How Many Solar Panels Do I Need?

Every solar quote you will ever receive is an answer to a question you probably have not asked yet. The question is not "how many panels fit on my roof." It is "how many kilowatt-hours do I use, and how many of those do I want to make myself." Once you have those two numbers, panel count is division. Everything else in a solar proposal is packaging around that arithmetic, and the reason quotes for the same house vary so wildly is that different companies quietly assume different answers to the second half of the question.

Here is the arithmetic, worked all the way through, plus the part almost nobody tells you until the first outage: a normal rooftop solar array does not keep your lights on when the grid goes down.

Quick Answer

Take your annual electricity use in kilowatt-hours from your utility bill, then divide it by your location's daily peak sun hours multiplied by 365 and by about 0.86 for real-world system losses. That gives the system size in kilowatts. Divide by your panel wattage to get panel count. For the US average household at 10,791 kWh per year, at a middling 4.5 peak sun hours, that is a 7.6 kW system, which is about 20 panels at 400 watts each or 17 at 450 watts. Real answers across the country land roughly between 13 and 25 panels for an average-consumption home, driven almost entirely by where you live. If your goal is backup power rather than offsetting a bill, the number is far smaller, often three or four panels, but you need a battery and the right inverter or the array is useless during an outage.

Key Takeaways

  • Start with your own kWh, not a national average. EIA puts the average US residential customer at 10,791 kWh a year, about 899 kWh a month, but the spread between states is more than double: Louisiana averages 14,774 kWh and Hawaii 6,178.
  • Sunlight, not roof size, sets the panel count. The same 10,791 kWh house needs roughly 25 panels in a cloudy northern climate and 15 in the desert Southwest. Nothing about the house changed.
  • Losses are real and already quantified. NREL's PVWatts calculator applies a default total system loss of 14 percent, covering soiling, shading, wiring, connections, mismatch, light-induced degradation, nameplate tolerance, and availability. Skipping that derate is how people end up short.
  • Panel wattage moved. Panels quoted for homes today are mostly 400 to 460 watts, up from the 250 to 350 watt modules of a few years ago, so older panel-count rules of thumb overstate how many you need.
  • Grid-tied solar shuts off in an outage. UL 1741 and IEEE 1547 require a grid-tied inverter to disconnect within about two seconds of losing the grid, so utility crews are not working on lines your roof is energizing. Without a battery and an inverter designed to island, your panels go dark exactly when you wanted them.
  • Backup sizing is a different calculation entirely. Offsetting a bill is sized on annual kWh. Keeping a refrigerator and lights alive is sized on daily kWh plus storage, and it is a much smaller, much cheaper system.

The Formula

There is one equation, and it has four inputs:

System size in kW = annual kWh ÷ (peak sun hours × 365 × 0.86)

Then panel count = system watts ÷ panel wattage, rounded up.

"Peak sun hours" is the useful fiction that makes solar math tractable. It is not how long the sun is up. It is the number of hours per day your location would need to receive full-strength noon sunlight to deliver the same total energy the real, slanting, partly cloudy day actually delivers. Across the continental United States that number runs roughly 3.5 to 6 depending on latitude and cloud cover, which is why the same house needs very different arrays in Seattle and Phoenix.

The 0.86 is the derate. NREL's PVWatts tool, which is the standard free reference for this, defaults to 14 percent total system losses. Its published component defaults are soiling at 2 percent, shading at 1 percent for an array with little or no shading, wiring at 2 percent, connections at 0.5 percent, light-induced degradation at 1.5 percent, nameplate rating at 1 percent, mismatch at 2 percent, and availability at 3 percent. Those are combined multiplicatively rather than added, which is why the total lands at 14 rather than 13. If your roof has real shading from a chimney or a neighbor's oak, your number is worse than the default, and that is exactly the kind of thing a site-specific PVWatts run catches and a napkin does not.

Step One: Find Your Actual Consumption

Pull twelve months of utility bills and add up the kWh. Twelve, not one, because a single bill from April tells you nothing about August. If your utility has an online portal it will usually chart this for you directly.

If you are sizing for a house you do not live in yet, or you want to sanity-check the bill against reality, do a load analysis instead. The method is the one Cooperative Extension teaches from the federal appliance guidance: the wattage is stamped on the bottom, back, or nameplate of most appliances, and where it is not, you find the current draw in amperes and multiply by the voltage. Most US household appliances run on 120 volts, and larger ones like dryers and electric cooktops run on 240. Then watts × hours used per day ÷ 1,000 = daily kWh.

For any major appliance there is a shortcut worth knowing: the yellow EnergyGuide label carries an "Estimated Yearly Electricity Use" in kWh, calculated under DOE test procedures. That is a far better number than any chart. A typical refrigerator label reads around 630 kWh per year, which works out to roughly 1.7 kWh a day.

