Guides

How many solar panels do you need? Size the system from your bill, then count

At average US electricity use, a home needs roughly 16 to 25 panels of 400 watts in the eight US cities modelled here, depending on climate. The method: take a year of electricity use in kilowatt-hours, divide it by what one kilowatt of panels produces in a year where you live, and divide that system size by the panel wattage. At the 2022 US average use of 10,791 kWh, that works out to 16 panels of 400 watts in Phoenix and 25 in Seattle.

System size is where solar gets sold by the square foot instead of by the bill. The right size comes from three things: how much electricity you actually use, what your utility pays for the surplus, and how much good roof you really have. Get those three right and the kilowatt number mostly chooses itself, and the panel count is one more division. This guide prints the production factors behind that division for eight US cities, modelled in the national laboratory's free PVWatts calculator with every input stated, so you can check an installer's count instead of taking it on trust. Nothing here is engineering or financial advice: the numbers are planning examples, and a roof-specific model and your own bills outrank them.

Start with the bill, not the roof

Annual kilowatt-hours from 12 months of bills is the denominator. Seasonal patterns matter too: a home that peaks with summer cooling matches solar production; a home that peaks with winter heating exports its summer surplus. The same annual usage can justify two different sizes.

If you want a reference point before your bills are in hand, the U.S. Energy Information Administration publishes one. In 2022 the average US residential customer bought 10,791 kWh of electricity, about 899 kWh a month. The spread between states is wide: Louisiana customers averaged 14,774 kWh and Hawaii customers 6,178 kWh, so the national figure is a teaching number, not a stand-in for your house. EIA adds a caution that matters for anyone sizing a second system or buying a home that already has one: these are purchases from the utility, and a home with net-metered panels buys less than it uses. Size from consumption, which means adding back what any existing panels produce, not from the utility's billed total alone.

What one kilowatt of panels produces: eight cities modelled in PVWatts

The divisor in every sizing calculation is the production factor: how many kilowatt-hours one kilowatt of panels delivers in a year at your location. It is the number most worth seeing printed on a proposal, because it moves the panel count more than any other input, because the same household needs very different systems in Seattle and Phoenix. Rather than quote a national rule of thumb, this guide ran the same 1 kW system through PVWatts version 8, the national laboratory's free production model, at eight city locations on October 28, 2026. The inputs were held identical everywhere so the only thing that changes is the weather: a standard module, a fixed roof mount, a 20 degree tilt facing due south, the PVWatts default of 14 percent total system losses, a DC to AC ratio of 1.2 and 96 percent inverter efficiency. PVWatts reported the weather source for every run as the NSRDB typical-meteorological-year file, so each figure describes a typical year, not any particular one.

The 20 degree tilt is a deliberate choice, not a guess at your roof. It is the roof pitch Lawrence Berkeley National Laboratory assumes for residential roofs when it estimates roof coverage in its 2026 data update. Your own roof's pitch and direction will move your factor, which is why the caption calls these planning examples; the point of the comparison is the spread between climates at identical inputs.

Modelled first-year output of 1 kW of panels in eight US cities Horizontal bars of PVWatts version 8 annual AC output per kilowatt DC, fixed roof mount at 20 degrees facing south with 14 percent losses: Seattle, WA 1,088, Boston, MA 1,302, Chicago, IL 1,308, Atlanta, GA 1,402, Houston, TX 1,421, Denver, CO 1,595, Los Angeles, CA 1,677, Phoenix, AZ 1,755 kWh per kW per year. 05001,0001,500Seattle, WASeattle, WA: 1,088 kWh per kW per year1,088Boston, MABoston, MA: 1,302 kWh per kW per year1,302Chicago, ILChicago, IL: 1,308 kWh per kW per year1,308Atlanta, GAAtlanta, GA: 1,402 kWh per kW per year1,402Houston, TXHouston, TX: 1,421 kWh per kW per year1,421Denver, CODenver, CO: 1,595 kWh per kW per year1,595Los Angeles, CALos Angeles, CA: 1,677 kWh per kW per year1,677Phoenix, AZPhoenix, AZ: 1,755 kWh per kW per year1,755First-year AC output, kWh per kW DC per year
Modelled first-year output of 1 kW of panels in eight US cities Same system everywhere: 1 kW DC, standard module, fixed roof mount, 20 degree tilt, 180 degree azimuth, 14 percent losses, DC to AC ratio 1.2, 96 percent inverter efficiency, typical-year weather. Planning examples, not a prediction for your roof. Source: National Laboratory of the Rockies (formerly NREL), PVWatts API version 8.5.0, NSRDB PSM V3 typical-meteorological-year weather; runs made October 28, 2026.
Show the eight PVWatts results as a table
CitykWh per kW per year
Seattle, WA1,088
Boston, MA1,302
Chicago, IL1,308
Atlanta, GA1,402
Houston, TX1,421
Denver, CO1,595
Los Angeles, CA1,677
Phoenix, AZ1,755

