Internet is available, but not every household subscribes
Most US households pay for internet: in the Census Bureau's 2025 survey the share with a broadband subscription runs from 89.7% in West Virginia to 95.6% in Colorado and California. Where a subscription is missing, a network often is not. In Springfield, Missouri, fiber is reported at 93.0%2 of homes and businesses, and 16.2%1 of households have no internet subscription, the highest share among large US cities.
Published November 13, 2026. Data: FCC National Broadband Map, availability as of June 30, 2026 (file revision of November 10, 2026); Census Bureau, American Community Survey 2025 1-year estimates.
Source: Census Bureau, American Community Survey 2025 1-year detailed tables; Census Bureau, Cartographic boundary file cb_2025_us_state_20m. Darker states have larger gaps; DC is the dot. Period: 2025.
Show the numbers
| State | Households without an internet subscription |
|---|---|
| West Virginia | 10.1% |
| New Mexico | 9.4% |
| Mississippi | 8.9% |
| Iowa | 8.8% |
| Arkansas | 8.3% |
| Kentucky | 8.0% |
| South Dakota | 8.0% |
| Louisiana | 7.9% |
| North Dakota | 7.7% |
| Alabama | 7.5% |
| Missouri | 7.3% |
| Wisconsin | 7.2% |
| Montana | 7.1% |
| Pennsylvania | 7.0% |
| Nebraska | 6.9% |
| Oklahoma | 6.9% |
| Minnesota | 6.8% |
| Indiana | 6.7% |
| Ohio | 6.7% |
| Wyoming | 6.7% |
| Maine | 6.6% |
| South Carolina | 6.6% |
| Illinois | 6.5% |
| Tennessee | 6.5% |
| Idaho | 6.4% |
| Kansas | 6.4% |
| Vermont | 6.4% |
| Michigan | 6.3% |
| Rhode Island | 6.3% |
| New York | 6.2% |
| District of Columbia | 6.0% |
| Virginia | 6.0% |
| North Carolina | 5.9% |
| Connecticut | 5.7% |
| Georgia | 5.5% |
| Hawaii | 5.5% |
| Texas | 5.4% |
| Utah | 5.4% |
| Arizona | 5.2% |
| Florida | 5.2% |
| Oregon | 5.2% |
| Maryland | 5.1% |
| New Hampshire | 5.1% |
| Alaska | 5.0% |
| Nevada | 5.0% |
| New Jersey | 4.9% |
| Delaware | 4.8% |
| Massachusetts | 4.7% |
| Washington | 4.5% |
| California | 4.4% |
| Colorado | 4.4% |
How many households have internet
The Census Bureau's American Community Survey asks households whether anyone pays for internet service and what kind. In the 2025 survey, West Virginia has the lowest share with a broadband subscription of any type, 89.7%; every other state is at 90.5% or above, and Colorado and California lead at 95.6%. The median state is at 93.5%1. The full table, with margins of error, is in internet subscriptions by state.
"Any type" includes a phone's cellular data plan. Subscriptions to a wired line (cable, fiber or DSL) are lower in every state: from 66.6% of households in Alaska, the lowest, to 84.0% in New Hampshire, the highest1. On the supply side, the FCC reports wired service of at least 100/20 Mbps at 92.3%2 of US homes and businesses as of June 30, 2026 (revision of November 10, 2026).
The two figures cannot be subtracted. The FCC counts locations on one date, and an apartment building is one location; the Census counts households across a calendar year and publishes each estimate with a 90% margin of error. "Available" means a provider reports it could install standard service within ten business days, a claim anyone can challenge on the FCC map. Read side by side, the two sources still show where the gap is a missing network and where it is a missing subscription.
Source: Census Bureau, American Community Survey 2025 1-year detailed tables; Census Bureau, Cartographic boundary file cb_2025_us_state_20m. Darker states have larger gaps; DC is the dot. Period: 2025.
