Methodology

Simulate years, don't divide. Dividing up-front cost by first-year savings is the standard shortcut and it is wrong in both directions: it ignores equipment degradation, which pushes payback later, and energy-price inflation, which pulls it earlier. Every payback figure here comes from a year-by-year cash-flow loop.

Charge for the losses. Energy is lost converting, storing and transmitting it. Models that quietly ignore round-trip and charging losses overstate savings by 10–15% before any other assumption is made.

Publish the unfavourable answer. Where the maths says a purchase does not pay back, the page says so. See our editorial policy.

System size is set to cover a year of electricity use, then costed and run forward for 25 years. The default baseline is EIA's state-level average monthly consumption per residential customer — the metric EIA actually publishes. It is not a per-household figure, and no state average describes any particular home: a large all-electric house can use several times what a small flat does in the same state. Where you know your own usage or bill, enter it — the calculator recomputes from your value and labels it as yours.

Exported power is the second-biggest lever after the electricity rate. About 60% of what a rooftop system generates goes to the grid rather than into the house, so what an exported kilowatt-hour earns is close to half the answer. Two different questions decide it, and this site keeps them apart because conflating them is how solar economics get overstated:

• The policy structure. Retail-rate net metering credits exports at the same price you pay. Net billing credits them at roughly what the utility avoids spending to generate the power. Some states set rates utility by utility, and some have no statewide rule at all. A programme called “net metering” does not necessarily pay the retail rate.
• The money. Where compensation is not at the retail rate we model it in cents per kilowatt-hour, not as a percentage of retail. Avoided cost tracks wholesale generation value; it does not rise because a state's retail rate is high. Treating it as a fixed share of retail overstated export value most in exactly the states where retail power is most expensive.

The modelled baseline of 6¢/kWh is anchored on the published residential export tariffs we verified with a regulator — Arizona's Resource Comparison Proxy rates. It is a baseline, not a measurement of your utility. Export compensation can vary by utility, tariff, installation date and customer class; where a state runs time-varying export rates, no single number reproduces them, and the figure is labelled modelled for that reason. Every state page shows which structure applies and what we assumed, and the slider lets you substitute your own rate.

Two details matter. Installed cost per watt falls as the system grows, because permitting, design and labour are near-fixed: a 3 kW roof genuinely costs more per watt than a 12 kW one. We scale from a 8 kW benchmark. And the annual benefit is capped at the customer's whole bill, because utilities do not write cheques for net-positive generation.

Fuel cost only — this compares what it costs to move the car, nothing else.

The division by 0.9 is the part most comparisons omit: you pay for the electricity that leaves the wall, not the amount that reaches the battery. Blending in public charging matters just as much — DC fast charging runs several times the residential rate, which is where the “my EV isn't actually cheaper” complaints come from.

Excluded: purchase price, depreciation, insurance, maintenance, tax credits, and the EV registration surcharges several states levy in place of fuel tax. Those often matter more than fuel — this sizes the fuel gap so you can weigh it against the price difference yourself.

Annual heat demand is derived from heating degree days and a modelled building heat-loss rate, then met two ways.

The full state-by-state consequence of this choice, including a test of the switching condition in every jurisdiction, is set out in heat pump vs gas furnace economics by state.

That climate-varying COP is the single biggest reason published heat-pump savings fail to show up on real bills. A unit works harder per unit of heat as it gets colder, and leans on resistance backup during cold snaps.

Excluded: the cooling benefit (a heat pump replaces your air conditioner too, which can flip the case), rebates, maintenance differences and any electrical panel upgrade.

A battery holds one charge and returns it about once a day. It can arbitrage a time-of-use spread or store solar you would otherwise export cheaply — not both with the same stored kilowatt-hour.

Dividing by 0.9 is deliberate: to deliver one usable kilowatt-hour you must buy about 1.11. Omitting that overstates arbitrage by roughly 11.1%.

Three value streams, never added together. A battery can arbitrage a time-of-use spread, or store solar it would otherwise export, but not both with the same stored kilowatt-hour — so the model takes the larger of the two, not the sum. Backup power is a third stream and is priced separately, never inside payback.

Storing solar is worth what you did NOT sell it for. This is the part most battery maths gets wrong. If power costs 30¢ and exports earn 6¢, moving one exported kilowatt-hour into later household use does not create 30¢ of value — you were already earning 6¢ for it. The gain is the retail purchase you avoid minus the export income you give up, and because delivering one kilowatt-hour means diverting 1.11 into the battery, that is how many exports you forgo.

One consequence is counter-intuitive and worth stating plainly: the less your utility pays for exports, the more a battery is worth. Full retail net metering is the arrangement that makes a battery hardest to justify, because the grid is already storing your power for free. Solar and battery read the same export compensation for a given state — they are pricing the same exported kilowatt-hour, once as income and once as an opportunity cost — so the two calculators cannot disagree about what it is worth.

Backup power is valued separately and never folded into payback. Keeping the lights on during an outage is real value, but it is insurance, not a return. Each state page lets you price an hour of kept power yourself and shows the result beside — not inside — the payback figure.

Where a page shows a plain-language verdict, the thresholds are tied to the product's own lifetime rather than invented bands. A payback is only meaningful relative to how long the equipment lasts.

For batteries the reference is the 10-year warranty, not the analysis horizon — a payback that arrives after the warranty expires is not a payback you can rely on.

Incentives move these numbers more than almost any modelling choice, and they change frequently. Our position is to model only what we can state precisely, and to say plainly what is excluded.

Federal residential clean-energy credit: 0% for new 2026 installations. Public Law 119-21, enacted 4 July 2025, terminated §25D for expenditures made after 31 December 2025. Under §25D(e)(8)(A) an expenditure is treated as made when the original installation is completed — so paying in 2025 for a system commissioned in 2026 does not preserve the credit. A system finished in 2025 is unaffected and is claimed on the 2025 return, and unused credit from an earlier qualifying year may still be carried forward; neither is a credit for a new 2026 installation, which is what this site models.

What the baseline represents: A homeowner who buys and owns a residential solar system whose installation is completed in 2026. Business clean-electricity credits such as §48E continue on their own timetable, but they are claimed by the owner of the system. In a lease or PPA the installer owns the equipment and may or may not pass the value through in the rate — that is not modelled here, and it is not a credit the homeowner receives.

State and utility incentives are not modelled by default — they are too local and change too often for a national dataset — but the battery and heat-pump calculators let you enter a rebate amount and see its effect immediately.

Proposed legislation is not modelled. H.R. 5862 (119th Congress), introduced 28 October 2025, would reverse the termination. It has not been reported out of committee or passed either chamber, so it is not reflected in these figures. Payback Math publishes the current-law baseline, never a forecast of what Congress might do.

Verified against Internal Revenue Service guidance on 2026-08-21 (Public Law 119-21, enacted 2025-07-04). Tax eligibility depends on individual circumstances: this is the baseline federal treatment for a typical homeowner-owned installation, not tax advice.

Nothing here should be read as a guarantee that any rebate is available to you. Confirm current programmes before spending money.

These are estimates from state averages, not quotes for your property. Results move with future energy prices, your actual usage, equipment and installation cost, financing, maintenance, local climate and incentives. The calculators expose the inputs that matter most precisely so you can test how sensitive the answer is to each one.

Where a figure is an informed estimate rather than a sourced measurement, our data sources page says so explicitly, field by field.