Why Is My Power Bill So Damn High?

Part 1 · The Bill

What you're actually paying for

Start with what's already happened, because you've been feeling it. In its 2023 rate case, Duke Energy Carolinas won three increases in a row: a typical household bill went up $10.04 a month in January 2024, another $4.19 in January 2025, and another $4.10 this past January, taking a 1,000 kilowatt-hour bill from $130.29 to $148.62. Fuel adjustments move separately on top of that, and Duke has announced it wants roughly another 800 million dollars in fuel costs too.

The next round is already scheduled. The comprehensive settlement announced July 17, now pending at the Utilities Commission, would add about $6.53 a month in 2027 and another $4.66 in 2028 (Duke figures those against a typical bill of about $157 once the fuel riders are counted, calling them 5.9 and then 3.6 percent), on top of whatever fuel is doing... and there is always a next rate case. Even those settled numbers are contested: the state's Attorney General announced on July 21 that he won't sign the deal, and his reasons land in Part 8. The reason there's always a next rate case took me a while to understand, and it turned out to be the whole story, so bear with me for a few screens.

duke's own numbers

$130

2023

$140

2024

+$10.04

$145

2025

+$4.19

$149

now

+$4.10

~$155

2027

+$6.53 pending

~$160

2028

+$4.66 pending

the rate-case piece of a typical 1,000 kWh bill, using Duke's own figures and estimates · solid bars happened · outlined bars add the pending settlement's increases, which Duke calculates off a ~$157 bill that includes fuel riders, so those totals are approximate · fuel swings separately, see below

and then there's the fuel line, which swings

+$19.10Sept 2023 fuel spike

−$8.96Jan 2025 give-back

?whenever gas prices move

One thing to know about that fuel line, because it matters later: it passes through to you at cost. Duke earns nothing on it, and loses nothing when it spikes, so you carry the swings while the company stays dry. Hold that thought for Part 5.

And if you want to see where each of these pieces sits on the actual paper Duke mails you, I took Duke's own published sample bill and annotated it: how to read your power bill, line by line. It's a good companion to everything below, because the most important number on the bill turns out not to be printed on it.

Sources: Duke's own release on the approved 2023 rate case ($130.29 to $148.62 by 2026) · NCUC Docket E-7 Sub 1329, comprehensive settlement announced July 17, 2026 · WUNC on the settlement figures ($6.53 and $4.66 on a typical $157.15 bill) · Gov. Stein on Duke's announced $800M fuel-cost request · the Sept 2023 fuel adjustment (+$19.10/mo) · the Jan 2025 fuel decrease (−$8.96/mo).

Part 2 · The Machine

How Duke actually makes its money

Collage illustration of a transmission tower standing on stacks of coins.

Here's the part I had wrong my whole adult life, and I'd bet most people have it wrong the same way. I assumed Duke profits by selling electricity, with some markup baked into every kilowatt-hour. The actual arrangement is stranger. The fuel Duke burns gets passed through to customers at cost, with no profit on it at all. It goes further than that: under the framework the legislature set up in 2021, Duke's residential revenue is decoupled from how much electricity households actually buy. Sell more, sell less, the approved revenue is the revenue. What Duke earns is a guaranteed return, 9.8 percent under the pending settlement, on the money it invests in physical stuff: the plants, the poles, the wires, and the upgrades. Regulators call that pile the rate base, and growing it is the whole business. There are side dishes, like small performance incentives under that same framework, but next to the return on rate base they're pocket change.

Building is what it's allowed to profit on.

Economists have warned since 1962 that a monopoly paid this way has every incentive to build more than it needs, for exactly that reason. And it can't stop, thanks to a boring accounting fact that explains a lot once you see it: the rate base shrinks every year as equipment depreciates, so a utility that stops building watches its earnings melt away on their own. Duke has promised Wall Street earnings growth of 5 to 7 percent a year, which in this business means growing the rate base, which means building. In February the company announced 103 billion dollars in planned construction, the largest program any American utility has ever put forward.

