Batteries and backup power

Do you need a battery? The three jobs storage actually does

You do not need a battery for solar panels to pay: Lawrence Berkeley National Laboratory counted 37 percent of new US residential solar systems in 2025 installed with one, so about 63 percent went in without, and a battery earns its place only when it has a job. There are three: keeping essentials running in an outage, avoiding exports your utility pays little for, and shifting solar energy into expensive evening rate periods. If your project has none of those problems, a battery is a costly solution looking for a question. This guide works out what one stored kilowatt-hour is worth under three export rules, with every input printed, so you can run the same test on your own tariff before anyone prices a battery for you. It is planning arithmetic, not financial advice, and it never counts the federal residential credit, which ended for systems placed in service after 2025.

The outage case

Backup is insurance when outages are rare and short, and a different calculation when they are long or someone depends on powered medical equipment. A standard grid-tied system shuts down in a long or large grid disruption, the Department of Energy explains, so the backup value comes entirely from the battery being designed to island the house. The outage guide covers why panels shut off, federal outage-length data, runtime arithmetic and the other two backup routes.

Price the backup job as what it is: insurance with a known premium. In the Q1 2025 benchmarks published by the U.S. Department of Energy, the modelled 8 kW rooftop system is $23,600 on its own and $36,720 with a 13.5 kWh battery, so the storage line is $13,120 (modelled market price, 2024 dollars). The national laboratory's Annual Technology Baseline analyses a residential battery over a 15-year lifetime. Spread $13,120 across those 15 years and the backup costs about $875 a year before any financing cost, any incentive, or any bill savings the battery also earns. That is the number to hold against your own outage history. A household that lost power for a few hours twice in a decade is paying $875 a year to avoid a short inconvenience. A household with a well pump, a home dialysis machine, or a freezer full of insulin, in an area where storms take the grid down for days, is buying something a monthly bill cannot measure.

Size the backup job from the loads, not the brochure. A critical-loads panel that carries the refrigerator, lights, Wi-Fi, a furnace blower and a medical device draws far less than whole-home backup, and the difference shows up in both the battery count and the electrical work. If backup is the only job, also check whether the battery will sit full most of the year waiting for an outage, or whether the installer is assuming daily cycling for savings it will not deliver on your tariff. Those are two different products, sold under one name.

The export case

Where export compensation sits far below the retail rate, storing a midday kilowatt-hour for evening use can be worth several times what selling it earns. This is the strongest pure-math case for batteries today, and it depends entirely on your utility tariff. Run the numbers against your actual export rate, not a national story; the California net billing battery arithmetic is one worked case.

The arithmetic is short once the inputs are on the table. A kilowatt-hour your panels export earns the export rate. The same kilowatt-hour stored and used that evening avoids buying a kilowatt-hour at retail, but only 85 percent of it comes back out of the battery, the representative round-trip efficiency in the national laboratory's 2024 Annual Technology Baseline. So the value of shifting is the retail rate minus the export rate divided by 0.85, per kilowatt-hour delivered. At the labelled inputs the net metering guide uses, a 30-cent retail rate and a 4-cent export rate, that is 30 cents minus 4.7 cents, or about 25.3 cents for every kilowatt-hour the battery delivers. The net metering guide explains the three export structures behind those rates.

The time-of-use case

Where evening electricity costs far more than midday power, a battery buys low and uses high without exporting anything. The value is the rate spread times the energy you actually shift, minus round-trip losses. Small spreads make slow paybacks.

The round-trip loss sets a hard floor under the spread. To deliver one kilowatt-hour in the evening, the battery has to take in 1 divided by 0.85, about 1.18 kilowatt-hours at midday. So shifting only breaks even when the evening price is more than the midday value divided by 0.85, which is 17.6 percent above it. With a labelled 25-cent midday value, the evening rate has to beat about 29.4 cents before a single cycle earns anything. A 45-cent evening rate clears that floor by about 15.6 cents per kilowatt-hour delivered. A 28-cent evening rate does not clear it at all, and a battery cycling daily on that tariff loses a little money every day while looking busy in the monitoring app.

One stored kilowatt-hour, valued under three export rules

The cleanest way to see which job your tariff gives a battery is to value a single kilowatt-hour of shifting under each export rule, using the same household, the same panels and the same battery. The inputs here are labelled illustrations, not anyone's current rates: a 30-cent retail price, a 4-cent export rate for net billing, a time-of-use plan charging 45 cents in the evening and crediting 25 cents at midday, and the 85 percent round-trip efficiency from the 2024 Annual Technology Baseline. Each case asks the same question. If the battery stores midday solar that would otherwise have been exported, and delivers it in the evening instead of grid power, how much better off is the household per kilowatt-hour delivered?

