Almost every answer to this question tells you to pull permits. Permits are a good record of something, and that something is not “which houses have solar panels on them”.
A permit says a jurisdiction approved an installation. It does not say the array was built, that it is still there, or that the roof it went on has not been replaced twice since. Meanwhile, systems get installed without permits, and permit coverage is uneven county by county in ways that have nothing to do with solar adoption.
There is a method that answers the actual question, and it is looking at the roof.
The three methods, ranked
Building permits. Public, cheap, and available at national scale. They tell you an approval happened on a date, which is genuinely valuable — it is the only one of the three methods that gives you a system’s age. As a presence test, though, it fails in both directions at once.
Utility interconnection records. The most authoritative source that a system was energised, and in some states it is published. The problem is coverage: it exists where a utility or regulator chose to publish it, in whatever format they chose, and it does not assemble into a national file. Excellent for one territory, unusable as a method.
Aerial imagery classification. A classifier looks at the current aerial photograph of a home and decides whether panels are on it. This measures the present tense. A system installed in 2015 and removed in 2024 does not count; a system installed without a permit does. It is the only method whose answer is about the house rather than about a filing cabinet.
The scale is real. Of 80,265,597 U.S. single-family roofs classified this way, 4,451,792 carry solar panels — 5.55% of the stock. That the roof count lands close to the installation-based estimates published by the industry is the best available evidence that both are roughly right. The full state and county breakdown is in homes with solar by state.
How far apart the two records actually are
This is not a theoretical gap. Joining permit records against roof classification on the same homes produces a large population where a solar permit exists and no array is visible on the roof today — the subject of solar permits that never got installed, where the rate tracks permit status in exactly the way you would want it to if the finding were real.
For your purposes the direction matters more than the size. If you build a “has solar” list from permits, some meaningful share of the doors you knock have no panels, and the opening line you rehearsed is wrong at the threshold. That is worse than a smaller list.
What the classifier gets wrong
Imagery is the best method available and it is not perfect. Three error modes are worth carrying into the field:
- Small arrays. Errors are largest on very small installations — a four-panel array clipped at the edge of a crop can be missed, and a run of skylights or a dark roof plane can be called a panel. Sub-ten-panel detections are the least reliable end of the distribution.
- Imagery age. Aerial imagery refreshes on a rolling basis and is not uniformly current. A system installed in the last few months may not appear yet, so any imagery-derived count is a slight undercount of today.
- Anything not on the roof. Ground mounts, carports and arrays on a detached garage are a different detection problem. A roof-classification list will under-report rural properties for this reason.
The practical answer is a confidence grade rather than a boolean. At state and county scale these errors largely wash out; at the level of one address they do not, and a rep at a door is always at the level of one address.
Five jobs that only exist on a solar home
- Battery retrofit. The clearest one. A house with panels and no storage is a qualified prospect for a product that did not exist, or did not pencil, when the array went in. Solar battery attachment rates covers how the attach rate has moved and where it has not.
- Re-roof under an array. A roofing job with a solar component, on a homeowner who has to solve both problems at once and usually has no idea who to call. Roof age plus visible panels is a two-signal list.
- Orphaned systems. Homes whose installer is no longer operating, left with a warranty nobody answers and a monitoring portal nobody maintains. This is a service and upgrade audience with unusually low competition — orphaned solar systems covers how the tiers differ and how teams work them.
- Panel and inverter upgrades. A system installed a decade ago is on its first inverter and its original modules. The permit date is what tells you which decade, which is the one thing permits are genuinely best at.
- Service contracts. Unglamorous, recurring, and essentially uncontested on any street where the original installers have moved on.
Build the list with two signals, not one
“Has solar” on its own is a poor filter, because on some streets it is most of the street. San Diego County is at 42.0% of classified roofs; Fresno, Kern and San Joaquin sit in the mid-thirties. A list of solar homes in those markets is close to a list of homes.
Pair it with a second signal and it becomes a reason to knock:
- Panels present + permit ten or more years old → upgrade and inverter conversations.
- Panels present + no storage → battery retrofit.
- Panels present + roof age past the local replacement interval → the re-roof-under-array job, which almost nobody is calling about.
- Panels present + installer no longer operating → service and monitoring.
- Panels present + recent sale → a new owner who inherited a system, often with no documentation and no idea who installed it.
Every one of those is a specific opening line, which is the actual deliverable. A list is only as good as the sentence it lets a rep say first.
Where Scout Data fits
Disclosure: Scout Data is our product, and roof classification is one of the things it does. List Builder holds a present-tense panel determination on U.S. single-family roofs, joined to permit history for system age and to ownership records, and delivers the result with phone numbers matched by name to the owner on title and scrubbed against the federal do-not-call registry. Scout shows the same determination on a map, which is how a canvasser sees which roofs on the street already have panels before knocking on any of them.
The mirror image of this page — finding the homes that do not have solar, which is a much harder claim to make honestly — is in how to find homes without solar.
Frequently asked questions
How do I find out which houses have solar panels?
Three methods exist and only one of them measures the roof. Building permits record that a system was approved; utility interconnection data records that one was connected; aerial imagery classification records whether panels are visible on the house today. The first two measure paperwork and the third measures the roof, which is why they disagree — and why the third is the one to build a list on.
What percentage of homes have solar panels?
5.55% of U.S. single-family homes — 4,451,792 of the 80,265,597 single-family roofs classified from aerial imagery. The spread by state is enormous: California is at 24.44%, Nevada 18.25%, Arizona 15.30%, while Alabama, Tennessee and North Dakota sit below 0.25%.
Can I get a list of homes with solar panels?
Yes, and the useful version specifies how the determination was made. A list built from permits is a list of homes that pulled a permit, which is neither all the homes with panels nor only the homes with panels. A list built from current imagery is a list of roofs with panels visible now. Ask which one you are being sold, because they answer different questions and are priced as if they were the same.
Why would I want homes that already have solar?
Because several of the best-margin jobs in the trade only exist on a house that already has an array: battery retrofit, a re-roof underneath panels, service and monitoring on a system whose installer is gone, and panel or inverter upgrades on a decade-old system. These are also the least-contested doors in solar, because most of the industry is still selling first systems.