So your bill mostly tracks the payment plan on everything Duke has approval to build, plus the guaranteed margin on top, far more than it tracks the price of coal or gas in any given month. There is an official check on all this: Duke can't build anything big without a certificate from the Utilities Commission, and spending the commission finds imprudent can be disallowed. But ask what the commission weighs when it decides whether a plant is needed, and the answer is Duke's own demand forecast, which is where Part 4 picks up. The approvals keep coming, because asking for them is the growth strategy. Here's the whole cycle, drawn out:

the business model

Promise Wall Street profits will grow 5 to 7 percent, every year, forever

↓

That means growing the rate base, the pile of stuff earning the guaranteed 9.8 percent

↓

Get a forecast that justifies buildingthe permission slip... see Part 4

↓

Build the new plants, and keep the old ones running and earning toowhether or not the demand shows up

↓

The payment plan, plus the margin, lands on your bill

↺ and then it goes again... that's the 3-rate-cases-in-7-years part

Sources: settlement ROE of 9.8% per NCUC Docket E-7 Sub 1329 · Averch & Johnson, "Behavior of the Firm Under Regulatory Constraint," American Economic Review (1962) · HB 951 (S.L. 2021-165) (performance-based regulation: residential decoupling, performance incentives, multiyear rate plans) · Fortune on Duke's $103B capital plan

Part 3 · The Law

They named it the Power Bill Reduction Act

I'll give them this much, the name has confidence. Senate Bill 266 became law in July 2025 over the Governor's veto, and it changed the rules in three ways that matter for the machine you just read about. It lets Duke charge customers for a plant that's still under construction, years before it makes any electricity. It repealed the 2030 carbon checkpoint that would have forced the old coal plants to retire on schedule, a repeal that researchers at NC State and, yes, Duke University independently priced at up to 23 billion dollars in added fuel costs through 2050. And it shifted more of the fuel bill onto households, roughly 19 percent more of the burden by one independent analysis.

I want you to have the other side's number too, because you'll hear it: supporters projected the repeal could save ratepayers up to 15 billion dollars in avoided compliance costs by 2050, working from a Public Staff estimate in that neighborhood. The catch is what that math assumes. The savings only show up if natural gas stays cheap for the next twenty-five years, and the modeling behind the estimate didn't account for gas price swings or the rising cost of gas turbines. The NC State and Duke University analyses that did price the fuel risk came out 23 billion dollars in the other direction. Both are projections riding the same bet on cheap gas, and you already know from Part 1 who carries the fuel line when that bet goes wrong: you do, at cost, while the company stays dry.

Governor Stein vetoed it and said plainly that it moved costs onto families. The Senate overrode him on July 29, 2025, with every Republican senator voting yes, including the one who represents Stokes and Forsyth, and every Democrat present voting no. The House finished it the same day. The roll calls are public, and worth a look.

The official Senate roll call transcript for Roll Call 431 on SB 266, The Power Bill Reduction Act, showing the veto override passed with 30 Republican ayes, zero Democratic ayes, and 18 Democratic noes on July 29, 2025. — the actual state record. note the line that says "Ayes (Democrat): None"

Senate veto override · Roll Call 431

30 AYES, all Republican18 NOES, all Democrat

House veto override · Roll Call 586

74 AYES (71 R, 3 D)46 NOES, all Democrat

Sources: SB 266 (Session Law 2025-78) · Senate Roll Call 431 · House Roll Call 586 · WUNC's coverage of the NC State cost analysis and the residential fuel-cost shift · Gov. Stein's office on NC State and Duke University independently finding $23B · Carolina Journal on the supporters' $15B savings projection · Inside Climate News on what the Public Staff estimate left out

Part 4 · The Demand Story

About those data centers

Step 3 of the cycle needs a forecast, and the data centers are it. In this system a demand forecast works like a permission slip: it's the document that turns "we want to build" into "we get to build, and you get to pay for it." Duke's estimate of what new large customers will need, mostly data centers, has grown every time the company has updated it, and Duke now says data centers could be a quarter of everything it serves in the Carolinas by 2030. On the strength of that projection, it plans roughly 14 gigawatts of new generation by the end of the decade, and it gets to keep the old plants running for the same reason.

3.9GW

the large-load forecast two years ago

6GW

the forecast by May of last year

8GW

the forecast now, for 2035

Maybe the demand shows up. But notice who carries the bet if it doesn't: under the new law, construction charges land on your bill while the concrete is still wet. The July settlement did add a refund rider, and it's a real concession: if certain planned projects don't get finished on time, that money comes back with interest. Look at what it insures against, though. It covers construction running late. It doesn't cover the plants getting built right on schedule for demand that never shows up... in that case the money stays spent, and you keep paying the return on it. Duke is meanwhile offering special rate terms to attract the data centers themselves, which tells you which customer is being courted, and it isn't the house at the end of a gravel road in Quaker Gap.

The Attorney General has been pushing on exactly this seam. His office argued that data centers and other giant new customers should carry the costs they create instead of blending them into everyone's bills, and the July settlement does include a fast-track proceeding to write new rules for large-load customers. He still didn't sign the deal, and his caution about that proceeding is worth carrying with you: the proposal moves in the right direction, but "the details will determine whether families are actually protected."

