What kWh per kW means, and what a typical city makes
By Emily CarterPublished Data as of 2026-10-27: PVWatts® v8, typical-year weather
kWh per kW is the electricity, in kilowatt-hours, that one kilowatt of solar panels makes in a year. Because it is measured per kilowatt, it compares places regardless of system size. Across the 100 cities on this site the median is 1,450 kWh per kW a year, and the middle half of the cities sit between 1,335 and 1,613.
Two units in one number
The kW in the name is the panels' rated power: the direct-current output printed on each module, added up for the whole array. Ten 400-watt panels make a 4 kW array. The kWh is energy delivered over time, counted after the inverter has turned the panels' direct current into the alternating current a home uses. So the figure already includes the losses between the roof and the wall socket.
Those losses follow the settings we use for every city: 14% system losses (soiling, wiring, mismatch between panels and similar), an inverter efficiency of 96% and a DC-to-AC ratio of 1.2. The panels face south at a 20° tilt. That reference roof sits behind every headline number on the site, so two cities are always compared on identical hardware.
To get a whole system, multiply. Denver's figure is 1,602 kWh per kW, so a 6 kW array there makes about 1,602 × 6 = 9,612 kWh a year, which the city page rounds to 9,610. A 4 kW array would make about 6,410 kWh. The yield estimator does this sum for any size from 0.5 to 100 kW.
What a typical city makes
Most of the 100 cities land in a fairly narrow band: 58 of them make between 1,300 and 1,600 kWh per kW a year, 27 make 1,600 or more, and 15 make less than 1,300. The median is 1,450 and the mean 1,470.
The ends of the range are far apart. El Paso, TX makes 1,835 kWh per kW and Anchorage, AK 951 kWh, a ratio of 1.93×. Only four cities fall below 1,200: Anchorage, AK, Seattle, WA, Portland, OR and Pittsburgh, PA. A figure around 1,450 is ordinary for the United States; above 1,700 puts a city among the 12 at the top of the rankings, all of them in California, Arizona, Nevada, New Mexico or west Texas.
The same figure, three other ways
Per day. Divide by 365 and you get the average day. The median city makes 3.97 kWh per kW per day. City pages show this as “on an average day”, and the estimator multiplies it by your system size.
Capacity factor. Divide by the 8,760 hours in a year and you get the share of the panels' rated power they deliver on average, nights included. Denver's 18.29% means 1 kW of panels averages about 183 watts around the clock. Across the 100 cities the capacity factor runs from 10.85% in Anchorage to 20.95% in El Paso, with a median of 16.6%.
Per month. Monthly figures are the same measure for one month. The highest single month in the data is July in San Jose, CA, at 185.3 kWh per kW; the lowest is December in Anchorage, AK, at 8.8.
| City | kWh per kW a year | Per day | Capacity factor |
|---|---|---|---|
| El Paso, TX | 1,835 | 5.03 | 20.9% |
| Denver, CO | 1,602 | 4.39 | 18.3% |
| Charleston, SC | 1,455 | 3.99 | 16.6% |
| Boston, MA | 1,275 | 3.49 | 14.6% |
| Seattle, WA | 1,088 | 2.98 | 12.4% |
| Anchorage, AK | 951 | 2.60 | 10.8% |
The table runs from the top of the ranking to the bottom. Charleston, SC sits at the middle (50th of 100).
Sunlight on the panel explains almost all of it
Each city record also holds the average solar radiation reaching the panel surface, in kWh per square metre per day. Across the 100 cities the correlation between that and yearly output is 0.99, close to a perfect 1. Put simply, kWh per kW is very nearly a measure of how much sunlight lands on a south-facing 20° roof over a typical year. The model's treatment of each city's weather accounts for the small remainder.
The extremes show it. El Paso receives 6.66 kWh/m² a day on the panel and Anchorage 3.15. That is why the ranking of cities by output is close to a ranking by sunshine, and why the sunniest cities are where they are.
What the figure leaves out
The weather behind every number is a typical meteorological year: a composite of real months chosen to represent long-term conditions, not a forecast and not any particular year. A given year can come in higher or lower.
The roof is also an idealised one. There is no shading from trees, chimneys or neighbouring buildings, and the panels sit at exactly 20° facing due south. A real roof differs in pitch and direction, and snow, dust and the equipment chosen all move the result. For direction and tilt, the data does cover five alternatives; the direction guide shows how much each one changes output, and the estimator applies them to any city.
Read kWh per kW as the yardstick it is: a fair way to compare places on the same hardware, and a starting point for one particular roof.
Sources
- PVWatts® Calculator API v8, National Laboratory of the Rockies (NLR, formerly NREL): developer.nlr.gov. 600 model runs (100 cities × 6 layouts), retrieved 2026-10-27.
- Weather: NSRDB typical-meteorological-year data returned by PVWatts® (NSRDB PSM V3 GOES tmy-2020 3.2.0, NSRDB PSM V4 Polar tmy-2023 1.0.0).
- City list and locations: U.S. Census Bureau, Vintage 2024 Population Estimates for incorporated places (SUB-EST2024) and U.S. Census Bureau, 2024 Gazetteer Files - Places (census.gov).
- Every figure in this guide is computed by SunYieldMap from those model runs. Settings and limits are in the methodology.