What percentage of homes have solar? Rooftop counts by state

Almost every published figure for “how many homes have solar” is built from installation records — utility interconnections, incentive claims, manufacturer shipments. Those are good numbers, but they measure paperwork. This page measures roofs: an image classifier looks at the current aerial photograph of every single-family home in the United States and decides whether panels are on it.

The answer, as of August 18, 2026: 4,451,792 of 80,265,597 classified single-family roofs carry solar panels — 5.55% of U.S. single-family homes. That the rooftop count lands this close to the installation-based estimates published by the industry is the best available evidence that both are roughly right.

Percentage of homes with solar, by state

All 48 states and the District of Columbia where roof classification covers at least 60% of the single-family stock. “Homes classified” is the denominator — the number of roofs actually looked at, not an estimate of the housing stock.

Hawaii and Alaska are missing on purpose. Hawaii almost certainly leads the country on rooftop solar, but only 22.8% of its single-family roofs are classified today and the classified subset is not a random sample — publishing a rate from it would produce a number that is both very high and meaningless. Alaska is at 0.1% coverage. We would rather have two holes in the table than two wrong rows.

What the ranking actually tracks

Not sunshine. Alabama, Mississippi, and Tennessee get more usable sun than Connecticut, Massachusetts, and New Jersey, and sit at the bottom of this table while the northeastern states sit near the top. The ordering follows retail electricity prices, net-metering rules, and state incentives far more closely than it follows irradiance.

Texas is the clearest illustration. It has more classified single-family roofs than any other state — 7.85 million, slightly more than California — and 315,348 of them have panels, for a rate of 4.02%. California, with a comparable roof count, is at 24.44%. Six times the adoption on the same sized housing stock is a policy outcome, not a weather outcome.

The counties with the most solar homes

San Diego County is the outlier worth staring at: 42.0% of its single-family roofs have panels, the highest of any large county in the country, and it holds essentially as many solar homes as Los Angeles County on a quarter of the housing stock. Inland California — Fresno, Kern, San Joaquin, Riverside — clusters in the mid-thirties, where hot summers and high bills meet cheap suburban roof area.

What a 5.55% national rate means if you sell solar

It means the market is still overwhelmingly unpenetrated, and that the national average is nearly useless for planning. In Alabama, 99.87% of roofs have no panels and almost none of them will get any this year. In San Diego, the market is closer to saturation than any sales deck admits, and the money has already moved to batteries, re-roofs under existing arrays, and the orphaned systems left by installer bankruptcies.

The useful cut is never the state rate. It is the street: which roofs near this one already have panels, which are old enough to need replacing first, and which homeowner is actually reachable. That is the same classification, run one address at a time.

Methodology

• Universe. Active single-family residential properties in Scout Data’s national property database — 85,072,153 of them. Condominiums, multi-family buildings, mobile homes, and non-residential parcels are excluded.
• Denominator. The 80,265,597 of those homes whose current roof has been classified from aerial imagery, which is 94.4% of the single-family stock. Homes with no imagery classification are excluded from both numerator and denominator rather than counted as “no solar” — the mistake that makes most rooftop estimates too low.
• Numerator. Homes where the classifier finds solar panels on the current roof image. This is a present-tense measure: a system installed in 2015 and removed in 2024 does not count, and a system installed without a permit does.
• Exclusions. Hawaii (22.8% classified), Alaska (0.1%), and Guam are excluded from all figures including the national rate.
• Known error modes. The classifier has a tail in both directions. It can miss a small array clipped at the edge of a crop, and it can call a run of skylights or a dark roof plane a panel. Errors are largest on very small arrays, which is why we treat sub-10-panel detections as unverified for commercial work. At state and county scale these largely wash out; at the level of one address they do not, which is why individual records carry a confidence grade.
• Imagery age. Aerial imagery is refreshed on a rolling basis and is not uniformly current. A system installed in the last few months may not appear yet, so every figure here is a slight undercount of today.

Derived August 18, 2026. If a number here looks wrong for a market you know well, tell us — that is usually how we find a coverage gap.

Frequently asked questions