Restore the Electricity Credits

Put the clean electricity production and investment credits back on their original schedule to 2035, reversing the 2025 phase-out.

AI evaluation · not yet reviewed by a human

This evaluation was produced and sourced by an AI model; a human review is still pending. Figures and conclusions may still change. The review log is at the foot of the page.How review works →

Two technology-neutral credits pay generators either for each unit of low-carbon electricity produced or for a share of what the plant cost to build. The 2025 reconciliation law brought both to an early end for wind and solar, tightened the deadline for qualifying, and added ownership restrictions. Restoring them means returning to the schedule that ran to 2035, with the same eligibility rules and the same rates. Nothing else about the electricity system changes: the same interconnection queues, the same state regulators, the same market design. This evaluation looks ten years ahead from 2026.

Balance

Better for the future · 0.68 previous scale

Balance on the previous scale. The Bilanz 2.0 simulation is not yet available for this evaluation. The category comes from the share of the debate on the pro side (r).

For 61 · 68 % Against 29 · 32 %

Size class: large Scale of this evaluation: Normalised Impact — unitless, calibrated to this topic. For comparison: one point here is worth roughly 2 billion euro per year. The credits buy avoided carbon dioxide at almost exactly 100 euro a tonne, which is the price this site uses. That makes the carbon a wash and leaves the outcome to the two things that come with it: lower electricity bills and cleaner air. Anyone who prices a tonne lower will read this balance differently. How we score →

Arguments for

Arguments against

7 arguments evaluated · Scoring v1.3 Δ absolute +32

Arguments — For

4 arguments · top 3 shown

A hundred and sixty million tonnes a year

34of 100

Wind and solar additions are projected to fall by well over half through 2035 without the credits. What replaces that generation is mostly gas, and increasingly coal. The gap in emissions between the two futures is measured in hundreds of millions of tonnes by the mid-2030s.

Value 7 · ClimateImpact 8Plausibility 6

▸ Show reasoning & sources ▾ Hide reasoning & sources

The stream is carbon dioxide that stays out of the atmosphere, which this site places in the class it uses for the environment, above money and below life and health. The weight does not rise because the emissions are American or fall because they are one country's share of a global total: a tonne is a tonne. What is priced is the physical quantity, not the harm any particular person suffers from it, and the harms are the reason the quantity matters rather than a second stream to add. The cleaner air that comes with burning less gas and coal is a genuinely separate stream, because it harms people where it happens rather than everywhere, and it is counted separately below. The value sits in the upper middle of the scale, at the level this site uses for the environment.

Independent modelling of the 2025 law puts American emissions 315 to 574 million tonnes higher in 2035 than they would have been, with power sector emissions 19 to 79 percent above the baseline ^[1]. That figure covers all the law's energy provisions, including the vehicle and household credits, which this measure does not restore; the electricity credits carry roughly seven tenths of it, giving about 310 million tonnes in 2035. The gap opens gradually as plants that would have been built are not, so an average across the decade counted here is a little over half the endpoint: 160 million tonnes a year, in a range from 80 to 280. For scale, American energy-related emissions were 4,904 million tonnes in 2025 and rising ^[5]. The modelling compares two futures rather than crediting the policy with everything clean that gets built, so generation that would have happened without the credits is already netted out. A tonne is valued here at 100 euro, the cost of avoiding it elsewhere. The Impact is the largest in this debate and, at the price used here, almost exactly equal to what the credits cost the budget.

