DOE Didn't Ban Your Appliances. It Raised the Bar for the Next Rule, by Two to Six Times

On July 2, the Department of Energy announced it would 'permanently end home appliance and equipment mandates.' The press release named names: air conditioning units, gas stoves, washing and drying machines, water heaters, refrigerators, 'and other products Americans rely on every day.' It reads like a repeal.

It is not one. The Notice of Proposed Rulemaking (NOPR), a formal draft rule published for public comment (RIN 1904-AF72, docket EERE-2025-BT-STD-0001), cancels no standard currently in force. Your refrigerator rule, your dishwasher rule, your water-heater rule all survive intact. What the document rewrites is the arithmetic DOEDOEThe U.S. Department of Energy, the federal agency that sets minimum efficiency standards for home appliances. uses to judge every future standard: a change to the agency's 'Process RuleProcess RuleDOE's internal rulebook that decides how and when a new appliance-efficiency standard can be considered, the rules for making rules.,' the internal rulebook that decides when a new efficiency requirement is even allowed to be considered.

That gap, between a press release that bans and a rule that only changes the math for the next rule, is a familiar shape. The loud claim describes the future; the binding mechanism is a definition. So we asked one narrow question and went to the text to answer it, and we caught two things the face numbers hide.

When DOE lifts the floor from 0.3 to 2 quads, how much does the screen actually tighten once you correct for a change of units and for the fact that the test is an 'or', and which products does it newly block?

The short answer: the floor did rise, but not the clean 6.7x the face numbers suggest, and not across the whole program. Two corrections cut the drama down to something real but bounded. Here is how we got there.

Every standard DOE sets must, by statute, the Energy Policy and Conservation Act, or EPCA, produce a 'significant conservation of energy.' EPCA never defines 'significant.' That single undefined word is the lever. Whoever defines it controls which rules can proceed. The point is not academic: a threshold placed high enough disqualifies a proposed rule before DOE even runs the costly engineering and economic analysis.

The 2020 Process Rule, finalized in the first Trump term (85 FR 8626), filled the gap with a bright line: to count as significant, a standard had to save 0.3 quadrillion BTUs ('quads') of site energysite energyThe energy an appliance actually uses at your home, at the plug or the burner. over 30 years, or cut a product's site-energy use by at least 10%. A quad is about 1% of everything the US uses in a year. The Biden DOE deleted that line in December 2021 and reaffirmed the deletion in later action. The obvious question was whether 2026 simply restores 0.3, or sets something harder.

We read the proposal. Section IV.F.2 proposes to treat 'significant energy savingssignificant energy savingsThe legal threshold a proposed standard must clear to be worth pursuing; EPCA never defines it, so DOE sets the number.' as at least a 10-percent reduction in full-fuel-cycle energy over 30 years, or a 2-quad reduction in full-fuel-cycle energy over 30 years. The quad prong moved from 0.3 to 2.0. On its face, that is a 6.7-fold increase (2.0 divided by 0.3). DOE also says it may consider alternatives, inviting comment on thresholds anywhere from 0.5 to 5 quads, or from 5 to 15 percent.

So 2.0 is nearly seven times higher than 0.3 on the quad prong. But before we could say how much deeper the new screen reaches, we had to catch two things the raw comparison hides, one that shrinks the drama, one that narrows who it hits.

The first correction cuts against the drama. The 2026 floor is explicitly measured in full-fuel-cycle (FFC) energyfull-fuel-cycle (FFC) energySource energy plus everything else lost upstream: fuel extraction, processing, and transport (and, for gas, pipeline leakage), always a bigger number than site energy., which adds everything upstream of your home: the fuel burned at the power plant, generation and transmission losses, extraction, and processing. FFC is always a bigger number than the site energy an appliance uses at the plug or the burner.

The honest comparison therefore hinges on how the 2020 floor was measured, and here the proposal answers its own question. Its own history section states that the 2020 rule set the threshold at '0.3 quads or 10 percent site savings over 30 years' (85 FR 8675). The old floor was site energy; the new floor is full-fuel-cycle energy. That is not our inference; it is the text. It also means the drama is real but smaller than 6.7x, because to compare like with like you must first inflate the old 0.3 to its FFC equivalent.

By how much depends on the fuel. For natural gas, source energy roughly equals site energy, so 0.3 quads of site energy becomes only about 0.33 quads FFC once upstream losses are added. For grid electricity, the conversion is far larger, roughly three times, because most of the fuel's energy is lost turning it into power and moving it, so 0.3 site becomes about 0.9 quads FFC. The 2020 floor, in the new currency, is therefore a band from about 0.33 to about 0.9, not a single 0.3.

Divide the 2.0-quad proposal by that band and the real tightening lands between roughly two times (for an all-electric product) and about six times (for a gas product), not the 6.7x the unit-blind arithmetic implies. That is the corrected headline: two to six times, with a central estimate near three. A meaningful lift, but not the order-of-magnitude jump the face numbers advertise.