Read the figure as a range of planning factors rather than eight answers. Seattle's 1,088 kWh per kW and Phoenix's 1,755 differ by about 61 percent, which means a Seattle household needs about 61 percent more panel capacity than a Phoenix household with the same bill. Six of the eight cities fall between 1,300 and 1,700 kWh per kW. These runs are a best case for orientation, an unshaded roof facing due south, so they sit at or above the top of the state planning ranges this site records on its state pages: Boston at 1,302 against 1,150 to 1,300 for Massachusetts, Atlanta at 1,402 against 1,150 to 1,350 for Georgia, Phoenix at 1,755 against 1,550 to 1,750 for Arizona. A roof that faces away from south or carries shade lands lower, as the east and west runs in the ground versus roof mount guide show, so on an ordinary roof the lower figure is the safer divisor until a roof-specific model exists. When an installer's proposal implies a factor well outside the band for your region, ask which inputs produced it: a very high factor usually means low assumed losses or optimistic shade, and a very low one may mean a poor roof plane that a different layout would avoid. You can reproduce any of these runs yourself in the free PVWatts web calculator by typing your address and the inputs in the caption, then changing the tilt and azimuth to match your roof.

Count the panels: three divisions and a round-up

With annual use and a production factor in hand, the count takes three divisions. None of them needs a calculator more advanced than a phone, and an installer should be willing to show each one on the proposal.

  1. Annual use divided by the production factor gives the system size in kilowatts DC. For the US average of 10,791 kWh at Boston's modelled 1,302 kWh per kW, that is 10,791 divided by 1,302, or about 8.3 kW.
  2. System size divided by the panel wattage gives the panel count. A panel's wattage is printed on its datasheet and in the equipment list of your quote; this guide uses 400 W and 450 W only as illustrations. At 400 W, the unrounded 8.29 kW is 8,290 watts divided by 400, or 20.7 panels.
  3. Round up to a whole panel, and then accept that the roof layout may round again. Twenty-one 400 W panels make 8.4 kW. If the roof plane fits 20, the system is 8.0 kW and covers a little under the full year.
  4. Divide the finished size back into the use figure as a check. 8.4 kW at 1,302 kWh per kW models about 10,900 kWh in year one, against 10,791 kWh used: close to a full annual offset, before the export rate has had its say.
Panels needed for the 2022 US average use of 10,791 kWh a year, by modelled city factor (rounded up to whole panels; panel wattages are illustrations)
CitykWh per kW per yearSystem size for 10,791 kWh400 W panels450 W panels
Seattle, WA1,0889.9 kW2523
Boston, MA1,3028.3 kW2119
Chicago, IL1,3088.2 kW2119
Atlanta, GA1,4027.7 kW2018
Houston, TX1,4217.6 kW1917
Denver, CO1,5956.8 kW1716
Los Angeles, CA1,6776.4 kW1715
Phoenix, AZ1,7556.1 kW1614

The table runs the same division for all eight cities. The same average household needs 25 panels of 400 W in Seattle and 16 in Phoenix, and choosing 450 W panels trims each count by one or two. That is the whole honest answer to "how many panels": the count follows the bill and the sunshine, and the wattage of the panel only changes how many pieces the same kilowatts are cut into. A higher-wattage panel does not make a household need fewer kilowatts. Higher-wattage panels are often physically larger, so what fits more kilowatts onto the same roof is module efficiency, and that matters only when the roof, not the budget, is the limit.

The same division also runs backwards, and that is the fastest audit of a proposal already on your table. Divide the proposal's first-year production by its system size and you get the production factor it assumed. As a labelled illustration, a Boston-area proposal for 8.0 kW that promises 12,400 kWh in year one implies 1,550 kWh per kW, about 19 percent above the 1,302 that PVWatts modelled for a due-south roof in Boston at default losses. That gap is not proof of an error. It is a question with a specific answer: which tilt, azimuth, loss total and weather file produced the higher figure, and would the installer stand behind it in a production guarantee? The production estimate guide walks through those inputs one at a time.