▸ Show calculation ▾ Hide calculation

	Additional emissions in 2035 without the 2025 law's energy provisions [1]	range 315 to 574 million tonnes	445 million tonnes
×	Share carried by the electricity credits Setting, range 55 to 85 percent: the rest is the vehicle and household credits, which this measure does not restore	70 %	311 million tonnes
×	Average across the decade rather than the endpoint Setting, range 30 to 70 percent: the gap opens gradually as plants that would have been built are not	52 %	160 million tonnes a year
×	Value of a tonne the cost of avoiding a tonne elsewhere, which is the rate this site uses	100 euro	16 billion euro
÷	Normalised Impact scale of this evaluation	2 billion euro a point	8

Score 8 Impact × 7 Value × 6 Plausibility ÷ 10 = 34 of 100

The chain is short and the contested link is the middle one. That subsidised generation displaces gas and coal is not in dispute; that removing the subsidy reduces how much gets built is what the modelling estimates. The counterfactual is the credit schedule as it stood before 2025, which is a written rule rather than a scenario, and the comparison is made inside a capacity-expansion model that has been run against outturns for a decade. The confounder that matters is technology cost: if wind, solar and battery costs keep falling as they have, projects clear without the credit and the gap between the two futures narrows to little. That possibility is exactly what the wide range reflects, and it is unresolved — no design can settle it, because it is a question about the future rather than about the past. Two independent modelling teams reach the same order of magnitude, which is the strongest thing that can be said ^[1]^[4]. Reverse causation does not arise. The Plausibility is at the top of what a projection can carry: the mechanism is close to arithmetic and the size depends on a cost path nobody can observe yet.

evidence basis: Projection · P ceiling 6 identification: Mechanistic · rung ceiling 6

Counterfactual: the credit schedule as it stood before the 2025 law, as written. Design: mechanistic — a capacity-expansion model comparing two policy futures (Rhodium Group [1]), corroborated by an independent team [4]; no natural experiment exists. Confounder: technology costs falling far enough that projects clear without the credit, which would close the gap; unresolved and reflected in the 80 to 280 million tonne band. Direction: no reverse causation. Ceiling: projektion 6.0 binds, and mechanistic gives the same. Free-riding — credits paid to projects that would have been built anyway — is already netted by the two-futures comparison rather than deducted here.

Electricity bills that rise less

14of 100

A generator that receives a credit per unit produced can bid lower, and wholesale prices follow. Household energy spending is projected to be 78 to 192 dollars higher in 2035 without the credits. Energy is a bigger share of a poor household's budget than of anyone else's.

Value 5 · Household budgetsImpact 5.1Plausibility 5.5

▸ Show reasoning & sources ▾ Hide reasoning & sources

The stream is money households do not pay for electricity, priced at the middle of the scale like any other. It is not a resource saving but a transfer: the credit comes from the federal budget and part of it reaches customers as lower prices, and the paying side of that is counted against this measure. What makes the transfer worth something is where it lands. Electricity is close to a fixed cost for a household — the fridge runs whatever the income — so it takes a far larger share of a small budget than a large one, and a euro of relief is worth correspondingly more there. That weighting is in the Impact, not here, and it is not counted a second time as a fairness argument. The value is the middle of the scale, because the stream is money and the budgets it lands in are priced in the Impact.

The two credits together score at 183 billion dollars over ten years — 28 billion for the production credit and 155 billion for the investment credit — which is 18.3 billion dollars a year, or 15.8 billion euro at 1.16 dollars to the euro ^[2]. Half of that is assumed to reach customers as lower wholesale and retail prices, in a range from three tenths to seven tenths, giving 7.9 billion euro a year; the remainder stays with the owners of the plants and is counted against this measure below. Across 132 million households that is about 60 euro a year each, which sits at the low end of the 78 to 192 dollar figure that independent modelling gives for all the 2025 provisions together in 2035 ^[1]. Electricity spending is roughly three times as large a share of income in the bottom fifth as in the top, so a euro of relief here is weighted at 1.3 rather than 1.0. The result is 10.27 billion euro a year. The Impact is the second largest on the pro side, and it is the half of the credit that reaches people rather than plants.