One warning about future load. If an electric vehicle or a heat pump is anywhere in your plans, size for the house you will have, not the one you have now. A car is the single largest load most households will ever add, and it is far cheaper to install four extra panels during the original job than to come back and re-permit a second array.

Step Two: Run the Numbers for Your Latitude

This table uses the US average of 10,791 kWh per year, the 0.86 derate, and rounds panel counts up to a whole panel. Find the row that matches your region's peak sun hours, or better, run your own address through PVWatts, which uses real historical irradiance data for your specific coordinates rather than a regional guess.

Peak sun hours Roughly where System size Panels at 400 W Panels at 450 W
3.5 Pacific Northwest, upper New England, Great Lakes 9.8 kW 25 22
4.0 Northeast, Upper Midwest 8.6 kW 22 20
4.5 Mid-Atlantic, Ohio Valley, much of the middle of the country 7.6 kW 20 17
5.0 Southeast, southern Plains 6.9 kW 18 16
5.5 Texas, inland California, Front Range 6.3 kW 16 14
6.0 Arizona, Nevada, New Mexico 5.7 kW 15 13

Read the extremes of that table honestly, because they are the real lesson. Two identical families with identical bills need a 9.8 kW array in one place and a 5.7 kW array in another. That is a difference of ten panels and many thousands of dollars, and it has nothing to do with the house, the appliances, or the panels. Anyone quoting you a panel count before they know your zip code is guessing.

How Many Solar Panels to Power a House, Really

"Power a house" is doing a lot of work in that phrase, and it usually means one of three quite different projects.

Offset the annual bill

This is the standard grid-tied residential install and the one the table above sizes. You produce roughly as many kWh over a year as you consume, exporting surplus at midday and pulling from the grid at night and in winter. The grid is functioning as your battery, which is why this is the cheapest version by a wide margin, and also why it does nothing for you in an outage.

Cover the essentials during an outage

Completely different arithmetic, and a much smaller system. You are not sizing to annual consumption. You are sizing to the daily kWh of a short list of loads, plus storage to carry them through the night.

Load Assumed daily kWh Notes
Refrigerator 1.7 From a typical EnergyGuide label at 630 kWh/yr. Read yours.
Chest freezer 1.0 Skip if you do not have one. Full freezers hold cold far longer.
LED lighting 0.3 Modern bulbs draw about 10 watts each. Lighting is nearly free now.
Phones, laptop, router, radio 0.5 Communications tier. Small and high value.
Fans or a well pump 0.5 Highly variable. A well pump can dwarf this line.
Essentials total 4.0 Your list will differ. Build it from your own labels.

At 4.0 kWh a day and 4.5 peak sun hours, the array works out to 4.0 ÷ (4.5 × 0.86), which is about 1.03 kW, or three panels. Three. That is the whole array for a household essentials tier, and it is the number that surprises people who have been quoted twenty.

The catch is that the panels are now the cheap part. You need a battery sized to carry those loads overnight and through the grey days that follow most storms, which for a 4 kWh daily draw means something in the 5 to 10 kWh class if you want a comfortable margin rather than a nervous one. Storms are the worst case for solar in exactly the moment you need it: the weather that knocked the grid down is usually still overhead the next morning. Size the battery for the outage, and the array only for recovering the battery.

Go fully off-grid

Now you are sizing for the worst month of the year rather than the annual average, because there is no grid to lean on in December. In practice that means running the calculation with your December peak sun hours instead of your annual figure, and it commonly lands two to three times larger than the grid-tied number, plus several days of battery autonomy. Off-grid is a design problem, not a shopping problem. Get a real load analysis done.

The Part That Catches Everyone: Solar Does Not Work in a Blackout

This is the single most important thing on this page for anyone buying solar with resilience in mind, and it is routinely glossed over in sales conversations.

A standard grid-tied inverter is required to shut down when the grid goes down. UL 1741 and IEEE 1547, the interconnection standards every residential inverter is certified against, require the inverter to stop feeding the line within roughly two seconds of detecting that it has been islanded from the utility. The reason is the same reason you never backfeed a generator into a wall outlet: a line crew has to be able to treat a downed conductor as dead, and a house quietly energizing the neighborhood transformer can kill someone.

⚠️ SAFETY: Do not attempt to defeat or bypass your inverter's anti-islanding protection to get power during an outage. It exists to keep utility line workers and neighbors from being electrocuted by your system, and disabling it is both dangerous and a code violation. If you want power during an outage, the correct answer is hardware designed for it: a hybrid or battery-backup inverter with an automatic transfer switch, installed by a licensed electrician. Verify against IEEE 1547, UL 1741, and your local electrical code.