Two households at the ends of EIA's state range show how far the count can move. At Hawaii's average of 6,178 kWh a year, a home on a 1,302 kWh-per-kW roof needs about 4.7 kW, or 12 panels of 400 W. At Louisiana's 14,774 kWh, the same factor gives about 11.3 kW, or 29 panels. Neither is a typical result for those states, because the real factor there differs from Boston's; the point is that use alone can more than double the count.

Exports are worth what the tariff says

Under full retail net metering, an exported kilowatt-hour is worth roughly a consumed one, and sizing near full annual usage makes sense. Under net billing or low export rates, every oversized kilowatt produces cheap exports. In those territories, size toward what the home uses while the sun is up, and treat remaining roof as future capacity, not a sales target.

The roof votes last

Shade, orientation, obstructions, and roof age cap what is sensible. A smaller array on the good plane beats a bigger array spread across poor ones, and any array over a tired roof inherits the roof problem. Size to the roof you have after the roof conversation, not before it.

The roof check is arithmetic too. The Department of Energy and NREL consumer guide Own Your Power (2009) planned on about 100 square feet of unshaded, south-facing roof or yard space for every kilowatt, so the 8.3 kW Boston example asks for something like 830 square feet of good roof before fire setbacks and vents. The same allowance turns the $2.95 per watt benchmark into $29.50 per square foot of roof, one of three per-square-foot readings of a single price that the square-foot conversion guide sorts out. Treat that as an upper bound on a modern quote, not a current rule: Lawrence Berkeley National Laboratory reports that median module efficiencies have risen about 0.4 percentage points a year over the long term, and that residential system sizes and module efficiency have each risen roughly 50 percent since 2010, which it reads as efficiency gains driving larger systems on roofs where space and shade are often the binding limit. Your installer's layout drawing, with each panel's outline on each plane, is the real measurement.

Laboratory data also gives a sense of how much of a roof a typical system covers. In the same 2026 update, residential arrays in 2025 typically covered 16 to 39 percent of total roof area, with a median of 22 percent on single-story homes and 31 percent on multi-story homes, where the same panels sit on a smaller roof. If a proposal claims to cover far more of your roof than that, check the setbacks and the shaded planes it has included; if it covers far less and still falls short of your use, the binding limit is the roof, and that is the case where higher-wattage panels or a ground mount earn a look.

Check the size against what other households install

After the arithmetic, a reality check. Lawrence Berkeley National Laboratory's U.S. Distributed Solar and Storage 2026 Data Update, published in August 2026, reports that the median US residential system installed in 2025 was 7.7 kW, up 0.5 kW from 2024, with most systems between roughly 5 and 12 kW, the 20th to 80th percentile. It also notes that California, roughly half of the residential sample, had a relatively small median size. The 8.3 kW Boston example sits just above the national median, and every result in the city table above lands inside the 5 to 12 kW band.

Use the band as a question, not a target; a new California house is the exception where the starting size comes from a building-code formula rather than a bill, and the Title 24 new-home solar formula is worked there for all 16 climate zones. A proposal for 4 kW on a home with a 12,000 kWh bill is either deliberately sized to daytime use, which is a legitimate choice on a low export rate, or limited by the roof, and the proposal should say which. A proposal for 14 kW on a home with an average bill should name the load it expects to add or the export rate that makes the surplus worth producing. Both can be right. Neither should be a surprise you discover by dividing the numbers yourself after signing.

A worked sizing check you can repeat on your bills

Take twelve months of bills and add the annual kilowatt-hours. The example on this page uses a home consuming 10,800 kWh a year in a climate producing about 1,400 kWh per kW each year. Dividing usage by production gives roughly 7.7 kW for a full annual offset on paper. That arithmetic is the starting point every installer should show you, with your production factor named and its tool identified, before any roof measurement enters the conversation.

Then ask what the last kilowatts are for, using the export rules above. In the example, a 6 kW system sized toward daytime load can pay back sooner than the 7.7 kW full-offset design while leaving roof room for a future addition. Neither size is virtuous on its own; the tariff decides which one your money should buy.

Finish with the roof vote from the section above, adding fire setbacks to shade, orientation and roof age. Price one smaller system alongside every recommended system. The comparison is information any installer expects to provide, and it tells you whether the recommended size came from your bill or from your roof area.

  • Size from twelve months of bills, not one summer month.
  • Ask which kilowatts serve daytime load and which only export.
  • Price a smaller system next to every recommended system.