▸ Show calculation ▾ Hide calculation

	Cost of the two credits [2]	28 billion dollars for the production credit and 155 billion for the investment credit over ten years	18.3 billion dollars a year
÷	In euro	1.16 dollars to the euro	15.8 billion euro
×	Share reaching customers as lower prices Setting, range 30 to 70 percent: the rest stays with the owners of the plants and is counted against this measure below [1]	50 %	7.9 billion euro
×	Weight of a euro of electricity relief Setting, range 1.1 to 1.6: electricity takes about three times as large a share of income in the bottom fifth of households as in the top	1.3	10.27 billion euro
÷	Normalised Impact scale of this evaluation	2 billion euro a point	5.1

Score 5.1 Impact × 5 Value × 5.5 Plausibility ÷ 10 = 14 of 100

Two things carry this figure and they differ in firmness. That a production credit lowers the price at which a generator will bid, and that lower bids lower wholesale prices in a marginal-cost market, is market mechanics rather than a contested finding, and the counterfactual — the same market without the credit — is what the modelling compares against ^[1]. What is estimated is the pass-through: how much of the credit reaches customers rather than staying with plant owners. That share has been studied for other energy subsidies and comes out anywhere between a third and three quarters depending on market structure, which is the range used here. The confounder that matters is retail regulation: in much of the country, retail prices are set by state commissions rather than by wholesale markets, so a wholesale saving may reach customers slowly or not at all. That is unresolved and is the main reason this figure is not scored higher. Reverse causation does not arise. The Plausibility is at the upper end of what a projection can carry: the mechanism is market arithmetic, and how much of it reaches a household bill is genuinely open.

evidence basis: Projection · P ceiling 6 identification: Mechanistic · rung ceiling 6

Counterfactual: the same electricity market without the credits, as modelled [1]. Design: mechanistic — chain named (credit → lower bids → lower wholesale prices → lower bills), with the pass-through share estimated from studies of other energy subsidies rather than measured for these. Confounder: state retail regulation, under which a wholesale saving may reach customers slowly or not at all; unresolved. Direction: no reverse causation. Ceiling: projektion 6.0 binds, mechanistic gives the same. The pass-through doubt sits in the three-to-seven-tenths band as well as in P, because the band prices the size and the regulation question prices whether the link works at all.

Cleaner air where the plants are

9.8of 100

Burning less gas and coal does not only avoid carbon dioxide. It avoids the sulphur dioxide, nitrogen oxides and fine particles that shorten lives in the counties downwind of a power station. Those harms land on identifiable people rather than on the atmosphere.

Value 7 · Air and healthImpact 2.8Plausibility 5

▸ Show reasoning & sources ▾ Hide reasoning & sources

The stream is air quality and, through it, asthma, heart disease and shortened lives near power stations. It sits between the environment and health, above money and below life counted alone. The people affected are identifiable in a way the climate stream's are not: they live in particular counties, and those counties are on average poorer than the country as a whole. That concentration is a fact about who bears the harm rather than about what it is worth, and no separate adjustment is made for it here. The carbon dioxide from the same combustion is a different stream with a different reach and is counted separately, which is the ordinary treatment on this site rather than a double count. The value sits in the upper middle of the scale, between the environment and health, because the stream is both.

The pollutants that harm people locally come out of the same chimneys as the carbon dioxide, so the quantity avoided moves with it: 160 million tonnes of carbon dioxide a year not emitted from gas and coal plants. Studies of American power sector decarbonisation put the local health benefit at roughly 30 to 60 dollars for every tonne of carbon dioxide avoided, because the plants that run at the margin are the dirtier ones and they sit near populations. Thirty-five euro a tonne is used here, in a range from 15 to 70. The result is 5.6 billion euro a year. Two things are left out and they run in opposite directions: the coal plants that would run more without the credits are the worst offenders per unit, which would raise this figure, and separate air rules constrain them independently, which would lower it. The Impact is a third of the climate gain, which is the usual proportion in American power sector studies and is the part of it that lands on people rather than on the atmosphere.