Under net billing, the answer is about 25.3 cents: the household avoids a 30-cent import and gives up an export worth 4 cents, grossed up by the round-trip loss to about 4.7 cents. Under the time-of-use plan, it is about 15.6 cents: a 45-cent evening import avoided, less a 25-cent midday credit grossed up to about 29.4 cents. Under one-for-one retail net metering, the answer is negative, about minus 5.3 cents. The export would have earned the full 30-cent retail credit anyway, so storing it gains nothing, and the 15 percent lost in the round trip is a pure cost. That last result surprises people, but it is the core of the decision: where exports are credited at retail, the grid is already acting as a lossless battery, and the physical battery is competing with it.

What one kilowatt-hour delivered from a battery is worth under three export rules Horizontal bars of the value per kilowatt-hour delivered from storage, labelled illustrative inputs with 85 percent round-trip efficiency: net billing 25.3 cents, time-of-use 15.6 cents, retail net metering minus 5.3 cents. -50510152025cents per kWh deliveredNet billing (4-cent export)25.3Time-of-use (45 cents evening, 25 midday)15.6Retail net metering (30 cents both ways)-5.3: the round-trip loss
What one kilowatt-hour delivered from a battery is worth under three export rules Labelled illustration, not current rates: retail 30 cents, net billing export 4 cents, time-of-use 45 cents evening and 25 cents midday credit. Value per kWh delivered = evening price minus (midday value divided by 0.85). No incentive and no federal credit included. Source: round-trip efficiency of 85 percent from National Laboratory of the Rockies (formerly NREL), 2024 Annual Technology Baseline, Residential Battery Storage, checked October 29, 2026; arithmetic by SolarRoost.
Show the three cases as a table
Export ruleMidday value given up (after 15% loss)Evening import avoidedValue per kWh delivered
Net billing4 cents / 0.85 = 4.7 cents30 cents25.3 cents
Time-of-use25 cents / 0.85 = 29.4 cents45 cents15.6 cents
Retail net metering30 cents / 0.85 = 35.3 cents30 centsminus 5.3 cents

Scale it to a year and set it against the storage line. Suppose the battery delivers 10 kilowatt-hours every evening, which means storing about 11.8 kilowatt-hours of midday solar and losing about 1.8 in the round trip, for 365 days. Under net billing that is about $923 a year. Under the time-of-use plan, about $569. Under retail net metering, about $193 a year worse off than exporting. Against the $13,120 storage line in the Q1 2025 benchmark, the simple paybacks are about 14.2 years under net billing and 23.1 years under the time-of-use plan, and there is none under retail net metering. The Annual Technology Baseline analyses these batteries over 15 years. So even the strongest case here pays back just inside the analysis life on bill savings alone, before financing, before any capacity fade, and before an incentive that might shorten it.

Two cautions keep this honest. Ten kilowatt-hours every evening of the year is generous: winter days may not leave 11.8 kilowatt-hours of surplus solar to store, and a household away on weekdays may not use 10 in the evening. A home with no grid connection has to plan around exactly those short days, which is why the off-grid solar cost guide sizes its panels to December output rather than the yearly average. Fewer delivered kilowatt-hours mean a longer payback, in direct proportion. And the payback moves with the price you actually pay for the battery, which the home battery cost guide breaks into line items. Put your own rates, your own evening use and your own battery price into the same three lines. If the result is a payback longer than the warranty, the battery is a backup purchase that happens to earn something, and it should be judged that way.

Why a battery can lose money under retail net metering

Under one-for-one retail net metering, an exported kilowatt-hour earns a credit worth a kilowatt-hour of imports later. The grid, in effect, stores your surplus without loss and hands it back at the same price. A physical battery cannot beat that on bill arithmetic, because it returns about 85 percent of what it takes in. That is why the negative bar in the chart above is not a quirk of the inputs: on a flat-rate tariff that credits exports at the same price you pay for imports, at every hour, shifting your own solar through a battery costs you the round-trip loss on every kilowatt-hour you cycle. On a time-of-use rate the midday credit and the evening price differ even with one-for-one credit, so the spread test in the time-of-use section applies instead.

