Original research
Smart meter statistics: the infrastructure is built, the prices are not
Interval metering and time-varying tariff enrolment across all 51 US jurisdictions, measured against the right denominator.
Written by HyreSolar Research team Research and analysis
Audited by HyreSolar Research team Data audit and fact check
The finding
Before the number: which denominator this is measured against
Smart meter penetration is one of the easiest statistics in American energy to get slightly wrong, and the error is always in the same direction. There are more residential electricity meters in the United States than there are residential electricity accounts: 147,612,180 against 143,144,185, a difference of 4,467,995 meters, or 3.1%. A single customer can hold more than one meter: a second dwelling, a detached garage or workshop on its own service, a separately metered water heater or electric vehicle circuit.
EIA reports both numbers, on different schedules. Divide the AMI meter count by the meter count from the same schedule and you get 83.8%. Divide the same AMI count by residential customer accounts from the Sales schedule and you get 86.5%, 2.62 percentage points higher, for no reason other than mixing two schedules.
We use the meter denominator throughout. It is the only one that makes the ratio a statement about meters rather than an artefact of two different surveys, and it is why the headline figure here sits below several widely quoted ones. If you are comparing our number with another source's, check that first.
The arithmetic, in four steps
- 1 Take the residential meter total from the Advanced Meters schedule: 147,612,180
This column is the utility's own count of every residential meter on its system, whatever type. It is reported on the same sheet as the AMI count, by the same filer, for the same year.
- 2 Take the residential AMI meter count from the same sheet: 123,758,084
AMI means advanced metering infrastructure: a meter that records consumption in intervals and communicates them back without a visit. It is reported separately from AMR, which reads a cumulative register remotely but records no interval data.
- 3 Divide: 83.8% of residential meters are interval meters
Leaving 23,854,096 residential meters that are AMR or standard. HyreSolar analysis of EIA-861 2024.
- 4 Do not divide by customer accounts, which would give 86.5%
The numerator counts meters and the denominator would count accounts. The two differ by 3.1%, so the ratio inherits the difference and overstates penetration by 2.62 points. This is the most common error in published AMI figures.
The same numerator, two denominators
| Basis | Numerator | Denominator | Result | What it is a statement about |
|---|---|---|---|---|
| Meters (used here) | 123,758,084 AMI meters | 147,612,180 residential meters | 83.8% | The share of the metering estate that records intervals |
| Accounts | 123,758,084 AMI meters | 143,144,185 residential accounts | 86.5% | A ratio between two schedules, which is not a penetration rate |
Both figures are computed from EIA-861 2024. The numerator is identical; only the denominator changes.
The gap is 2.62 percentage points. It is small enough to look like rounding and large enough to move a state several places in a ranking, which is why it is stated here rather than buried in a note.
What is installed
On the meter basis, 83.8% of American residential meters record interval data. The rollout is effectively complete in 8 jurisdictions, which are above 95%, and 21 are above 90%. The median jurisdiction sits at 87.5%.
District of Columbia is the most fully metered at 99.5%, followed by Maine and Pennsylvania, both at 97.6%. This is what a finished programme looks like: a regulator approved a full deployment, the utility replaced essentially every meter, and there is nothing left to do.
10 jurisdictions are still below half. Rhode Island is the lowest in the country at 1.1% (11 interval meters per thousand) with Massachusetts at 10.4% and Wyoming at 39.3%. In each case the reason is regulatory rather than technical: a state commission that has not approved a full deployment, or has approved one that is not yet built.
What is actually used
Now the same picture for tariffs. 15,832,411 residential customers (11.1% of the 143,144,185 accounts in the country) are enrolled in a time-of-use or otherwise dynamic tariff. Set the two waffles side by side and the argument of this page is finished before a table appears: 838 squares of capability, 111 squares of use.
Put differently, of every hundred interval meters installed in the United States, roughly 13 are attached to an account that pays a price varying by hour. The remaining 87 record interval data that is used for billing a flat rate, for outage detection, for load research and for remote connection, all real uses, none of which changes what a household pays for the electricity it consumes at 6pm.