What you need instead is a system built to island deliberately: a hybrid inverter or a battery system with a backup gateway, which physically disconnects your house from the grid and then runs your protected circuits as a small island. That is a design decision made at purchase. Retrofitting it later is possible but usually means replacing the inverter, so if backup power is part of why you are buying solar, say so out loud on the first call and make sure the proposal names the backup hardware.

The tradeoffs between this and an engine generator are covered in Generator vs. Solar: What Actually Keeps the Lights On, and if you are also weighing a fuel-burning backup, What Size Generator Do I Need? runs the equivalent sizing math on the other side of that decision.

Roof Reality: Orientation, Tilt, and Shade

The table gives you a panel count. Your roof decides whether those panels can go where they need to.

  • Orientation. In the northern hemisphere, south-facing is best. East and west faces still produce usefully, generally at a meaningful discount, which means more panels for the same output. North-facing roof planes are mostly not worth it.
  • Tilt. Most residential roofs are close enough to optimal that this is a second-order concern. Ground mounts and flat roofs give you the freedom to actually optimize it.
  • Shade. The one that ruins arrays. A single chimney shadow crossing a string can cost far more production than its area suggests, depending on the array's electrical topology. Microinverters and optimizers reduce but do not eliminate the penalty. This is what a real site assessment is for.
  • Area. A 400 watt panel occupies roughly 20 to 30 square feet. Twenty of them is on the order of 400 to 600 square feet of unobstructed, correctly oriented roof, which is more than many houses actually have on the good side.

If the roof cannot hold the array the math asks for, you have three honest options: reduce consumption first, accept partial offset, or go to a ground mount. Reducing consumption is almost always the cheapest kilowatt-hour available. Every 1,000 kWh a year you stop using is roughly two fewer panels you have to buy, mount, wire, and eventually replace.

FAQ

How many solar panels do I need to power my house?

For a US household at the national average of 10,791 kWh a year, somewhere between 13 and 25 panels at current panel wattages, depending almost entirely on how much sun your location gets. Run your own annual kWh through the formula above rather than trusting any single number, because household consumption varies by more than a factor of two across states, and sunlight varies by nearly as much.

How many solar panels do I need for a 2,000 square foot house?

Square footage is the wrong input and this is the most common version of the question. A 2,000 square foot house with gas heat, gas water heating, and two occupants might use 6,000 kWh a year. The same house with a heat pump, an electric water heater, a hot tub, and an EV might use 25,000. Those are a nine-panel system and a thirty-five-panel system in the same building. Use kilowatt-hours, not floor area.

Will solar panels power my house during a blackout?

Not on their own. A standard grid-tied system is required by UL 1741 and IEEE 1547 to shut down within about two seconds of losing the grid, for the safety of utility line crews. You need a battery plus a hybrid or backup-capable inverter to keep any power during an outage, and that has to be specified when the system is designed.

How many solar panels do I need to run a refrigerator?

Using a typical EnergyGuide figure of 630 kWh a year, about 1.7 kWh a day, and 4.5 peak sun hours, roughly 0.44 kW of panels, so two 400 watt panels with a little margin. But a fridge is a motor load that cycles, so the limiting factor is usually the inverter's surge capacity and the battery, not the panels. Solve the surge and storage first and the array is almost an afterthought.

Do I need a battery?

Only if you want power during an outage, or your utility's export compensation is poor enough that storing your own midday surplus beats selling it. For pure bill offset under favorable net metering, a battery is usually not the cheapest kilowatt-hour. For resilience, it is not optional, because without it the array shuts down with the grid.

What about winter and cloudy days?

Production drops substantially and seasonally, which the annual-average calculation already absorbs for a grid-tied system: you overproduce in July and underproduce in January, and the grid smooths it. For off-grid or backup sizing there is no smoothing, so you size on the worst month, not the average one. Cold itself is not the enemy. Panels are slightly more efficient in cold air. Short days and cloud cover are what cut the output, along with snow sitting on the glass.

Is it better to buy more panels than I need?

Modest headroom is sensible, because consumption tends to rise (EVs, heat pumps, a home office) and panels degrade slowly over decades. Large overbuild is usually wasted money under net metering rules that pay less for exported power than they charge for imported power. Check what your specific utility pays for exports before you decide, because that single number changes the economics of overbuilding more than anything else in the proposal.

The Bottom Line

Panel count is arithmetic wearing a costume. Get twelve months of kWh off your bills, get your real peak sun hours from PVWatts rather than a regional guess, derate by about 14 percent, divide, and you have an answer you can check any installer's proposal against. Then decide, before you sign anything, which project you are actually doing: offsetting a bill, or keeping the refrigerator running when the grid drops. They are different systems, and the second one is smaller, cheaper, and completely dependent on hardware most standard quotes do not include. Ask the question out loud. Be prepared. Stay savage.

References