There are real exceptions, and each one changes the job rather than the arithmetic. Where retail credits expire monthly or at an annual true-up, surplus that would otherwise be forfeited has no value to lose, so storing it costs nothing extra; the true-up guide explains how those settlements work. Where a utility program pays for battery dispatch, the battery has a second job with its own payment, covered in the utility battery program guide. And outage backup is worth the same on any tariff. What does not hold is the sales claim that a battery boosts savings under retail net metering because you use more of your own solar. On that tariff, using more of your own solar through a battery is worth less than exporting it.

If you already have a retail net metering agreement, also check whether adding storage changes it. PG&E's Advice Letter 5559-E, approved in 2019, is one example of a tariff setting conditions for existing net metering customers who add storage, and the grandfathering guide reads what the rules say before you assume either way.

Storage cannot rescue a weak solar quote

Buying storage to fix a payback problem created by an oversized system or a bad quote. Storage raises project cost; it does not rescue weak solar economics.

The decision in three questions

Put the three jobs in order and the decision gets short. Ask about safety first, because it is the one job no tariff can price. Then ask whether exports are paid well below imports. Then ask whether the evening price clears the midday value by more than the round-trip loss. The first "yes" names the battery's job, and the job tells you how to size and judge it. If every answer is no, the battery is not earning its place yet, and the cheaper move is to keep the option open: a battery-ready inverter or pre-wiring at install time can be a smaller bill than retrofitting later, and the add-on guide linked below covers what a retrofit involves.

Does this home need a battery? Three questions in order A decision tree. Question 1: would an outage endanger a person or a must-run load? Yes leads to a backup job sized to critical loads. No leads to question 2: does an export earn far less than an import costs? Yes leads to a shift job valued at retail minus export divided by 0.85. No leads to question 3: is the evening price above the midday value divided by 0.85? Yes leads to a time-of-use job. No leads to not yet: pre-wire and keep the option open. 1. Would an outage endanger aperson or a must-run load?Backup: size to criticalloads, price as insurance2. Does an export earn far lessthan an import costs?Shift: retail minusexport divided by 0.853. Is the evening price abovethe midday value / 0.85?Time-of-use: count thespread after the 15% lossNot yet: pre-wire at installand keep the option openYesYesYesNoNoNo
Does this home need a battery? Three questions in order Answer in order; the first yes names the battery's job. The 0.85 divisor is the representative round-trip efficiency, so a time-of-use spread must exceed 17.6 percent of the midday value before shifting earns anything. Source: SolarRoost decision rule; round-trip efficiency from National Laboratory of the Rockies (formerly NREL), 2024 Annual Technology Baseline, checked October 29, 2026.
Show the decision tree as a table
QuestionIf yesIf no
1. Would an outage endanger a person or a must-run load?Backup job: size to critical loads, price as insuranceGo to question 2
2. Does an export earn far less than an import costs?Shift job: value is retail minus export divided by 0.85Go to question 3
3. Is the evening price above the midday value divided by 0.85?Time-of-use job: count the spread after the 15 percent lossNot yet: pre-wire at install and keep the option open

A battery can hold more than one job, and the jobs stack: a home on net billing with a well pump has both a shift job and a backup job, and the shift savings partly pay for the insurance. But stacking only works if the battery is sized and operated for both. A battery drained every evening for bill savings may be half empty when a storm arrives at night, so ask how much capacity the installer reserves for backup, and whether you can change that reserve yourself. If your existing system came first, the guide to adding a battery later covers the coupling choice.

Price the battery on its own line

Do not ask storage to rescue weak solar economics, and do not count on a utility battery program payment to do it, since the household rates that guide reads work out to roughly $330 to $825 a year for a battery reliably sharing 3 kW, against a $13,120 storage line in the 2025 benchmark, and if you already have an older system, check whether a battery affects its legacy net metering first. A battery raises project cost; it does not fix an oversized system or an expensive quote. Get the solar price per watt defensible first, in the estimator and against competing quotes, then decide whether storage earns its separate place. If storage is coming later, pre-wiring and a compatible inverter during the solar install can cut the retrofit bill, which is a cheaper way to keep the option open. Whatever the tariff says, check which battery incentives are open before pricing the storage line; the battery incentives guide reads each program's own status page.