Show the numbers
| State | Households without an internet subscription |
|---|---|
| West Virginia | 10.1% |
| New Mexico | 9.4% |
| Mississippi | 8.9% |
| Iowa | 8.8% |
| Arkansas | 8.3% |
| Kentucky | 8.0% |
| South Dakota | 8.0% |
| Louisiana | 7.9% |
| North Dakota | 7.7% |
| Alabama | 7.5% |
| Missouri | 7.3% |
| Wisconsin | 7.2% |
| Montana | 7.1% |
| Pennsylvania | 7.0% |
| Nebraska | 6.9% |
| Oklahoma | 6.9% |
| Minnesota | 6.8% |
| Indiana | 6.7% |
| Ohio | 6.7% |
| Wyoming | 6.7% |
| Maine | 6.6% |
| South Carolina | 6.6% |
| Illinois | 6.5% |
| Tennessee | 6.5% |
| Idaho | 6.4% |
| Kansas | 6.4% |
| Vermont | 6.4% |
| Michigan | 6.3% |
| Rhode Island | 6.3% |
| New York | 6.2% |
| District of Columbia | 6.0% |
| Virginia | 6.0% |
| North Carolina | 5.9% |
| Connecticut | 5.7% |
| Georgia | 5.5% |
| Hawaii | 5.5% |
| Texas | 5.4% |
| Utah | 5.4% |
| Arizona | 5.2% |
| Florida | 5.2% |
| Oregon | 5.2% |
| Maryland | 5.1% |
| New Hampshire | 5.1% |
| Alaska | 5.0% |
| Nevada | 5.0% |
| New Jersey | 4.9% |
| Delaware | 4.8% |
| Massachusetts | 4.7% |
| Washington | 4.5% |
| California | 4.4% |
| Colorado | 4.4% |
The wire reaches the door; the subscription does not
The sharpest cases are cities where wired service is reported almost everywhere. Among large US cities with a one-year survey estimate, the share of households with no internet subscription is highest in Springfield, Missouri, at 16.2% (±2.7 percentage points). There, fiber is reported at 93.0% of locations and 92.6% have two or more wired providers at 100/20 Mbps. Not all of those households are offline: only 4.5% report no internet access at all; the rest report having access without a paid subscription. In Chattanooga, where EPB reports fiber at 100.0% of locations, 11.2% of households (±3.0 percentage points) have no subscription, the 4th highest share. Dayton (12.3%, 3rd highest) and Albany (11.1%, 5th highest1) have wired 100/20 service at 96.5% and 96.5%2 of locations.
At the other end are suburbs with the same reach. In Carlsbad, California, wired 100/20 service is reported at 97.7%2 of locations and 0.7% of households (±0.5 percentage points) have no subscription, the lowest share among large cities. Sandy Springs, Georgia (1.1%) and Frisco, Texas (1.6%1) are also near the bottom.
| Area | Wired 100/20 Mbps | Two or more wired choices | Households subscribing to cable, fiber or DSL | Households relying only on a cellular data plan | Households without an internet subscription |
|---|---|---|---|---|---|
| Springfield, Missouri | 99.2% | 92.6% | 67.8% ±3.8 | 13.1% ±2.4 | 16.2% ±2.7 |
| Chattanooga, Tennessee | 100.0% | 98.4% | 74.6% ±2.9 | 12.6% ±2.5 | 11.2% ±3.0 |
| Dayton, Ohio | 96.5% | 61.7% | 67.3% ±3.9 | 16.9% ±4.3 | 12.3% ±3.5 |
| Albany, New York | 96.5% | 0.4% | 72.9% ±3.5 | 12.4% ±2.8 | 11.1% ±3.1 |
| Carlsbad, California | 97.7% | 44.0% | 82.3% ±3.8 | 13.8% ±3.9 | 0.7% ±0.5 |
| Sandy Springs, Georgia | 99.9% | 77.9% | 87.8% ±2.1 | 8.4% ±2.9 | 1.1% ±0.8 |
| Frisco, Texas | 98.7% | 91.3% | 89.9% ±1.8 | 6.6% ±2.2 | 1.6% ±0.9 |
Source: FCC, National Broadband Map, data as of June 30, 2026; Census Bureau, American Community Survey 2025 1-year detailed tables. FCC file revision of November 10, 2026. Shares are of broadband serviceable locations (homes and businesses), not people. ± is the 90% margin of error.
The number of wired networks does not sort these cities. Springfield and Chattanooga have two or more wired providers at nearly every location and are near the top for non-subscription; Albany has a second wired provider at only 0.4% of locations and sits just below them; Carlsbad, with two or more at 44.0%2, is at the bottom. Where every address can order a line, the share that does is a question about households, which the FCC map does not describe.
A phone plan in place of a home line
Many households counted as subscribed have no line to the home at all. The Census survey counts a cellular data plan as a broadband subscription, and in Detroit 20.8% of households rely on one with no other internet subscription, the 13th highest share among large cities. Wired 100/20 service is reported at 99.2%2 of the city's locations. In Frisco the cellular-only share is 6.6%1.