▸ Show calculation ▾ Hide calculation

	Carbon dioxide not emitted from gas and coal plants [1]	from the argument above	160 million tonnes a year
×	Local health benefit per tonne avoided Setting, range 15 to 70 euro: American power sector studies put it at 30 to 60 dollars a tonne, because the plants running at the margin are the dirtier ones and sit near populations	35 euro	5.6 billion euro
÷	Normalised Impact scale of this evaluation	2 billion euro a point	2.8

Score 2.8 Impact × 7 Value × 5 Plausibility ÷ 10 = 9.8 of 100

That burning less fuel emits less of everything is chemistry, and the counterfactual is the same generation mix without the credits. What is not measured here is the ratio between local pollution and carbon dioxide for the specific plants that would run more. That ratio varies by an order of magnitude between a modern gas plant and an old coal unit, and which of them fills the gap depends on where and when. No source gives the figure for this measure; the 30 to 60 dollar band comes from studies of American power sector decarbonisation generally and is carried across. The confounder that matters is the separate air rules — limits on sulphur dioxide, nitrogen oxides and particulates that apply whatever the credits do — which would prevent part of what this argument counts. Those rules are themselves being loosened, which cuts the other way, and neither effect is resolved here. Reverse causation does not arise. The Plausibility is at the top of the range for a claim whose chain is fully visible and whose size is carried over from studies of something adjacent rather than measured here.

evidence basis: Mechanism · P ceiling 6 identification: Mechanistic · rung ceiling 6 band: Chain closed, unevidenced · P 4–5

Counterfactual: the same generation mix without the credits. Design: mechanistic — chain named (less gas and coal burned → less local pollution → fewer deaths and illnesses), with the pollution-to-carbon ratio carried across from studies of American power sector decarbonisation rather than measured for this case. Confounder: separate air rules constraining the same pollutants independently, which would prevent part of the effect; unresolved, and cutting the other way as those rules are themselves loosened. Direction: no reverse causation. Ceiling: mechanistic 6.0 binds. Band: chain closed but unevidenced — every link is named and the counter-mechanism is stated on both sides; only the measurement for this case is missing.

Nothing measured argues against the claim; what is missing is the local pollution ratio for the specific plants that would run more. The counter-mechanism, separate air rules catching part of it, is named and offset by those rules being loosened at the same time. Read back: cleaner air worth roughly a third of the climate gain follows from this measure about as often as it does not.

Open: Hourly emissions data are reported by every American power plant. Matching them against the generation the credits actually induce would give the ratio directly and could carry P to 6.

Factories that have something to make

3.2of 100

Around 110 billion dollars of announced battery, panel and component plants were sized for a market the credits were meant to create. Much of that capacity is now waiting for demand that will not arrive. Restoring the credits gives it work.

Value 6 · OutputImpact 1.2Plausibility 4.5

▸ Show reasoning & sources ▾ Hide reasoning & sources

The stream is capital producing what it was built to produce rather than standing partly idle, and the workers hired for the difference. It belongs to the class this site uses for economic systems and prosperity. What is counted is the output, not the money already spent building the plants, which is gone either way. Nor is anything counted for the fact that the jobs are in particular places: a job is a job on this scale, and where it sits is a political matter rather than a welfare one. Whether the owners or the workers capture the gain does not change the weight. The value sits in the middle-upper part of the scale, at the level this site uses for economic output.

About 110 billion dollars of announced manufacturing capacity — batteries, panels, components — was sized against the demand the credits were expected to create, and the modelling of the 2025 law counts it as at risk ^[1]. A quarter of that running below its design output is used here, in a range from a tenth to a half: 27.5 billion dollars of underused investment, or 2.75 billion dollars a year across the decade, which is 2.4 billion euro. What the figure does not assume is that the plants close or that the market disappears; demand grows without the credits too, and the loss counted here is only the difference between capacity built and capacity used. It also does not count the plants that were announced and then cancelled before construction, because capital that was never committed is not lost. The Impact is the smallest on the pro side and it is the one that falls on particular towns rather than on the country in aggregate.