There's a separate field guide on the data-center push if you want to go deeper on that piece.

the forecast doubled in two years

Sources: Canary Media on Duke's data-center forecasts · Duke's 2025 Carolinas Resource Plan · Canary Media on the proposed data-center tariff · the July 17 settlement's refund rider for late projects · AG Jackson's July 21 statement (data-center cost protections and the fast-track large-load proceeding)

Part 5 · A Note For My Green Friends

Why the old coal plants won't die

Collage illustration of a coal plant smokestack beside a lake at dusk.

If you care about the climate side of this, you've probably wondered why a company would fight to keep 1970s coal plants running while asking the legislature to repeal the state's carbon targets. The answer is less ideological than it looks. It's the same machine as everything else on this page.

Remember the fuel line from Part 1: it passes through at cost, so burning expensive, inefficient coal costs Duke's shareholders nothing... you pay for it, and SB 266 shifted more of that burden onto households at the same time. Meanwhile an old plant is still rate base. Its remaining book value keeps earning the guaranteed return, and every "reliability upgrade" that keeps it limping along is fresh investment earning 9.8 percent. Retiring it early would mean writing chunks of that off, on a schedule Duke didn't choose.

The 2030 carbon checkpoint was that schedule. It would have forced the old plants into retirement whether or not it suited the balance sheet, which is exactly why it had to go. With the checkpoint repealed and federal rules loosening, Duke gets to run coal into the 2040s, collect on the upgrades, and still build everything new for the data centers, with the up-to-23-billion-dollar fuel tab from the repeal landing on ratepayers, where it never touches the profit math. The carbon plan lost an argument about rate base, and the science was never really in the room.

Sources: fuel pass-through per NCUC ratemaking (annual fuel rider dockets) · NC State and Duke University analyses of the interim-target repeal (up to $23B) and the ~19% residential fuel-cost shift, via WUNC · WUNC on the coal retirement delays

Part 6 · The Warning Label

South Carolina already ran this experiment

The charge-during-construction idea has a track record one state away. South Carolina passed its version in 2007. Customers there paid roughly two billion dollars, across nine rate increases, for two nuclear reactors that were abandoned half-finished in 2017. The utility's CEO went to prison for misleading people about the project, and the legislature repealed the law.

Collage illustration of a half-built cooling tower with coins falling into a hole at its base.

$2B

paid by South Carolina customers during construction

separate rate increases along the way

watts anyone ever got out of that hole in the ground

Our new law is the same species as theirs. And Duke has already flagged where a project like that could land: this past December 30 it filed an early site permit application with the Nuclear Regulatory Commission for land near the Belews Creek coal plant, on the lake in eastern Stokes County. An early site permit is a first step, not a construction permit... it reserves the option of small reactors there, and Duke says it hasn't decided to build. But South Carolina's reactors started as paperwork too, and if this one goes forward, the new law is how it gets financed. The plant, the lake, and what Duke has planned out there deserve a full page of their own, and I'm working on it.

Sources: South Carolina Base Load Review Act (2007, repealed 2018) · V.C. Summer Units 2 & 3 abandonment, 2017 · United States v. Kevin Marsh (SCANA), guilty plea, 2021 · Duke's Dec 30, 2025 early site permit application for Belews Creek

Part 7 · Other Ways To Do This

This part is actually fixable

The setup we have, an investor-owned monopoly earning a guaranteed return on whatever it builds, is a choice, and other places have chosen differently. Hawaii passed a short law in 2018 ordering its utility commission to break the link between utility revenue and construction, so the utility there earns more when the lights stay on and rooftop solar gets connected quickly, and nothing extra for pouring concrete nobody asked for. Colorado makes its biggest utility win an open competition before it builds anything new. Nebraska, which nobody has ever accused of being a commune, has no investor-owned electric utility at all: the whole state runs on public power districts with elected boards, and their rates sit below the national average.

You don't even have to leave the district

Surry-Yadkin EMC, the electric co-op serving parts of Stokes and Forsyth along with three neighboring counties, answers to its 27,000 members instead of shareholders. When it collects more than it needs, the margins go back to members as capital credits. Their power company mails them money. Ours mails Charlotte a dividend.

