Run the standard arithmetic on a typical 2026 install and you get a clean answer. A 7.2 kW system at $2.60 per watt is $18,720. At 1,400 kWh per installed kW it makes 10,080 kWh a year, and at the 18.44 cents per kWh the EIA reported for May 2026 that is $1,859 of electricity. Divide, and payback lands at 10.1 years.
That is the number our own assumptions post stress-tests one variable at a time, and it is the honest base case for what the calculators model. This post is about what they do not model at all.
Four lines. Every one of them is ordinary, none of them is in the division above, and together they move the answer by years rather than months.
One: Not Every Kilowatt-Hour Is Worth Retail
The $1,859 figure assumes all 10,080 kWh offset a purchase at 18.44 cents. They do not, because your house is not awake when your roof is busiest.
EnergySage puts a typical self-consumption rate for a battery-less home at around 30%, and the seasonal split in a case study it cites is instructive: 35.69% in autumn and winter, dropping to 27.68% in spring and summer, precisely when production peaks. The rest is exported.
What the export is worth depends entirely on your tariff. Under old-style net metering it was worth retail, which is why the simple division used to be defensible. Under the export-rate tariffs replacing it, it can be a third of retail or less.
| Export treatment | Value of 10,080 kWh | Simple payback |
|---|---|---|
| Full retail for everything | $1,859 | 10.1 years |
| 30% self-consumed, export at 10c | $1,264 | 14.8 years |
| 30% self-consumed, export at 6c | $981 | 19.1 years |
Look at the spread before you look at the middle. This single line, with nothing else changed, moves payback further than price per watt, degradation and panel efficiency combined. It is also the line you have most control over, because shifting load into daylight raises the self-consumed share for free.
Two: The Inverter Does Not Last 25 Years
Panels are warrantied for 25 to 30 years. String inverters carry 10 to 12. That mismatch is not a defect, it is the product design, and it means a replacement falls inside almost every payback window anybody quotes.
Budget $1,500 to $2,500 for a string inverter swap including labour, more for roof-level microinverters where the cost is access rather than hardware. A wall-mounted unit in the Fronius Primo 8.2-1 class is the cheap version of this event, because an electrician can reach it standing up. And read the warranty rather than trusting it: EnergySage notes that many manufacturers cover neither shipping nor the labour to reinstall, and that where labour is covered it is frequently capped, up to $500 being a common figure.
On the 14.8-year line above, a $2,000 inverter at year 12 adds roughly another 1.6 years. It arrives at the worst possible moment, just before the system was supposed to start being free.
Three: Insurance and Property Tax
Two small recurring lines that no calculator asks about.
Most homeowner policies will cover a roof-mounted array as part of the dwelling, but the insured value of the dwelling went up by $18,720, and premiums follow insured value. Call it $50 to $150 a year depending on carrier and state.
Property tax depends entirely on where you live. Plenty of states exempt residential solar outright; plenty do not, and an assessor who adds the system to your assessed value creates a bill that runs for the life of the array. Check your own state before modelling either answer, because this one is genuinely binary.
At the low end this pair is $50 a year and rounding error. At the high end it is $400 and it eats a fifth of your annual saving.
Four: Operations, Which Is Mostly Cleaning
Panels need very little. Very little is not nothing.
Soiling losses are real and local, which is why we built a soiling loss estimator rather than publishing one number. Somewhere between a hose and a professional clean once or twice a year, plus a monitoring subscription if yours is not free, plus the occasional service call, and you are at $150 to $300 a year on a system this size.
That is under 2% of installed cost annually, which is genuinely low for a capital asset. It is also about 15% of your annual saving in the retail-rate scenario and nearer 25% once export rates are applied.
Putting the Four Together
Same house, same panels, same sun. Only the missing lines added.
| Version | Annual net saving | Payback |
|---|---|---|
| Calculator answer | $1,859 | 10.1 years |
| Export rates applied | $1,264 | 14.8 years |
| Plus insurance, tax, O&M at $350 | $914 | 20.5 years |
| Plus insurance, tax, O&M at $200 | $1,064 | 17.6 years |
| Realistic middle, with inverter at year 12 | $1,219 | 15 to 17 years |
That last row is worth spelling out, because a range with no arithmetic behind it is just a mood. It assumes self-consumption pushed to 40% by shifting load into daylight, a 12 cent export rate, and $250 a year of ownership costs: 4,032 kWh at retail is $744, 6,048 kWh exported is $726, minus $250 leaves $1,219 a year and a 15.4-year payback. Add a $2,000 inverter at year 12 and you are at 17.0. Change any one of those three inputs and the range moves, which is the point. Anybody handing you a single decimal for a 25-year horizon is selling something.
Why This Matters More Than the Optimistic Number
Here is the opinion, and installers will hate it: the 8-year quote is doing more damage to residential solar than any policy change of the last two years.
A household told 8 years that gets 16 does not conclude that electricity prices moved. It concludes it was mis-sold, and it says so to everyone who asks. A household told 16 that gets 16 recommends its installer. The industry's own payback numbers got three years longer when Section 25D expired, and it survived that fine, because the change was explained. What it does not survive is arithmetic that quietly omits the export rate.
So model your own four lines before you sign anything. Ask the installer what export rate their estimate assumes, what the inverter replacement is budgeted at, and whether your state assesses solar. If the answer to any of the three is a pause, you have just learned what the estimate is worth.
Summary
A 7.2 kW residential system at $2.60 per watt divides out to a 10.1-year payback, and that division omits four ordinary costs. Export rates matter most: at a typical 30% self-consumption and a 10 cent export rate the same system saves $1,264 a year rather than $1,859, pushing payback to 14.8 years. A string inverter warrantied for 10 to 12 years against panels warrantied for 25 to 30 adds a $1,500 to $2,500 replacement inside the window. Insurance, property tax where it applies, and $150 to $300 of annual cleaning and monitoring take another bite. Model it honestly and a realistic range is 15 to 17 years, which is still a good deal on hardware that runs for 25 and a far better thing to promise than an eight-year figure the house will never see.