▸ Show calculation ▾ Hide calculation

	Announced manufacturing capacity at risk [1]		110 billion dollars
×	Share running below its design output Setting, range 10 to 50 percent: export markets and state procurement could absorb part of it	25 %	27.5 billion dollars
÷	Spread across the decade counted here	10	2.75 billion dollars a year
÷	In euro	1.16 dollars to the euro	2.4 billion euro
÷	Normalised Impact scale of this evaluation	2 billion euro a point	1.2

Score 1.2 Impact × 6 Value × 4.5 Plausibility ÷ 10 = 3.2 of 100

The mechanism is not in doubt: capacity built against an expected market has less to do when the market does not appear. The counterfactual is the same announced plants under the credit schedule as it stood. What has no source is the size — nobody has published how much announced capacity is now surplus, because the answer depends on how fast the market grows without the credits, which is the same open question that runs through this whole debate. The counter-mechanism is real and only partly answered: export markets and state-level procurement could absorb capacity that the federal credits no longer support, and the quarter used here is meant to reflect that possibility rather than settle it. There is also a risk of counting the same thing twice, since some of the output these plants would make is already inside the emissions figure; that is avoided by counting only the idle capacity and not the electricity it would have helped generate. Reverse causation does not arise. The Plausibility is below the middle because the size of this loss has no source behind it and the counter-argument that other markets absorb the capacity is unresolved.

evidence basis: Mechanism · P ceiling 6 identification: Mechanistic · rung ceiling 6 band: Chain closed, unevidenced · P 4–5

Counterfactual: the same announced plants under the pre-2025 credit schedule. Design: mechanistic — chain named (credits withdrawn → market smaller than planned → capacity runs below design), with no published figure for the surplus. Confounder: export markets and state procurement absorbing the capacity; partly reflected in the quarter used, not resolved. Direction: no reverse causation. Disjointness: only idle capacity is counted, not the electricity these plants would help generate, which is already in pro-1. Ceiling: mechanistic 6.0 binds. Band: chain closed but unevidenced — links named, counter-mechanism addressed by the low share, measurement missing.

Nothing measured argues against the claim; what is missing is a published figure for surplus capacity. The counter-mechanism — export markets and state procurement filling the plants — is addressed by taking only a quarter of the announced capital. Read back: about half the time, roughly a quarter of the announced capacity runs below what it was built for.

Open: Plant-level utilisation appears in manufacturers' quarterly filings; two years of those against announced design capacity would replace the setting with a measurement and could carry P to 6.

Arguments — Against

3 arguments

Sixteen billion euro a year from the budget

24of 100

The tax committee scores the two credits at 183 billion dollars over ten years, most of it the investment credit. That is money the federal government does not collect, in a year when it is already borrowing 5.8 percent of national output. The credits are uncapped: the cost follows whatever gets built.

Value 5 · Public financesImpact 7.9Plausibility 6

▸ Show reasoning & sources ▾ Hide reasoning & sources

The stream is federal money, priced at the middle of the scale like any other euro. A tax credit is spending in the shape of a deduction, and this site treats it as such: revenue not collected is money not available. Public money is not treated as cheaper or dearer than private money, and the deficit is not a second harm on top of the spending, because borrowing shifts who pays and when and that shift is inside the euro figure. Where the money ends up — with customers as lower bills or with plant owners as returns — is counted on the receiving side, once as a gain and once as a loss. The value is the middle of the scale, the level this site uses for public money whatever it is spent on.

The tax committee scored the clean electricity production credit at 28 billion dollars and the clean electricity investment credit at 155 billion over ten years, together 183 billion, or 18.3 billion dollars a year ^[2]. At 1.16 dollars to the euro that is 15.8 billion euro a year, in a range from 12 to 22 billion. The width of that range is not measurement error but design: neither credit has a cap, so the cost follows whatever gets built, and the scores assume a build-out path that the same modelling shows to be uncertain by a factor of two ^[1]. The euro carries the standard weight of one for public money. What the money buys is counted in the four arguments above, so this argument is the payment and nothing else. The Impact is the largest against and it is almost exactly equal to the carbon dioxide it buys, which is what makes this debate turn on everything else.