I want to be straight about the limits, because the places that tried a hostile government takeover of their big utility have mostly lost those fights, and I'm not pitching one. What a state senator can actually push for is duller and more useful: pay the monopoly for performance instead of construction, make it compete for the right to build anything big, and protect the co-op and municipal models that already work here.

capital credits are a real thing... ask around

Sources: Hawaii Act 5 (2018) and the Hawaii PUC's performance-based regulation framework (Docket 2018-0088) · Colorado all-source competitive solicitations · Nebraska public power (all utilities consumer-owned since the 1940s) · Surry-Yadkin EMC

Part 8 · The Fine Print We'd Rather You Hear From Us

About that 3.7 percent

If you follow this stuff, you might reasonably object that the system just worked. The number in the coverage is a 3.7 percent average annual increase, and the story behind it is real: Duke walked in asking for close to 18 percent more from residential customers over two years, the Public Staff pushed back on behalf of customers, and the settlement landed at roughly 9.5 percent for households instead, the 5.9 and 3.6 from Part 1. The pushback was real, and the people who did it deserve the credit.

Then, four days after the settlement was announced, the state's top consumer lawyer looked at the deal and walked. Attorney General Jeff Jackson said on July 21 that his office won't sign it. His team ran its own independent analysis of what Duke actually needs to build to meet demand and concluded, in his words, that Duke "overshot the mark": the company "can come down significantly and still meet demand." His office also says the profit rate is still too generous. The settlement sets Duke's return on equity at 9.8 percent; the Attorney General backed 7.4, and priced the gap at about 1.37 billion dollars of customer money over two years. The Utilities Commission gets the final call, expected by November.

Think of it like pie

Duke has told Wall Street its profit will grow 5 to 7 percent every year, and that promise is load-bearing... miss it and the stock slides and the bonuses dry up. Duke's profit is a slice of everything it builds, and the settlement trimmed the angle of that slice. The promise didn't budge. So there's really only one move left: bake a bigger pie.

The 103 billion dollar construction plan is the bigger pie. The pie tin is your bill.

Sit with what the Attorney General's finding actually says. His office isn't quibbling over the paperwork; its independent analysis concluded the planned build-out is bigger than the demand justifies. "Overshot the mark" is the bigger pie, described by a lawyer. The settlement even schedules the 2027 and 2028 increases as payments tied to construction getting done. I'd read the 3.7 as the machine running quietly rather than the machine being fixed.

the percentage decoder

One deal, and the coverage of it has produced at least nine different percentages, quoted interchangeably. They sort into two piles: numbers about your bill, and numbers about Duke's profit. Here's every one of them, decoded.

Numbers about your bill

18%

Where Duke opened: the residential increase it asked for over two years when it filed the case in January.

11.6%

Where the ask stood after the first, partial settlement in mid-July (7.5 percent, then 4.1).

9.5%

Where the comprehensive settlement landed for households: 5.9 percent in 2027 plus 3.6 percent in 2028. Those are the $6.53 and $4.66 from Part 1.

3.7%

The very same deal, averaged per year across every customer class, factories included. This is the number in the friendlier headlines.

Numbers about Duke's profit

10.1%

Duke's return on equity today: the guaranteed profit rate on the shareholder-funded share of everything it builds. This is the dial all the ones below are fighting over.

10.95%

The return Duke asked the commission to raise it to.

9.8%

The return the settlement would set instead, the number running through Parts 2 and 5 of this page.

7.4%

The return the Attorney General's office says would still get everything built, about 1.37 billion dollars less of your money over two years.

5–7%

The yearly profit growth Duke has promised Wall Street. Every other number on this list answers to this one.

clip and save for the next round of headlines

Sources: NCUC Dockets E-7 Sub 1329 and E-2 Sub 1380 · the comprehensive settlement announced July 17, 2026 (9.8% ROE, multi-year rate plan with construction-contingent increases and a refund rider) · AG Jackson's July 21 "No Deal" statement (18% opening ask, 11.6% revision, 9.5% settlement, 7.4% ROE position, $1.37B) · WCCB's July 22 interview ("overshot the mark," decision expected by November) · WSOC on the 11.6% partial settlement · Duke Energy investor guidance

The Part Where This Becomes Your Business

The fix runs through the legislature

The Utilities Commission votes on the settlement this fall, and the commissioners are appointed, so you don't get a say in that room. But every rule the commissioners have to follow got written up the street at the General Assembly, and every fix in Part 7 is legislation. Hawaii's entire reform started as a bill a few pages long, ordering its commission to pay the utility for performance instead of construction. North Carolina could pass the same kind of law tomorrow. What it needs is legislators who'll sponsor it, and the ones who represent us right now voted for the current rulebook instead.

Which means the fight over your power bill is a state legislature race. That happens to be the race I'm running, and if a power bill is what brought you here, this is where it goes.

Donate Volunteer Get the updates

This page cites its work. Every claim traces to a roll call, a docket, or a published filing, and if you find something wrong, email hello@andycantwin.com and I'll fix it and say so publicly.

Why is my power bill so damn high?