Three ways the answer is no

This is the most fundamental failure and it has nothing to do with price or size. Moving a kilowatt-hour through a pack rated 90 percent ^[3] means buying about eleven percent more energy than you get back. If the expensive hour is not priced above the cheap hour by at least that margin, every cycle ends behind where it began.

On a genuinely flat tariff with the pack charged from the grid, the same price goes in and comes out, so the margin is zero and every cycle loses the round trip. Charged from solar surplus instead, the pack gives up the export credit that surplus would have earned, and if that credit sits below the flat price by more than the round-trip loss the model finds a margin even on a flat tariff; if it does not, the loss returns. On a shallow time-of-use plan the margin can be positive but too small to matter. Wherever it cannot cover the loss, the correct output is a sentence, not a large number of years. The remedies are a different tariff, a cheaper source to charge from, or, close to the line, a pack that loses less in the round trip, because the break-even price is the charging price divided by that efficiency. Our arbitrage tool prints the break-even peak price so this can be tested in about ten seconds against your own bill.

It is worth being explicit about what this means geographically, because a state’s average price answers a different question. Here are the five dearest and five cheapest jurisdictions for residential electricity in May 2026, from the federal statistics agency’s archived monthly release, with the median for scale:

In May 2026 the median jurisdiction sat at 16.3¢/kWh ^[6]. The live version of that federal table has since moved on to later months; these figures stay at May 2026 and link to the archived issue that printed them. Now notice what this table cannot tell you: not one of these figures is a spread. A household in the dearest jurisdiction on a flat rate has no gap between two grid prices to work with, so a pack charged from the grid cannot gain there, while a household in a mid-priced one with a steep evening peak can. Average price is the wrong axis, and a calculator that sizes savings from it is measuring the wrong thing.

That is not a criticism of the source. The federal table is doing exactly what its own title scopes it to do: one price, for one state, for one customer sector, for a named month, in cents per kilowatt-hour. Every one of those qualifiers is load-bearing. What the table cannot offer, because averaging a month of days removes it, is the gap between two hours of the same day — and that gap is what the grid-charging comparison in this section turns on. (A pack charged from solar surplus is compared with the export credit it gives up instead, and an outage value you enter is counted separately.) No amount of care with a state figure recovers a number that was taken out before publication.

Suppose the tariff does work. This calculator still sets a ceiling on what the pack can earn within its warranty, and it comes from the warranty’s cycle count. 6,000 covered cycles ^[1] at 10 usable kilowatt-hours ^[2] is a fixed quantity of energy, and that energy is worth at most the widest margin on your tariff’s published prices, on every single cycle, with perfect dispatch and no missed days.

Multiply those together and compare against the net price. If the price is higher, the battery cannot repay it from bill savings within its warranted cycles at those prices, however it is dispatched. That is an upper bound within this model rather than a forecast, and the model’s own estimate sits below it because capacity fades and not every day earns the widest spread. It does not count cycles after the warranty ends, which a pack may still deliver at reduced capacity without cover, or a later change in the tariff.

The payback worksheet prints this ceiling next to your net cost for exactly this reason. It is the fastest way to discover that a quote is not merely expensive but outside what the equipment can return within its warranted cycles at the published prices the calculator uses, and it takes one multiplication.

That inverts the usual advice about buying bigger. Both packs listed here carry 6,000 covered cycles — the smaller at 5.0 kWh usable ^[5] and the larger at 10.0 kWh. Warranted throughput therefore scales with capacity while the cycle budget does not, which sounds like an argument for the larger unit and is one only if the evening consumes it.

In this model, capacity above the window adds to the price and to the ceiling without adding a single earned kilowatt-hour. The ceiling test therefore cannot tell an oversized pack from one sized to the window at the same price per kilowatt-hour; the year-by-year payback, which counts only what the evening uses, is where the oversized pack falls behind.

This is the only one of the three that is genuinely a judgement call. The arithmetic clears, the cumulative saving does eventually overtake the cost, and the question is whether the date is one you are prepared to underwrite.

Two things push it later than a flat annual saving would suggest. Capacity fades, so a flat annual saving overstates the later years. And the warranty ends — on a term, or on a cycle count, whichever comes first — so a payback landing after that point is being paid by an uncovered asset that may need replacing before it gets there.

Our rule of thumb, stated as ours: a payback date beyond the warranty is not a payback. If the arithmetic only clears in year eighteen on a fifteen-year warranty, the honest reading is that the savings case has failed and the purchase needs a different justification.

Quite a lot, and it is worth saying so plainly rather than leaving the impression that storage is a mistake. A battery keeps a freezer cold and a medical device running through an outage, silently and with no fuel to store. It removes the exposure to evening price rises that a household on a steep peak otherwise carries indefinitely. It allows an all-electric home to ride through the hours when importing is most expensive.

None of those enters the payback figure unless you put a value on it yourself: the payback worksheet adds an outage value only when you enter outage hours and what an outage-served kilowatt-hour is worth to you, and it adds nothing by default. If you do, keep that value visible rather than letting it work as a fudge factor, and be honest with yourself about which argument is actually carrying the purchase.