Month by month: why a right-sized system still runs short in winter

An annual offset hides a seasonal swing that decides how your bills feel. In the Boston PVWatts run, one kilowatt produced about 143 kWh in July and 59 kWh in December, a ratio of about 2.4 to 1. Scale that to the 8.3 kW example and the system models about 1,185 kWh in July and 489 kWh in December, against the household's average month of 899 kWh. Over the year the two sides roughly balance, at about 10,800 kWh produced against 10,791 used, but the household is a net exporter from spring through early autumn and a net buyer through the dark months.

Whether that pattern costs anything depends on how your utility settles the account. Under annual netting with retail credit, summer surplus carries into winter and the swing mostly disappears from the bills; the true-up guide shows how the year-end settlement works and how much to set aside. Under monthly netting or a low export rate, summer surplus is paid at the export value and winter shortfall is bought at retail, and a system sized to the annual total can still leave noticeable winter bills. The net metering guide explains which structure your utility uses. Ask the installer for production by month, set it beside twelve months of use, and look at the gap in each month, not only at the annual percentage.

Seasonal load matters as much as seasonal sunshine. A home whose use peaks in summer with air conditioning matches the production curve; a home that peaks in winter with electric heat exports in exactly the months it needs least. Two homes with identical annual use can therefore justify different sizes, and the monthly table, not the annual one, shows which kind of home you have.

Read the production estimate behind the size: losses, clipping, and year twenty

The kilowatt figure on a quote is a nameplate rating, not a promise of output. Module ratings are measured under standard test conditions, and the U.S. Department of Energy notes that real sunlight is usually weaker and real cells usually hotter than those test conditions, a gap that the heat guide converts into percentages for six cities and kilowatts for one. Two quotes can therefore carry the same kilowatt number and still promise very different production, because the difference sits in the assumptions behind the estimate: the weather file, your tilt and azimuth, the loss total, the inverter pairing, and whether output is allowed to fall with age. Ask for those inputs in print before you argue about the size itself.

NREL PVWatts, the modelling tool this site points to, makes the loss stack visible. In the PVWatts Version 5 Manual, total system losses carry a default of 14 percent, built from separate items such as soiling, shading, mismatch, wiring, connections, light-induced degradation, nameplate rating, and system availability. The total is combined multiplicatively, not by adding the items, which is why the printed total is smaller than a simple sum would suggest. The practical test is comparison. In a labelled hypothetical, hold everything else equal and change only the loss total from 14 percent to 7 percent: the estimate rises by about 8 percent, from 10,800 kWh to about 11,679 kWh, a gap of about 879 kWh a year with no extra panel on the roof. A quote that assumes unusually low losses needs a site-specific reason, such as measured shade turned into a loss figure, not optimism.

Next, read the pairing between panels and inverter. PVWatts treats the DC to AC size ratio and inverter efficiency as separate inputs, and the PVWatts version 8 runs on this page use a ratio of 1.2 with 96 percent inverter efficiency. On the worked example here, a 7.7 kW DC array at a 1.2 ratio pairs with an inverter of about 6.4 kW AC. That smaller AC number is not automatically a mistake. Around the sunniest midday hours the array can briefly produce more than the inverter can pass, and that excess is clipped, while the larger array produces more in the morning, afternoon, and cloudy weather. Clipping is a design trade-off to price, not a defect to fear or a bonus to ignore.

Finally, ask which year the offset claim describes. The Department of Energy puts module degradation at about 0.5 percent per year and system lifespans at 20 to 30 years. In a labelled illustration using this page's example, a system producing 10,800 kWh in year one at that rate produces about 10,324 kWh in year ten and about 9,819 kWh in year twenty, roughly 91 percent of its first year, if household usage stays flat. Sizing exactly to year one therefore plans a growing shortfall, while sizing up today buys that future shortfall back at today's price, and the top slice may be paid at a low export rate in its early years. Neither answer is automatic. What matters is that the proposal states its degradation assumption and the year its offset percentage refers to, so you choose the trade-off instead of discovering it.

Before you sign, get the estimate inputs on one page: location and weather source, tilt, azimuth, module type, the loss total and its main items, the DC to AC ratio, inverter efficiency, the degradation rate, and production by month. Two proposals compared on different inputs are not yet a price comparison. Re-running PVWatts with your own roof angles and the same inputs for both sizes is the fastest way to see whether a bigger system is buying production you will use, or production the tariff will discount.