What the first storage benchmark added to a home system

NREL's first residential storage benchmark (Q1 2016, published 2017) found adding a small 3 kW / 6 kWh battery to a 5.6 kW home system raised the modeled project price from $15,581 to $27,703 for a DC-coupled system or $29,568 for an AC-coupled one, about 1.8 and 1.9 times the solar-only price. The current benchmark tells the same story at today's representative size: in the Q1 2025 PV cost benchmarks, an 8 kW system is modeled at $23,600 for solar alone and $36,720 with a 13.5 kWh battery (modeled market price, 2024 dollars). That gap buys outage protection and export strategy, not extra sunshine. A household with two outages a decade is buying comfort; a household on net billing with cheap midday exports may be buying real bill savings. Same battery, different verdicts.

More buyers are adding batteries, and why that is not your reason

Storage is becoming a normal line on residential quotes. Lawrence Berkeley National Laboratory's 2026 data update found that the share of new residential solar systems installed with a battery rose from 25 percent in 2024 to 37 percent in 2025, driven largely by California but with large increases in Arizona and Texas, and with Hawaii and California keeping the highest rates. The lab also reported a median residential battery of 13.5 kilowatt-hours in 2025, similar to the year before, and noted that these sizing trends rest mainly on California data. Among cash purchases, median prices for paired solar and storage were $2.1 per watt of solar higher than for solar alone.

Read those figures as a description of tariffs, not as a recommendation. California's Net Billing Tariff, adopted by the state commission in Decision 22-12-056, pays exports at hourly values from its Avoided Cost Calculator instead of the retail rate, which is exactly the export case on this page, and the decision itself says those price signals promote installing storage with solar. A rising national attachment rate tells you that more buyers now face the export case or want backup. It does not tell you that your utility pays exports below retail, or that your evening spread clears the round-trip loss. Note also what the lab says its attachment rates leave out: batteries added to existing solar systems and stand-alone batteries are not counted, so the share of solar homes that eventually add storage is not visible in the 37 percent figure.

The popularity of one battery size matters for a different reason. A 13.5 kilowatt-hour median means many quotes arrive sized to one product rather than to your loads. That can be right, but it should be a conclusion from your critical-loads list and your evening use, not a starting assumption. The California net billing guide works the hour-by-hour case where the attachment growth comes from.

Power, energy, and losses: read the spec sheet in three numbers

A battery quote may lead with one big number, and that number is the least useful one on the page. A home battery is described properly by two ratings together: power in kilowatts (kW), which is how fast it can deliver energy at one moment, and energy capacity in kilowatt-hours (kWh), which is how much it can hold in total. The U.S. Department of Energy makes the same point in its Solar-Plus-Storage 101 explainer: a storage system is referred to by both its capacity and its storage time, and naming it by energy size alone is, in the explainer's words, less ideal. Divide capacity by power and you get duration at full power. The representative residential system in the National Renewable Energy Laboratory (NREL) 2024 Annual Technology Baseline is 5 kW and 12.5 kWh, so the arithmetic is 12.5 divided by 5, which equals 2.5 hours at full power. Run that same battery at a steady 1 kW draw from a critical loads panel and the nameplate arithmetic stretches to 12.5 hours, before the adjustments below. Same battery, very different outage story, and the difference is your load, not the brochure.

The second number is round-trip efficiency, which NREL defines as the ratio of useful energy output to useful energy input. For residential battery storage, the 2024 Annual Technology Baseline adopts 85 percent as a representative round-trip efficiency, based on Cole and Karmakar (2023). Put plainly, that is planning arithmetic, not a promise for your unit: for every 10.0 kWh you put into a battery at that efficiency, you get 8.5 kWh back out, and 1.5 kWh is lost in the round trip. Apply that before you value stored energy for evening use or for a time-of-use shift. If you price all 10.0 kWh at your evening retail rate, you have overstated the value of the shift by the lost share, and a small rate spread can disappear entirely once the loss is counted. Ask the installer which efficiency figure, on which basis, applies to the exact unit quoted, and run your own math at that figure and at 85 percent to see whether the verdict changes.

The third number is what the battery holds over time, not on day one. NREL notes that degradation is a function of the usage rate in its modelling, which assumes roughly one cycle per day for these systems, and that its fixed operating cost figures include battery augmentation so a system can operate at its rated capacity through its 15-year analysis lifetime. In plain terms, a battery cycled hard every day for time-of-use shifting is being used differently from one held mostly for outages, and its capacity years from now will not match the nameplate. That is why the warranty terms matter as much as the capacity figure: read what capacity the warranty actually covers, over how many years or cycles, and what happens when the battery falls below that line. Then size from your critical loads list at the degraded, efficiency-adjusted figure, not at the fresh nameplate. If the loads that truly matter still fit inside that smaller, honest number, the battery is sized for the job you named. If they only fit at nameplate, with no losses counted, you are shopping on a best case.