In Dayton, 16.9% of households rely on a phone plan alone and another 12.3%1 have no subscription, in a city with wired 100/20 service at 96.5%2 of locations. A phone plan is a subscription in the survey and serves some households fully; it is not the same service as a home connection, and the FCC map does not count mobile service as fixed broadband.
Where the network is the limit
Rural counties show the opposite pattern. In Park County, Colorado, wired 100/20 service is reported at only 14.1% of locations and 65.2% are unserved by the BEAD definition, yet only 6.7% of households (±2.4 percentage points) have no subscription in the five-year survey for 2020–2024. 27.9% subscribe to satellite. Bonner County, Idaho, is similar: wired 100/20 at 34.4%2 of locations, 25.7% of households on satellite, 9.4%4 with no subscription.
Compare two counties with wired service at nearly every location. In the Bronx, wired 100/20 is reported at 99.3% of locations and 13.1% of households (±0.5 percentage points) have no subscription. In Russell County, Alabama, the figures are 99.1%2 and 20.2% (±2.8 percentage points4). A county with little wired service can have a smaller unsubscribed share than one with wired service everywhere.
| Area | Wired 100/20 Mbps | Households subscribing to cable, fiber or DSL | Households with satellite internet | Households relying only on a cellular data plan | Households without an internet subscription |
|---|---|---|---|---|---|
| Park County, Colorado | 14.1% | 53.3% ±4.3 | 27.9% ±4.1 | 12.9% ±3.5 | 6.7% ±2.4 |
| Bonner County, Idaho | 34.4% | 49.5% ±3.0 | 25.7% ±2.3 | 15.1% ±2.1 | 9.4% ±1.6 |
| Bronx County, New York | 99.3% | 65.0% ±0.7 | 4.9% ±0.3 | 18.9% ±0.5 | 13.1% ±0.5 |
| Russell County, Alabama | 99.1% | 54.3% ±2.5 | 10.2% ±1.7 | 19.3% ±3.0 | 20.2% ±2.8 |
Source: FCC, National Broadband Map, data as of June 30, 2026; Census Bureau, American Community Survey 2020-2024 5-year detailed tables. FCC file revision of November 10, 2026. Shares are of broadband serviceable locations (homes and businesses), not people. ± is the 90% margin of error.
At the state level, Alaska is the only state where the two measures roughly match: wired 100/20 service is reported at 65.8%2 of locations and 66.6%1 of households subscribe to a wired line. West Virginia shows both gaps at once.
Source: FCC, National Broadband Map, data as of June 30, 2026; Census Bureau, American Community Survey 2025 1-year detailed tables. Availability is a share of locations as of June 30, 2026 (FCC); subscriptions are a share of households from the American Community Survey, 2025 (whiskers: 90% margin of error). Different units and dates: read the gap as a sign, not an exact number.
Show the numbers
| Measure | Share | MOE |
|---|---|---|
| Wired 100/20 Mbps available (FCC, share of locations) | 81.9% | |
| Households subscribing to cable, fiber or DSL (Census) | 68.6% | ±1.1 |
| Households relying only on a cellular data plan | 13.5% | ±0.7 |
| Households with no internet subscription | 10.1% | ±0.7 |
West Virginia has wired 100/20 service at 81.9% of locations, below the national 92.3%2, and the highest share of households with no subscription of any state, 10.1% (±0.7 percentage points1).
What offline households say
The NTIA's Internet Use Survey asks households where no one uses the internet at home for their main reason. In November 2023 three of those reasons came in the same order in every state: no need or interest first, cost second, no available service third. In Maryland 70.5% gave no need or interest and 12.3% cost; in Arkansas the split was 49.7% and 25.1%. Alaska had the highest share naming no available service, 10.0%5.
| State | Households where no one uses the internet at home | Offline households: main reason is no need or interest | Offline households: main reason is cost | Offline households: main reason is no service available |
|---|---|---|---|---|
| Maryland | 10.3% ±2.6 | 70.5% ±12.7 | 12.3% ±9.2 | 0.0% |
| New Hampshire | 8.9% ±2.0 | 70.4% ±10.6 | 8.4% ±7.8 | 1.9% ±3.3 |
| Arkansas | 16.3% ±2.2 | 49.7% ±5.1 | 25.1% ±4.8 | 4.4% ±3.5 |
| Mississippi | 17.0% ±2.2 | 59.1% ±8.0 | 21.7% ±5.9 | 4.8% ±2.3 |
| Alaska | 17.1% ±3.1 | 55.8% ±10.1 | 14.3% ±6.9 | 10.0% ±7.7 |
Source: NTIA Data Explorer "Analyze Table". Period: November 2023. ± is the 90% margin of error.