Sizing for load you will add: an EV or a heat pump

Planned electrification is a legitimate reason to size above today's bill, and the method is the same division with a new use figure. As a labelled illustration, suppose an electric vehicle will add 3,000 kWh a year; your own figure is your annual miles multiplied by the vehicle's rated kilowatt-hours per 100 miles, divided by 100, plus a margin for charging losses. At Boston's 1,302 kWh per kW, 3,000 kWh needs about 2.3 kW more, or six more 400 W panels. The same arithmetic applies to a heat pump, with one difference: its load lands in winter, when the monthly table above shows production at its lowest, so under monthly netting a heat pump is the load solar matches least well.

If the new load is years away or uncertain, buying the panels now is not the only option. Designing for expansion, with conduit, roof layout, and inverter and panel capacity left ready, keeps the later addition cheap, and the expansion guide covers the inverter, busbar and utility checks an addition faces. The caution runs the other way where an older system enjoys a grandfathered export rate, because some tariffs treat a later addition as a reason to move the account to current rules. Price both paths, size-up-now and add-later, on the same day, and ask which export rules each would fall under.

Sizing a 10,800 kWh household

A home using 10,800 kWh a year in a 1,400 kWh-per-kW climate needs about 7.7 kW to offset its full annual usage on paper. Under retail net metering, that full offset is roughly rational. Under a low export rate, the last 2 kW produces mostly cheap exports, and a 6 kW system sized to daytime load can pay back sooner while leaving roof room for the future.

The 100% offset that pays at export rates

Sizing to 100% of annual usage because the phrase sounds complete, in a territory where the exported half of that production is paid at a fraction of retail.

What to gather before sizing

  • Get 12 months of bills, not one summer month, and size against annual usage.
  • Ask for the production estimate by month, not just per year.
  • Price one smaller system alongside the recommended one. The comparison is information.
  • If exports pay poorly, size toward self-consumption and ask for the math, not maximum roof coverage.

System size questions

How many solar panels does an average house need?

At the 2022 US average use of 10,791 kWh a year, PVWatts runs for this guide put the answer at about 16 panels of 400 W in Phoenix, 20 in Atlanta, 21 in Boston and 25 in Seattle. The method is annual kWh divided by your location's kWh per kW per year, divided by the panel wattage, rounded up. Use your own twelve months of bills and your quote's panel wattage for your count.

Can I start small and add panels later?

Usually, within inverter and panel compatibility limits (the guide to adding panels to an existing system works the inverter and busbar arithmetic), and installers often design for it. It costs less to leave conduit, roof layout, and electrical capacity ready than to retrofit them. Ask for expansion readiness in writing if growth is likely.

What does 100% offset actually mean?

It means the system is modelled to generate as much electricity over the year as the home uses. Whether that is valuable depends completely on what the utility pays for the surplus produced in sunny months. At retail credit it is fine; at wholesale export prices it is expensive generosity.

Why does my installer want to cover the whole roof?

Because system size is easy to sell by area, while the right size comes from annual usage, the export rate, and the good roof you actually have. Under a low export rate, extra panels mostly produce cheap exports. Ask for the sizing arithmetic from your bills, then compare a smaller design priced the same day. The roof should vote last, not first.

What if my usage grows later, with an EV or a heat pump?

Known future load is a legitimate reason to size up or to build expansion-ready, especially where a later addition could cost an older system its grandfathered export rate. Ask for conduit, roof layout, and electrical capacity that leave room to add panels, and get that readiness in writing. Guessing at panels today for load you may never add ties up money now; designing the roof so additions stay cheap keeps both options open on real numbers.

Why is the inverter in my quote rated below the panel total?

Because the two numbers measure different sides of the system. Panels are rated in DC kilowatts and the inverter in AC kilowatts, and designers often pair them at a DC to AC ratio above one; the PVWatts runs on this page use 1.2. On a 7.7 kW DC array that ratio means about 6.4 kW AC. Brief midday output above the inverter limit is clipped, in exchange for stronger production at other hours. Ask for the ratio in print rather than judging the pairing by the two nameplates alone.

Do higher-wattage panels mean I need a smaller system?

No. The kilowatts you need come from your use and your location's production factor. Higher-wattage panels cut the same kilowatts into fewer pieces, so the panel count drops, but the roof area the array needs depends on module efficiency, not wattage. Efficiency only pays when the roof is the limit; on a roof with room to spare, compare the price per watt, not the wattage per panel. The panel choosing guide runs the five-number comparison, including the roof-space test, on your quotes.

Sources and verification

Production factors on this page are PVWatts planning examples at the stated inputs, run October 28, 2026; panel wattages are illustrations. Your roof's estimate should come from a site-specific model with your tilt, azimuth, and shade, and your usage from a full year of bills. This guide is general information, not engineering, financial or tax advice.