030. For ratepayers the structural meaning is allocation: capacity costs, transmission buildouts and reliability margins are socialized by formulas written when load growth meant suburbs, not campuses, and the $29.4 billion line is what those formulas produce before regulators rewrite them. The policy fights now starting in PJM states, large-load tariffs, bring-your-own-generation rules, are the rewrite, and their outcome decides whether the 2027 forecast’s 18.68 cents is a ceiling or a waypoint.

What a household does with a structural story

The levers are the winter-heating guide’s, sharpened by trajectory. The 5.4 percent price rise lands hardest on electric-heated homes, the fuel lottery’s growing column, and the East Coast expectation of up to 7 percent annual growth makes rate-plan choice, time-of-use windows, fixed-rate locks where deregulated markets offer them, worth an hour this winter. Efficiency buys more than it did: every kilowatt-hour cut returns 18-plus cents against 14 two years ago, repricing the payback on heat-pump tuning, insulation and the strip-heat discipline our heating coverage details. And the same buildout driving the bill is the one repricing memory chips and, as our China renewables coverage tracks, reshaping generation economics globally: the household meter is one terminal of the year’s single biggest industrial story, and it is the terminal with the least negotiating power, which is what the December auction and the state tariff dockets are actually about.

A note on dueling windows and why both are in our table

n

Readers comparing coverage will meet two versions of the PJM wholesale jump, +40.9 percent to $111.61 through September and +46 percent to $116.53 through July, and both are the market monitor’s own arithmetic at different cutoffs, the later window softer because autumn prices eased off summer peaks. We print the pair deliberately, as with the gold tracker’s conflicting highs: a reader who knows the window can reconcile any headline, and a reader who doesn’t can be told any story. The same discipline applies to the attribution figures, 9.8 percent of wholesale in one grid versus 4 percent of national residential increases over five years measure different pools over different periods, and the sleight of hand in most coverage of this fight, on both sides, is quoting whichever pool suits the argument without naming it. The numbers above are dated and pooled; that is the entire methodology, and in a policy fight this monetized, it is the part most worth copying.

Our analysis

Two attributions, both true, define the fair debate. Four percent of national residential increases but ten percent of PJM wholesale and $29.4 billion of capacity cost: data centers are a modest share of yesterday’s bill and the dominant driver of the marginal, capacity-constrained dollar, and policy written for either number alone will misprice the other.

An auction at its cap and still short is the number to watch. Administered maximum prices with a 6,831 MW shortfall mean the market wants to clear higher; December’s higher cap and the $5-10 billion projection say ratepayers’ capacity line rises before any new turbine spins, and the 2027 forecast already books it.

Practical advice

  • Know your cents: 18.17 is the national mean; your bill’s rate line, times usage, is the only number that matters.
  • Electric-heat households: this is the fuel column rising on a schedule; the heating guide’s efficiency levers now pay 18 cents per saved kWh.
  • Deregulated-market customers: price fixed-rate locks before December’s auction result flows into offers.
  • Time-of-use plans reward exactly what the grid now needs, off-peak load; check the spread before dismissing them.
  • Watch your state’s large-load tariff docket; it is the one proceeding that decides who pays the buildout.

FAQs

What does electricity cost in 2026? 18.17¢/kWh residential average, up from 17.24 in 2025, about $1,925 a year for a typical home; EIA forecasts 18.68¢ for 2027.

Are data centers raising my bill? Measurably and partially: ~4% of national residential increases 2020-25, but 9.8% of PJM’s wholesale cost and $29.4B in capacity charges, concentrated where the buildout is.

What happened in the PJM auction? The 2028-29 auction cleared at its $325/MW-day cap, $16.4B total, while falling 6,831 MW short of the reliability target; the December auction carries a $337.83 cap.

How big is data-center demand? 183 TWh in 2024, over 4% of US use, projected to 426 TWh by 2030; Virginia already routes 26% of its supply to the industry.

Why does the commercial-residential crossover matter? Cost-allocation formulas assume households are the anchor class; 2026 is the first year they aren’t, and the tariff rewrites follow.

What can I actually do? Rate-plan shopping before December’s auction flows through, time-of-use where the spread pays, and efficiency that now returns 18+ cents per avoided kWh.

See our winter heating cost analysis, our Black Friday memory-crisis analysis, our China renewables coverage, and our natural gas tracker and how we research prices.

Published 21 November 2026. Sources: EIA, Monitoring Analytics and attribution-study compilation (15 Nov 2026), PJM capacity-market explainer with auction figures, checked 21 November 2026. The two wholesale windows (Jan-Jul vs Jan-Sep) differ as noted; bill arithmetic is ours on EIA’s stated price and usage.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top