21 of 51 jurisdictions have time-varying enrolment below 1%. 37 are below 5%. Only 8 are above a quarter of households. The median is 1.8%, which is to say that in a typical American state a time-varying electricity price is a rounding error.
The geography of the metering estate
AMI penetration does not have the geography most people expect. It is not a coastal-versus-interior story and it is not a rich-state-versus-poor-state story. It is a regulatory story, and the map is close to a map of which state commissions approved a full deployment in the 2010s.
The clearest cluster is New England, and it runs the wrong way. Rhode Island 1.1%, Massachusetts 10.4%, Connecticut 21.7%, New Hampshire 21.8%, Vermont 78.1%, Maine 97.6%, 4 of the 6 below 25%. These are not states with unsophisticated grids or low electricity prices. They are states whose commissions took a sceptical view of the cost case for a full meter replacement, in several cases explicitly.
That matters here more than it would elsewhere, because those same states have high rooftop solar adoption. Across the 10 jurisdictions with the highest household solar penetration, average interval metering is 71.0%, against a national 83.8%. Connecticut is 6th in the country on solar adoption at 6.7% of households and 49th on interval metering at 21.7%. Massachusetts is 8th on adoption and 50th on metering.
At the other end, Tennessee has interval metering on 95.4% of residential meters and rooftop solar on 0.0027% of households. The infrastructure and the adoption are close to unrelated, and the correlation coefficients below say the same thing more formally.
Installing the meter does not change the price
The natural assumption is that time-varying tariffs follow meters: put in the infrastructure and the rate design follows. Across the 51 jurisdictions that is barely true. The correlation between AMI penetration and time-of-use enrolment is Pearson 0.222, positive, weak, and nothing like the near-one you would expect if metering were the binding constraint.
District of Columbia is the extreme case: 99.5% of meters record intervals and 0.0% of customers are on a time-varying rate. Every meter in the district can do it and essentially nobody is billed that way. Pennsylvania, Maine and Tennessee are all above 95% on metering and below 1% on enrolment.
Montana makes the opposite point. It has interval metering on only 28.4% of meters, one of the lowest in the country, and 17.0% time-varying enrolment, higher than 42 of the other 50 jurisdictions. Time-of-use billing predates AMI: a mechanical two-register meter that switches at a fixed hour has been available for decades, and some utilities have run those tariffs for a long time on legacy hardware.
Maryland is the one jurisdiction where the two line up as the theory predicts, at 86.5% metering and 69.3% enrolment, the highest in the country. It got there by default enrolment rather than by installing more meters, which is the actual lesson of this chart. Metering is a precondition, not a cause.
The 8 jurisdictions where time-varying pricing is normal
Enrolment is extraordinarily concentrated. 8 jurisdictions have more than a quarter of residential customers on a time-varying rate; the other 43 do not. In almost every one of the 8, the reason is the same: the tariff is the default for some or all residential customers rather than something a household opts into.
Maryland leads at 69.3%. Delaware follows at 55.5% and Missouri at 51.9%. California, the largest residential market in the country and the one whose export tariff design has changed the most, sits at 36.1%, high by national standards and far from universal.
The pattern to notice in the table below is that a high enrolment rate almost never comes with the highest metering rate. Maryland is 28th of 51 on AMI. Delaware is 10th. What separates these states from District of Columbia is not hardware.
The 12 highest time-varying enrolment rates
| # | Jurisdiction | On a time-varying rate | Interval metering | Share of the metered base priced by time | Solar penetration |
|---|---|---|---|---|---|
| 1 | Maryland | 69.3% | 86.5% | 80% | 4.42% |
| 2 | Delaware | 55.5% | 93.6% | 59% | 2.98% |
| 3 | Missouri | 51.9% | 87.6% | 59% | 1.02% |
| 4 | Colorado | 47.2% | 92.5% | 51% | 7.03% |
| 5 | Arizona | 39.1% | 93.4% | 42% | 10.11% |
| 6 | Michigan | 38.9% | 96.8% | 40% | 0.52% |
| 7 | California | 36.1% | 86.6% | 42% | 14.71% |
| 8 | Oklahoma | 33.2% | 92.8% | 36% | 0.89% |
| 9 | Montana | 17.0% | 28.4% | 60% | 1.64% |
| 10 | Louisiana | 9.8% | 89.0% | 11% | 1.60% |
| 11 | Illinois | 9.6% | 93.6% | 10% | 1.91% |
| 12 | North Dakota | 8.8% | 60.9% | 14% | 0.02% |
HyreSolar analysis of EIA-861 2024. The fifth column is time-varying enrolment divided by AMI penetration, and is approximate: the two are measured against accounts and meters respectively.