▸ Show calculation ▾ Hide calculation

	Cost of the two credits over ten years [2]	28 billion for the production credit, 155 billion for the investment credit	183 billion dollars
÷	Per year	10	18.3 billion dollars a year
÷	In euro Setting, range 12 to 22 billion euro: neither credit is capped, so the cost follows whatever gets built	1.16 dollars to the euro	15.8 billion euro
×	Weight of a euro in the federal budget the standard weight for public money on this site	1.0	15.8 billion euro
÷	Normalised Impact scale of this evaluation	2 billion euro a point	7.9

Score 7.9 Impact × 5 Value × 6 Plausibility ÷ 10 = 24 of 100

The scores come from the tax committee's own estimates of the provisions the 2025 law changed, so the counterfactual — the pre-2025 schedule — is the one the scores were written against ^[2]^[3]. That a credit costs revenue is arithmetic rather than prediction. What is estimated is how much gets built and therefore claimed, and here the doubt is unusually wide in both directions because neither credit is capped: a faster build-out costs more and delivers more, a slower one costs less and delivers less, and the two move together rather than independently. That coupling is why the range here matches the range on the emissions figure rather than being set separately. The one asymmetry is that credits are paid to projects that would have been built anyway, which costs money without buying emissions; that free-riding is inside the score and is not netted out of it. Reverse causation does not arise. The Plausibility is at the top of what a budget projection can carry: the rule is certain, and only the amount that gets claimed is estimated.

evidence basis: Projection · P ceiling 6 identification: Definitional · no rung ceiling

Counterfactual: the pre-2025 credit schedule, which is what the tax committee's scores were written against [2][3]. Design: definitional — a credit costs revenue by arithmetic; the estimated element is how much is claimed. Confounder: none for the arithmetic; the build-out path drives the amount and is coupled to the emissions figure rather than independent of it. Direction: not applicable. Ceiling: projektion 6.0 binds; definitional carries no ceiling of its own. Free-riding is inside the score and is not netted out, which makes this figure if anything high relative to what the credits induce.

Half the money stays with the plant owners

3.5of 100

A subsidy lands where the market lets it land. The share of the credit that does not reach customers as lower prices stays with the developers and their investors, who are among the wealthiest holders of capital in the country. A euro is worth less there than in the budget it came from.

Value 5 · Household budgetsImpact 2.0Plausibility 3.5

▸ Show reasoning & sources ▾ Hide reasoning & sources

The stream is money again, at the middle of the scale. The objection is not that developers should not profit; it is that the euro ends up where it is worth less than where it started, and only that difference is a loss. The payment itself is already counted on the paying side in the argument above, so counting the whole sum here would be the same euro twice. Treating the ownership of clean energy assets as a separate question of fairness would price the same fact a third time, once as fairness and once as weight. The value is the middle of the scale, and the distributional point is carried entirely by the Impact.

Of the 15.8 billion euro the credits cost each year, half is assumed above to reach customers as lower prices; the other 7.9 billion stays with the owners of the plants. Those owners are utilities, infrastructure funds and their investors, who sit in the top tenth of American wealth holders, where this site counts a euro at half its worth at median income. The difference between that and the one the federal euro carries is 0.5, giving 3.95 billion euro a year of value dissipated in the passage, in a range from 2.4 to 5.5 billion. The range moves inversely with the pass-through share used in the argument on bills: if more reaches customers, less is lost here, and the two cannot both be large. What is not counted here is anything about who owns the plants beyond the weight, because ownership patterns change and the credits are not written to favour anyone. The Impact is small against the payment it comes from, because the gap in what a euro is worth at either end is narrow and only the gap is counted.