Once the three numbers are on the page, the price can be read against them. The home battery cost guide breaks a modelled 12.5 kWh system into its ten line items and converts an installed price into a cost per kWh delivered, which is the figure to set against the tariff cases above. If the battery will sit in an unheated garage, check the maker's operating window against your winter lows as well; the cold-weather guide tables four makers' documented limits.

Name the battery's job first

  • Write down what the battery is for: outages, export value, demand charges, or time-of-use shifting. If none apply, the answer may be not yet.
  • Count critical loads carefully. A battery that covers lights, fridge, Wi-Fi, and a medical device is smaller and cheaper than whole-home backup.
  • Ask for the battery quote as a separate line, never bundled invisibly into the solar price.
  • Confirm how your net metering or export structure treats stored energy before valuing it.
  • Run the three-line test on your own tariff: evening price minus (midday value divided by 0.85), times the kilowatt-hours you would really shift each evening.

Battery questions

Can I add a battery later instead?

Usually yes, at a higher per-unit cost than installing together. Your existing inverter decides whether it is a simple AC-coupled add-on, and the guide to adding a battery to existing solar covers coupling, permits and tariff risk. If you know storage is coming, pre-wiring and a hybrid inverter during the solar install can cut the later bill.

Will a battery run my whole house in an outage?

A battery the size of the 12.5 kWh representative system in the 2024 Annual Technology Baseline runs a steady 1 kW critical-loads draw for about 12.5 hours before losses, not a whole house for days. Running air conditioning or electric heat through an outage takes a much larger and more expensive system, and the outage backup guide works the runtime arithmetic for both cases. Your load list decides the size, so make it before shopping.

How big a battery do I need for outages?

Size it from a critical loads list, not from the whole house. Lights, refrigeration, Wi-Fi, and a medical device need a much smaller and cheaper battery than running air conditioning or electric heat through an outage. A 12.5 kWh battery at a steady 1 kW draw lasts about 12.5 hours before losses. Write the load list before shopping and ask installers to price against it.

Should the battery be priced inside the solar quote?

Ask for it as a separate line. Bundling storage invisibly into the solar price hides both the price per watt of the panels and the true cost of the battery. Separated pricing lets you judge the solar against the benchmark first, then judge storage against your export rate, your time-of-use spread, and your outage history on its own merits.

What does 85 percent round-trip efficiency mean for my math?

It means you do not get back everything you store. Round-trip efficiency is useful energy out divided by energy in, and NREL adopts 85 percent as a representative figure for residential battery storage in its 2024 Annual Technology Baseline, checked October 29, 2026. As labelled planning arithmetic: store 10.0 kWh and you get 8.5 kWh back, with 1.5 kWh lost. Value only the 8.5 kWh you can actually use in the evening, not the full 10.0 kWh you put in, and ask which efficiency basis applies to the unit you are quoted.

Is a battery worth it with retail net metering?

Not for bill savings at a one-for-one retail credit on a flat rate. When exports are credited one-for-one at the retail rate, the grid already returns your surplus without loss, while a battery returns about 85 percent of what it stores. At a labelled 30-cent rate, each kilowatt-hour cycled through the battery is worth about 5.3 cents less than exporting it. Backup value is unaffected, and credits that expire unused change the math, so check your tariff for both. On a time-of-use rate, run the spread test instead: the evening price has to beat the midday credit divided by 0.85.

How big does a time-of-use spread need to be for a battery to pay?

The evening price has to exceed the midday value divided by the round-trip efficiency. At the 85 percent representative figure in the 2024 Annual Technology Baseline, that is 17.6 percent above the midday value: with a 25-cent midday credit, the evening price must beat about 29.4 cents before shifting earns anything. The size of the spread above that floor, times the energy you actually shift, is the saving.

Sources and verification

Battery prices, incentives, and export rules move quickly. The cost figures cited are modeled benchmarks, not market prices: the storage cases are from NREL's Q1 2016 residential solar-plus-storage benchmark (published 2017) and the current-system figures are from the Q1 2025 PV system cost benchmarks published by the U.S. Department of Energy and its national laboratory partners, both re-checked October 29, 2026. The rates in the tariff examples are labelled illustrations chosen to show the arithmetic, not any utility's current prices; run the same lines with your own tariff before deciding.