8 of these 12 have solar penetration below the 3.55% national rate. Time-varying pricing is not, on this evidence, something that arrives with rooftop solar.
All 51 jurisdictions, 2024
Sorted by interval metering penetration, highest first. Rank 1 is the highest on each measure.
| Jurisdiction | AMI share of meters | AMI rank | On a time-varying rate | TOU rank | Share of metered base priced by time |
|---|---|---|---|---|---|
| District of Columbia | 99.5% | 1 | 0.0% | 50 | 0% |
| Maine | 97.6% | 2 | 0.7% | 34 | 1% |
| Pennsylvania | 97.6% | 3 | 0.1% | 47 | 0% |
| Michigan | 96.8% | 4 | 38.9% | 6 | 40% |
| Nevada | 96.0% | 5 | 1.3% | 30 | 1% |
| Tennessee | 95.4% | 6 | 0.3% | 40 | 0% |
| Texas | 95.1% | 7 | 2.7% | 20 | 3% |
| Hawaii | 95.0% | 8 | 4.5% | 16 | 5% |
| Florida | 94.5% | 9 | 0.1% | 48 | 0% |
| Delaware | 93.6% | 10 | 55.5% | 2 | 59% |
| Illinois | 93.6% | 11 | 9.6% | 11 | 10% |
| Arizona | 93.4% | 12 | 39.1% | 5 | 42% |
| Georgia | 93.0% | 13 | 1.8% | 26 | 2% |
| Alabama | 92.9% | 14 | 0.2% | 43 | 0% |
| Oregon | 92.9% | 15 | 1.4% | 28 | 2% |
| Virginia | 92.9% | 16 | 0.7% | 33 | 1% |
| Kansas | 92.8% | 17 | 2.0% | 25 | 2% |
| Oklahoma | 92.8% | 18 | 33.2% | 8 | 36% |
| Colorado | 92.5% | 19 | 47.2% | 4 | 51% |
| North Carolina | 92.5% | 20 | 2.9% | 18 | 3% |
| New Jersey | 90.8% | 21 | 0.4% | 36 | 0% |
| Wisconsin | 89.8% | 22 | 1.8% | 27 | 2% |
| Louisiana | 89.0% | 23 | 9.8% | 10 | 11% |
| South Carolina | 88.7% | 24 | 0.3% | 37 | 0% |
| Missouri | 87.6% | 25 | 51.9% | 3 | 59% |
| Minnesota | 87.5% | 26 | 2.6% | 21 | 3% |
| California | 86.6% | 27 | 36.1% | 7 | 42% |
| Maryland | 86.5% | 28 | 69.3% | 1 | 80% |
| Utah | 86.4% | 29 | 0.3% | 38 | 0% |
| Arkansas | 83.8% | 30 | 7.8% | 13 | 9% |
| Kentucky | 81.4% | 31 | 0.3% | 39 | 0% |
| Idaho | 81.3% | 32 | 1.3% | 29 | 2% |
| Indiana | 80.5% | 33 | 3.2% | 17 | 4% |
| Mississippi | 79.9% | 34 | 2.1% | 24 | 3% |
| South Dakota | 78.5% | 35 | 2.8% | 19 | 4% |
| Vermont | 78.1% | 36 | 2.5% | 22 | 3% |
| Alaska | 74.1% | 37 | 0.0% | 49 | 0% |
| New York | 74.1% | 38 | 2.5% | 23 | 3% |
| Washington | 69.4% | 39 | 0.3% | 42 | 0% |
| Ohio | 62.3% | 40 | 0.7% | 32 | 1% |
| North Dakota | 60.9% | 41 | 8.8% | 12 | 14% |
| Iowa | 45.9% | 42 | 6.0% | 14 | 13% |
| West Virginia | 44.7% | 43 | 0.3% | 41 | 1% |
| Wyoming | 39.3% | 44 | 0.6% | 35 | 2% |
| Nebraska | 31.1% | 45 | 0.2% | 44 | 1% |
| Montana | 28.4% | 46 | 17.0% | 9 | 60% |
| New Mexico | 28.4% | 47 | 0.9% | 31 | 3% |
| New Hampshire | 21.8% | 48 | 0.1% | 46 | 1% |
| Connecticut | 21.7% | 49 | 4.6% | 15 | 21% |
| Massachusetts | 10.4% | 50 | 0.2% | 45 | 2% |
| Rhode Island | 1.1% | 51 | 0.0% | 51 | 0% |
HyreSolar analysis of EIA-861 2024 Advanced Meters and Dynamic Pricing schedules. AMI is measured against the residential meter count; time-varying enrolment against residential customer accounts.