▸ Show calculation ▾ Hide calculation

	Share of the credit that does not reach customers the complement of the pass-through used in the argument on bills [2]	half of 15.8 billion euro	7.9 billion euro
×	Difference in what a euro is worth the money moves from a public budget, where this site counts a euro at 1.0, to the top tenth of wealth holders, where it counts at 0.5	1.0 minus 0.5	3.95 billion euro
÷	Normalised Impact scale of this evaluation	2 billion euro a point	1.98

Score 1.98 Impact × 5 Value × 3.5 Plausibility ÷ 10 = 3.5 of 100

The direction is not disputable: a subsidy that does not fully pass through stays with the recipient. The counterfactual is the same market without the credits, and the quantity is entirely determined by the pass-through share taken from the argument on bills, so this figure is as firm or as loose as that one. Where the two differ is that a pass-through estimate for one purpose is not automatically right for the other: some of what does not reach customers is competed away in higher land rents, equipment prices and interest costs rather than kept as profit, and that portion lands on other parties whose weights are not the top tenth. Nothing measures that split. The confounder is the same state retail regulation named above, which can hold customer prices fixed while wholesale prices fall, sending the whole saving to owners. Reverse causation does not arise. The Plausibility is below the middle because the split between profit and competed-away cost has never been measured for these credits.

evidence basis: Mechanism · P ceiling 6 identification: Mechanistic · rung ceiling 6 band: Chain open · P 3–3.5

Counterfactual: the same market without the credits. Design: mechanistic — the share is the complement of the pass-through taken from pro-2 and inherits its grounding; the further split between owner profit and costs competed away into land, equipment and finance is unmeasured. Confounder: state retail regulation holding customer prices fixed while wholesale prices fall, which would send the whole saving to owners; unresolved. Direction: no reverse causation. Ceiling: mechanistic 6.0 binds. Band: chain open, because the profit-versus-competed-away split carries the quantity and is unchecked. Parameter coupling: the pass-through share is one number shared with pro-2 and moves inversely there.

The chain is named but the link carrying the quantity — how much of the retained share is profit rather than cost competed away into land, equipment and finance — is unchecked, and the retail-regulation counter-mechanism is unresolved. Read back: about a third of the time, roughly half the credit stays with owners at the weight assumed here.

Open: Project-level returns for credit-supported plants against comparable unsupported ones would show how much of the credit is kept rather than competed away, and could carry P to 6.

The grid has to carry it

1.9of 100

Wind and solar produce when the weather allows, not when demand asks. Every additional unit of them needs transmission to move it, storage to shift it, and firm capacity to stand behind it. None of that is paid for by the credit.

Value 5 · Public financesImpact 1.1Plausibility 3.5

▸ Show reasoning & sources ▾ Hide reasoning & sources

The stream is real resources — copper, steel, batteries, gas turbines held in reserve — consumed to make variable generation usable. It is priced at the middle of the scale like any other money, and it is a genuine cost rather than a transfer, because the equipment exists and is paid for. Who pays is a mixture of network charges on bills and public investment, and both carry the same weight. The generation itself is not counted here as a cost, because it is what the measure is for and its value is on the other side. The value is the middle of the scale, because the stream is money spent on real equipment.

Restoring the credits raises clean capacity additions by something over half against what the 2025 law leaves, which is roughly 30 gigawatts a year of wind and solar ^[1]. At the capacity factors those technologies achieve, each year's addition produces about 79 terawatt hours, and the stock of additional generation builds up across the decade to something over 400 terawatt hours a year, averaging about 215. Estimates of what it costs to integrate variable generation — transmission, storage, and firm capacity held available — cluster around 10 euro per megawatt hour at these penetrations, in a range from 4 to 25 and rising steeply as the share grows. That gives 2.15 billion euro a year. The interconnection queues already hold 2,060 gigawatts, more than twice the installed fleet, which is a sign that the constraint is real and that the cost of relieving it is not zero ^[6]. The Impact is small against the credit itself, which is the ordinary finding at these penetrations and would not stay true at much higher ones.