The second correction narrows who gets caught. The screen is not a single hurdle. It is two hurdles joined by 'or': a standard is significant if it saves 2 quads FFC OR cuts energy use by 10 percent. A proposed rule fails only if it misses both.

That structure matters, because raising the quad number changes nothing for a rule that clears the percentage prong. A big-percentage efficiency jump on a small product still qualifies, no matter how modest its absolute quads. What the higher quad floor newly blocks is a specific middle band: rules that save a moderate but no-longer-'significant' absolute amount, above the old floor, below 2 quads, and also come in under 10 percent. High-absolute, low-percentage rules on big energy users, like some water-heater or central-air standards, are exactly the shape that can slip into that band.

So the change is not a wall across the whole program. It is a trapdoor under one quadrant of it. To see how large that quadrant is, we turned to DOE's own accounting.

DOE did the counting for us. In Section IV.F.2 it reports that a threshold of '10-percent reduction in FFC energy use over a 30-year period or a 2 quad reduction in FFC energy use over a 30-year period would retain 91.5 percent of the energy savings from the program while eliminating 35 percent of the rulemakings (30 of 86 rulemakings).' The blocked rules, it adds, contributed 'less than 9 percent of the total energy savings.'

Read those two numbers together and the middle-band story holds. The screen reaches wide across rulemakings, roughly a third of them, but shallow across savings, taking under a tenth. That is the mathematical signature of a screen aimed at low-yield rules: the ones it removes are, almost by definition, the ones that saved little and missed the percentage prong too. For comparison, DOE found the gentler 2020 site-energy floor would have retained 'over 95 percent' of savings. The new floor keeps somewhat less, 91.5 percent, which is the price, in savings, of the higher bar. Real, but bounded.

One caveat DOE itself flags: failing the screen in one rulemaking does not bar a future one for the same product. Its example is walk-in coolers, where a 2017 rule (0.85 quads FFC) and a later rule (1.6 quads FFC) each fell short alone, but together cleared 2 quads. Savings can be banked and combined, so 'blocked' often means 'deferred until the case is bigger,' not 'gone.'

A second lever gets far less attention, and it lives in a different file. Even a rule that survives the significance screen still has to be shown to be worth its costs, and that calculation runs on a discount ratediscount rateThe percentage used to convert future dollars into today's dollars; a higher rate makes future bill savings look smaller now., the percentage that converts future bill savings into today's dollars. Raise the rate and the same physical saving simply looks smaller on paper.

Crucially, this proposal does not set a discount rate. The rate and the broader valuation methods sit in a companion Analytic Methods Request for Information (FR doc 2026-13673), a document that asks questions rather than proposing numbers. It reviews the life-cycle-cost and national-impact analyses that rest on a discount rate, and it aligns itself with OMB Circular A-4OMB Circular A-4The federal guidance for how agencies do cost-benefit analysis; it has historically told them to discount future costs and benefits at both 3% and 7%., the cost-benefit guidance that has historically used both 3% and 7%. But it proposes no figure. Our chart above is an illustration of the mechanism, not a claim about what DOE will do.

The two levers also run on different clocks. The threshold that can mechanically block a rule gets a 30-day comment window; the methods RFI that could quietly devalue surviving rules gets 60 days, closing September 8, 2026. Anyone tracking the fight has to watch both dockets, on two timers.

Step back and the stakes clarify. Nothing you own is affected today. Existing standards remain, and EPCA's anti-backslidinganti-backslidingAn EPCA provision barring DOE from weakening an appliance standard once it is in force. provision legally bars DOE from weakening a standard once it is in force. For scale, RMI estimates that standards already on the books saved US households and businesses about $105 billion in 2024 alone. None of that is on the table here.

Because the in-force program is fenced off by law, the entire contest is over rules not yet written. And on that terrain the honest verdict is neither the press release's 'permanent end' nor a clean 6.7x tightening. It is this: a two-to-six-times higher bar, on a full-fuel-cycle basis, joined by 'or' to an unchanged 10-percent prong, that would have blocked about a third of past rulemakings while sacrificing under a tenth of their savings, with the sharper, quieter lever, the discount rate, still only a question in a separate file.

DOE argues it is on firm legal ground, citing NRDC v. HerringtonNRDC v. HerringtonA 1985 appeals-court decision that lets DOE define 'significant' savings, including as a percentage of a product's energy use, as long as the level reasonably serves EPCA's purposes. (which lets it set a significance threshold) and Loper Bright Enterprises v. Raimondo (which says courts, not agencies, now read ambiguous statutes). Whether the specific 2-quad number survives comment and litigation is unknown. What is knowable now is exactly what it does and does not do, and that turns out to be a narrower, more precise instrument than either its supporters or its critics have described.