Figures labelled HyreSolar analysis are computed by us from the EIA source files named below. EIA publishes the inputs; it does not publish these ratios or rankings.
Why an export tariff cannot exist without this
Smart meters belong in a solar research library for one reason. Every tariff change that has reshaped rooftop solar economics in the last five years (net billing, hourly export credits, avoided-cost export rates, capacity-based charges) requires a meter that records what happened in each hour. Without interval data, a utility can only know a household's net consumption over a month, and a monthly net is a retail-rate net metering credit whether anyone intended it or not.
What interval metering makes possible
Hourly export valuation. Crediting a kilowatt-hour exported at 1pm differently from one exported at 7pm requires knowing when it was exported. A cumulative register cannot say.
Time-of-use retail rates. The consumption side of the same problem, and the reason a solar system's value depends on which hours it offsets rather than on how much it generates.
Bill-level verification. A household can check an export credit against its own interval data, which is a meaningful consumer protection and did not previously exist.
Storage arbitrage. A battery that shifts export from midday to evening only earns anything if the tariff distinguishes the two. 5.9% of the US solar fleet has a battery; the tariff is what decides whether it pays. Battery statistics →
What it does not do
It does not change a household's rate by itself. 87% of American interval meters are attached to a flat tariff, and a flat tariff behaves identically whether the meter behind it records intervals or not.
It does not measure generation. The meter sits at the property boundary and records import and export. Self-consumed solar output never crosses it, which is why no dataset in this library can tell you what a rooftop system produced.
It does not imply the data is available to the customer. Access to one's own interval data varies enormously by state and utility, and EIA does not collect anything about it.
It does not settle the export question. What sits outside net metering, and what that does to the numbers, is the net metering page.
The vocabulary, precisely
Four terms are used loosely enough in coverage of this subject that a page reporting figures on them has to fix them first.
- AMI
- Advanced metering infrastructure. A meter that records consumption in intervals (typically 15, 30 or 60 minutes) and communicates them back to the utility over a two-way network. This is what "smart meter" usually means and it is the 123,758,084 figure on this page.
- AMR
- Automated meter reading. A meter whose cumulative register can be read remotely (by drive-by radio or a one-way network) but which records no interval data. It removes the meter reader and nothing else. AMR meters are counted in the 23,854,096 non-interval residential meters here, not in the AMI figure. Conflating the two is the second most common error in published smart meter statistics, after the denominator.
- Time-of-use, and dynamic pricing more broadly
- EIA’s Dynamic Pricing schedule counts residential customers enrolled in any tariff whose price varies with time: time-of-use blocks, critical peak pricing, peak-time rebates, variable peak and real-time pricing. We report the total, 15,832,411 customers, and do not break it down, because the schedule’s sub-categories are inconsistently populated between filers.
- Net billing
- An export arrangement in which electricity sent to the grid is credited at a rate set by when it was exported, rather than netted against consumption at the retail rate. It is arithmetically impossible without interval metering, which is the connection between this page and the rest of this research library.