▸ Show calculation ▾ Hide calculation

	Additional wind and solar the credits bring forward [1]	roughly 30 gigawatts a year	30 gigawatts a year
×	Additional generation, averaged across the decade as the stock builds	about 79 terawatt hours a year added, averaging 215 in place	215 terawatt hours a year
×	Cost of integrating it Setting, range 4 to 25 euro: transmission, storage and firm capacity held available; the published estimates were made at lower penetrations than these [6]	10 euro a megawatt hour	2.15 billion euro
÷	Normalised Impact scale of this evaluation	2 billion euro a point	1.08

Score 1.08 Impact × 5 Value × 3.5 Plausibility ÷ 10 = 1.9 of 100

The mechanism is engineering rather than behaviour and is not in dispute. What is contested is the number, and it is contested by an order of magnitude. Integration costs depend on the existing mix, the transmission already in place, and how much storage is built alongside, all of which vary enormously between American regions, and the published estimates were made for penetrations lower than the ones this measure would produce. The counterfactual is the same grid without the additional variable generation. The confounder that matters runs against this argument: the same credits support storage, which is one of the things integration requires, so part of the cost counted here is paid for by the measure itself. That is named and not resolved, and it is why the figure sits at the low end of the published range. Reverse causation does not arise. The Plausibility is below the middle because the integration cost at these penetrations has not been measured and the offsetting storage support is unresolved.

evidence basis: Mechanism · P ceiling 6 identification: Mechanistic · rung ceiling 6 band: Chain open · P 3–3.5

Counterfactual: the same grid without the additional variable generation. Design: mechanistic — chain named (more variable generation → transmission, storage and firm backup needed → real cost), with the cost per unit carried over from studies at lower penetrations. Confounder: the same credits supporting storage, so part of the integration cost is paid by the measure itself; named, unresolved, and the reason the figure sits at the low end. Direction: no reverse causation. Ceiling: mechanistic 6.0 binds. Band: chain open, because the cost per unit at these penetrations is unmeasured and the storage offset is unanswered.

The chain is named but the integration cost per unit at these penetrations has never been measured, and the counter-mechanism — that the same credits pay for the storage integration requires — is unanswered. Read back: about a third of the time, integrating this much variable generation costs roughly what is assumed here.

Open: Regional transmission operators publish the cost of network upgrades and capacity procurement attributable to new generation. Matching those against interconnected capacity would give the figure directly and could carry P to 6.

Summary

The striking thing about these credits is how close the arithmetic runs. They cost about sixteen billion euro a year and avoid about 160 million tonnes of carbon dioxide, which works out at almost exactly the hundred euro a tonne this site uses as the price of avoiding a tonne elsewhere. On carbon alone, in other words, they are neither a bargain nor a waste — they are the market rate. What tips the balance is everything that arrives with the carbon: bills that rise less, in households where electricity is a large share of a small budget, and cleaner air in the counties downwind of the plants that would otherwise run. Against that, half the money stays with the owners of the plants rather than reaching anyone's bill, and the grid has to be built out to carry what the credits induce. The result is a clear but not overwhelming case, and it rests on a carbon price that American agencies themselves put considerably higher.

Outlook — effect over time

Better for the future · 0.68 previous scale

Sources

Rhodium Group: What Passage of the One Big Beautiful Bill Means for US Energy and the Economy. rhg.com
Bipartisan Policy Center: 2025 Reconciliation Debate: Energy Provisions. bipartisanpolicy.org
Congressional Research Service: IRA Tax Credit Repeal in the FY2025 Reconciliation Law. congress.gov
Energy Innovation: The Economic, Consumer Cost and Pollution Impacts of Federal Energy Policy Changes. energyinnovation.org
U.S. Energy Information Administration: U.S. Energy-Related Carbon Dioxide Emissions, 2025. eia.gov
Lawrence Berkeley National Laboratory: Queued Up: Characteristics of Power Plants Seeking Transmission Interconnection. emp.lbl.gov

Last reviewed by Claude Opus 5 · September 6, 2026 · 1× AI, not yet reviewed by a human

September 6, 2026AI reviewClaude Opus 5First evaluation
Created for the English side: the credit cost and the avoided tonnes come out at almost exactly the site's carbon price, so the balance rests on bills and air quality.

Evaluations are produced with AI support and reviewed on a schedule for new developments; human passes are marked separately.How we review →