These answers come from a small group within each state's sample, so their margins of error are wide, often several times those of the Census figures above. They show the order of the reasons, not precise differences between states.
Fewer unsubscribed households than in 2019
The unsubscribed share fell in every state between the 2019 and 2025 surveys. Mississippi's fell from 23.0% (±0.7 percentage points) to 8.9% (±0.6 percentage points); West Virginia's from 18.6%6 to 10.1%1.
| State | Households without an internet subscription | Households without an internet subscription | Households without an internet subscription |
|---|---|---|---|
| Mississippi | 23.0% ±0.7 | 11.6% ±0.6 | 8.9% ±0.6 |
| New Mexico | 21.1% ±0.8 | 9.4% ±0.6 | 9.4% ±0.6 |
| Arkansas | 19.9% ±0.6 | 9.1% ±0.5 | 8.3% ±0.5 |
| Louisiana | 19.3% ±0.5 | 9.5% ±0.4 | 7.9% ±0.4 |
| West Virginia | 18.6% ±0.7 | 10.4% ±0.5 | 10.1% ±0.7 |
| New Hampshire | 10.5% ±0.7 | 4.6% ±0.4 | 5.1% ±0.6 |
| Colorado | 8.8% ±0.4 | 5.0% ±0.3 | 4.4% ±0.3 |
Source: Census Bureau, American Community Survey 2019 1-year detailed tables; Census Bureau, American Community Survey 2024 1-year detailed tables; Census Bureau, American Community Survey 2025 1-year detailed tables. Period: 2019. ± is the 90% margin of error.
The federal Affordable Connectivity Program, which had 22,591,9238 households enrolled in February 2024, ended on June 1, 2024. In the 2025 survey, the first full year without it, the unsubscribed share was below its 2024 level in most states; Mississippi's went from 11.6%7 to 8.9%1. The survey does not say why households subscribe, so these figures do not measure the program's effect.
Reading availability and subscription together
The FCC data contain no prices. The standing federal help with a bill is Lifeline, a $9.25 monthly discount for eligible households; eligibility is on the FCC's Lifeline page.
When a page quotes one coverage number for "households with internet", check which it is. FCC availability counts locations where a provider reports service. Census subscription counts households that pay for internet of any kind, phone plans included, and carries a margin of error. In most large US cities wired service is reported at nearly every location, so the households without a subscription there are inside the networks' reach. In rural counties with little wired service, satellite subscriptions fill much of the difference.
Questions about internet subscriptions and availability
What percentage of US households have internet?
The Census Bureau's 2025 American Community Survey puts the share of households with a broadband subscription of any type between 89.7% (West Virginia) and 95.6% (Colorado and California); the median state is at 93.5%. A cellular data plan counts as a subscription, so the share subscribing to cable, fiber or DSL is lower: from 66.6% in Alaska to 84.0%1 in New Hampshire.
Why don't people get internet when it is available?
In the NTIA's Internet Use Survey of November 2023, households where no one used the internet at home most often gave no need or interest as the main reason, ahead of cost; few said service was not available. The state-level answers carry wide margins of error.
Does a phone plan count as having home internet?
In the Census survey, yes: a cellular data plan is a broadband subscription. In Detroit 20.8%1 of households rely on one with no other internet subscription.
Sources
- Census Bureau, American Community Survey 2025 1-year detailed tables; 2025 (calendar-year sample); released October 29, 2026. api.census.gov
- FCC, National Broadband Map, data as of June 30, 2026; Availability as of June 30, 2026; released November 10, 2026. broadbandmap.fcc.gov
- Census Bureau, Cartographic boundary file cb_2025_us_state_20m. www2.census.gov
- Census Bureau, American Community Survey 2020-2024 5-year detailed tables; 2020-2024; released January 29, 2026. api.census.gov
- NTIA Data Explorer "Analyze Table"; November 2023 (also all prior waves 1994-2021). www.ntia.gov
- Census Bureau, American Community Survey 2019 1-year detailed tables; 2019. api.census.gov
- Census Bureau, American Community Survey 2024 1-year detailed tables; 2024. api.census.gov
- USAC, ACP Enrollments by County as of February 8, 2024; Enrolled households as of 2024-02-08 (Data Month January 2024 file). www.usac.org
Margins of error (±) are at the 90% confidence level. A difference is called significant only when it passes the Census Bureau's test at that level. How every figure is built: methodology. Found an error? Write to [email protected]; corrections are listed on the updates page.