Methodology
Sources and extraction
US Energy Information Administration, Form EIA-861, 2024 final release. Two schedules: Advanced Meters, for the AMI, AMR, standard and total residential meter counts, and Dynamic Pricing, for residential customers enrolled in a time-varying tariff. Both downloaded on 2 September 2026.
Both schedules are read by script. Their headers are stacked over several rows and the depth differs between them, so columns are addressed by the tuple of their ancestor labels and the sector leaf rather than by position, addressing by column index reads a data row as a header when a schedule changes shape between years.
Utility-level rows are summed to the state. Nothing on this page is transcribed by hand, and no figure in the prose is typed: each is a template expression reading the same JSON the tables read.
Why the denominator is meters and not accounts
The Advanced Meters schedule reports a residential meter total alongside the AMI count. That is the correct denominator, because it is the same population the numerator is drawn from, reported by the same filer on the same sheet for the same year.
The alternative denominator is the residential customer count from the Sales to Ultimate Customers schedule, which is what several published AMI figures use. Nationally it is 143,144,185 against 147,612,180 meters (3.1% fewer) because a customer can hold more than one meter. Using it produces 86.5% rather than 83.8%, an overstatement of 2.62 points.
The excess is not uniform across states, so the error is not a constant offset that could be ignored for ranking purposes. That is the practical reason to insist on it rather than treat it as a rounding argument.
The one denominator we could not make consistent
Time-varying enrolment is reported on the Dynamic Pricing schedule as a count of customers, and that schedule carries no meter total. So the 11.1% figure is measured against residential customer accounts while the 83.8% figure is measured against meters. We could have forced them onto a common denominator; we have not, because doing so would mean dividing one schedule's numerator by another schedule's denominator, which is exactly the error the section above is about.
The practical consequence is that the "share of the metered base priced by time" column, and the national 12.8% equivalent, are approximate. Both denominators are within 3.1% of each other nationally, so the column is accurate to roughly that tolerance, enough to compare states, not enough to quote to a decimal place. It is labelled as approximate everywhere it appears.
Definitions used in the figures
AMI penetration, residential AMI meters ÷ total residential meters, both from the Advanced Meters schedule, same year, same filers.
Time-varying enrolment, residential customers enrolled in any dynamic pricing programme ÷ residential customer accounts.
Share of the metered base priced by time, time-varying enrolment ÷ AMI penetration. Approximate, per the section above.
Correlation, Pearson product-moment across the 51 jurisdictions, each jurisdiction weighted equally. The coefficient against solar adoption is taken from the dataset's correlation matrix, which also reports a version excluding Hawaii and California as adoption outliers.
Limitations
A meter type is not a capability. EIA records whether a meter is AMI. It does not record the interval length, whether the utility retains the interval data, whether the customer can see it, or whether the billing system can act on it. A state at 95% AMI may or may not be able to run an hourly export tariff tomorrow.
Enrolment is not exposure. A customer counted as enrolled in a time-varying tariff may be on a rate whose peak and off-peak prices differ by very little. The schedule counts enrolment, not the size of the price signal, and a nominal differential does nothing to behaviour.
Default versus opt-in is not reported. The difference between Maryland at 69.3% and District of Columbia at 0.0% is almost certainly a default-enrolment policy, but EIA does not collect that field, so we describe it as the likely explanation rather than assert it.
Residential only. Every figure here excludes commercial, industrial and transportation meters. Commercial interval metering has been near-universal for far longer and including it would flatter the national figure substantially.
One year, no series. These schedules are reported here for 2024 only. This page cannot show a rollout curve, and we would rather say so than interpolate one.
Reproducing this
The inputs are two public EIA workbooks, the extraction is a script, and every derived ratio is defined above. If a figure here disagrees with one you have computed, the denominator is the first thing to check, and if the disagreement survives that, tell us and we will correct the page with a dated note.
What this means for anyone reading about export credits
The metering layer is no longer the constraint. 21 of 51 jurisdictions have interval metering on more than 90% of residential meters, and the national figure is 83.8%. Anywhere in that group, a regulator that wants to move from monthly netting to hourly export valuation has the measurement in place to do it.
What has not happened is the tariff. 11.1% of American households face a price that changes with the hour, and 87% of the interval meters in the country are billing a flat rate. Every argument about whether rooftop solar is fairly compensated is being had in a country that has already built the equipment to answer it and has, in most places, chosen not to switch it on.
Which way that resolves is a policy question and this dataset has no view on it. What the dataset does establish is that the answer will not be delayed by hardware.
Questions
What percentage of US homes have smart meters?
What is the difference between AMI and AMR?
Which state has the most smart meters?
Do I need a smart meter for solar panels?
How many households are on a time-of-use electricity rate?
Does having a smart meter mean my rate changes with the time of day?
Does smart meter penetration predict rooftop solar adoption?
How current is this and when does it update?
Written and audited by
HyreSolar Research
Primary-source research, data analysis and fact checking
We are a research desk, not a sales floor. We read the statute, the tariff, the code section, the federal filing or the manufacturer data sheet ourselves, and we publish the figure with the document it came from and the date we retrieved it. Where a number cannot be traced to a primary source, we publish the shorter page and say what we could not verify. That rule has cost us whole sections, and it is the reason the rest can be trusted.
- 160
- primary sources read and cited
- 220
- figures with a retrieval date
- 115
- federal and state government sources
- 66
- researched pages published
How this desk works
- Primary sources only. Statutes from the legislature’s own publishing system, federal data from the agency that collects it, code text from the adopted edition, manufacturer claims from the data sheet. We do not cite an article that cites a source; we go and read the source.
- Every figure carries its provenance. A named document and the date we retrieved it, so you can check it and so you know how old it is. Retrieval dates are not decoration: an EIA rate from May is a different fact from an EIA rate from August.
- We publish what we could not verify. Every research page carries a section naming the things we tried to establish and could not, and why. A paywalled standard, a state website that refused the request, a manufacturer that publishes no figure at all.
- We separate measurement from modelling from our own reasoning, and label which is which on the page. A laboratory measurement, an assumption inside a modelling tool and our own inference are three different kinds of claim and they are never presented as one.
- We do not sell solar, and we take no payment for placement, ranking or a favourable mention. Nobody buys a position on this site.
Data as of EIA-861 2024 final release. Authorship on this site is organisational: the analysis belongs to the desk rather than to a named individual, and we do not publish credentials we do not hold. Our editorial policy sets out how we source, date and correct what we publish.
Sources & retrieval dates
- US EIA, Form EIA-861, Advanced Meters (2024 final release) — Residential AMI, AMR, standard and total meter counts by utility and state. The source for every metering figure on this page, including the denominator. Retrieved 2 September 2026.
- US EIA, Form EIA-861, Dynamic Pricing (2024 final release) — Residential customers enrolled in time-of-use, critical peak, peak-time rebate, variable peak and real-time pricing programmes, by utility and state. Retrieved 2 September 2026.
- US EIA, Form EIA-861 — Net Metering (annual files, 2014–2024) — Utility-level net-metered installations, capacity and PV-paired batteries by state and customer sector. Retrieved 2 September 2026.
- US EIA, Form EIA-861 — Sales to Ultimate Customers (annual files, 2014–2024) — Residential revenue, sales and customer counts by utility and state. The basis for price, bill and consumption figures. Retrieved 2 September 2026.
- US EIA, Average Price of Electricity to Ultimate Customers by End-Use Sector — EIA’s published price series, used to validate our derivation. Agreement across all 357 overlapping state-years is within 0.005¢/kWh. Retrieved 2 September 2026.
A time-varying rate changes what solar is worth
If your tariff prices the evening differently from the afternoon, the value of a system depends on which hours it covers. The net billing calculator works that through on your own rate.
HyreSolar is an independent analysis and matching service. We are not an installer, lender or utility. When a reader asks to be introduced, installers may pay us a referral fee. That fee never buys ranking, scores or placement in research. Our editorial policy sets out the rules.