Blog
Online access in New York households
Census figures put 94.7 percent of New York City households on a computer and 90.5 percent on broadband. What those numbers count, and what they leave out.
Of the 3,334,088 households in New York City, 94.7 percent have a computer and 90.5 percent have a broadband internet subscription. Those are the Census Bureau's estimates for 2020 through 2024, published on its QuickFacts page for the city. They describe a place where being online is the norm, and they also describe a gap that is large in absolute terms: about one household in twenty has no computer, and about one in ten has no broadband subscription.
The arithmetic is worth doing once. Five-point-three percent of 3,334,088 households is roughly 177,000 households without a computer. Nine-point-five percent is roughly 317,000 households without a broadband subscription. Both figures are estimates from a survey, so read them as "on the order of," not as a head count. But the scale is clear: the share sounds small, and the number of homes behind it is larger than the population of many American cities.
What the survey counts
The figures come from the American Community Survey, the Census Bureau's continuing household survey. Two of its questions matter here, and each has a definition that is wider than the everyday sense of the words.
A computer in the survey includes desktops and laptops, but also smartphones, tablets and other portable wireless computers. A household that owns only a phone counts as a household with a computer. So the 94.7 percent figure does not mean nearly everyone has a keyboard and a screen on a desk. It means nearly everyone has some device that can run apps and browse the web.
A broadband subscription includes a cable, fiber-optic or DSL line, a satellite service, and also a cellular data plan on a phone or tablet. A household that has no home line but does have a data plan on a phone is counted as subscribed. That is why the gap between "computer" and "broadband" is only about four points: the survey treats a phone plan as a connection.
The distinction matters for how the numbers are used. A household that depends on a phone's data plan can send a text, check a bus time and open an email, but filling out a long online application, doing homework on a small screen or attending a video appointment is a different experience. The survey does not measure that difference. It records whether a connection of any kind exists.
The households behind the numbers
The same QuickFacts table gives the setting. The city has 3,334,088 households with an average of 2.48 people each. About 36.6 percent of residents are foreign-born, and 47.7 percent of people age 5 and over speak a language other than English at home. The owner-occupied rate is 32.8 percent, which means about two-thirds of households rent. The median household income is $80,483 in 2024 dollars, and the median gross rent is $1,821.
Each of these facts touches internet access in a way the headline numbers cannot show:
- Renters rarely choose their building's wiring. A tenant picks a provider from the ones that serve the address, and a landlord's decisions about the building shape which ones those are. In a city where two-thirds of households rent, the building matters at least as much as the household.
- Household size cuts both ways. At 2.48 people per household, one connection is shared by two or three people on average. A single phone plan stretched across a family is a different thing from a line with a router that several people can use at once.
- Language shapes which services are easy to use. With nearly half of residents speaking another language at home, the city publishes much of its public information in many languages. Notify NYC, the emergency alert service, sends messages in languages including Spanish, Chinese, Russian and Haitian Creole as well as American Sign Language, per the city's own page. An online service is only as accessible as the language it is offered in.
- Income and cost. A subscription is a monthly bill. The survey records whether a household has one, not whether it is a strain to keep.
Why the borough sizes matter
Boroughs are counties, and their populations differ widely. The 2020 Census counted 1,472,654 people in the Bronx, 2,736,074 in Brooklyn, 1,694,251 in Manhattan, 2,405,464 in Queens and 495,747 on Staten Island. The Census Bureau's estimates for July 1, 2025 are 1,406,332, 2,653,963, 1,664,862, 2,358,182 and 501,290 in the same order.
The QuickFacts figure quoted here is for the city as a whole. A citywide percentage hides whatever differences exist between a high-rise block in Manhattan, a two-family house on Staten Island and a rental building in the Bronx. The Census Bureau publishes finer breakdowns by borough, by census tract and by community district for anyone who wants to look at a particular area, and the city's planning department publishes population data for neighborhood-sized areas. Because neighborhood boundaries are not official in New York, the planning department's Neighborhood Tabulation Areas, built from whole census tracts, are the usual unit. The department says the names are not meant to be definitive. That caution is worth keeping in mind when a map of "internet access by neighborhood" gives a crisp-looking answer.
Say a reader wants to know whether the city's overall figure applies where they live. The right approach is to look up the tract or the neighborhood tabulation area, check the margin of error that comes with it, and treat a single number for a small area as a range. A tract has only a few thousand residents, and a survey sample in a tract is small.
Where people connect without a home line
Households without a subscription still use the internet, and the city offers several public places to do it.
The city has three separate public library systems: the Brooklyn Public Library, the Queens Public Library and the New York Public Library, which serves Manhattan, the Bronx and Staten Island. Each runs its own branches and its own website, and each lists the computer and Wi-Fi services its branches offer. The library cards in a three-system city guide explains how to get a card at each system, since the three do not share one. For the larger picture of how the systems differ, see the three library systems compared.
NYC Parks lists "Wi-Fi hot spots" among the facilities at many park pages, including the Coney Island Beach and Boardwalk page and the Rockaway Beach page. The Parks site shows which parks have them.
For anything about a city service, 311 takes calls, app requests and online requests, so a household without internet can still reach an agency by phone. The line is 311 inside the city and (212) 639-9675 from outside; the video relay and TTY numbers are on the city's 311 portal. using 311 effectively walks through what to say. The 311 and city services record lists how the line is organized.
A worked reading of the table
A reader can turn the QuickFacts rows into a short list of questions the table can and cannot settle. Take the two access rows and the rows around them.
First, the table says that 94.7 percent of households have a computer of some kind and 90.5 percent have a broadband subscription of some kind. It does not say how many have both, because the two shares overlap and the page gives them separately. A household can have a phone with a data plan and no laptop, a laptop and no subscription, or neither.
Second, the table says that 47.7 percent of people age 5 and over speak a language other than English at home. It does not say how well those households can use English-language websites, and it does not say which languages. The city's multilingual alert service and its translated forms exist because the answer is not "everyone is fine."
Third, the table gives a median household income of $80,483 and a median gross rent of $1,821. A median is the middle household, so half of households have less. For a household near the bottom of the range, a monthly internet bill competes with rent, food and transit, and the fare for a subway ride is $3 as of 2026. The figures cannot say how many households are in that position, but they explain why public places to get online, from libraries to parks, matter.
Fourth, the table counts households, not people. A household of five with one working laptop is one "household with a computer." The survey does not ask how many people share it.
That is the discipline the table asks for: use it for the question it answers, which is how many homes have a connection of some kind, and look elsewhere for the questions it does not.
What the figures do and do not say about neighborhoods
Neighborhoods in the outer boroughs, such as those around Jamaica and Flushing, are home to large immigrant communities and to many households that rent. The 36.6 percent foreign-born figure is for the whole city, and so is the language figure. These are context for why multilingual outreach matters; they are not a claim that any one neighborhood has more or less access than another. Only a tract-level table can answer that for a given area.
A better question than "which neighborhood is most connected" is "what kind of connection do households here have, and what do they use it for?" The Census Bureau's survey is built to answer the first part (computer, broadband, or neither). The second part, what people do with the connection, is left to local surveys and to the people who run services.
Reading the numbers next year
The estimates are five-year figures, which means each one pools survey responses from 2020 through 2024. That has two effects. A pooled estimate is steadier than a single year's figure, which helps for a city with many small areas. It also lags behind what has changed most recently, so a new program, a price change or a building getting wired will not show up for a while.
When the Census Bureau releases the next five-year period, the QuickFacts page will carry new figures with a new year range. Compare like with like: a 2020-2024 estimate against a 2021-2025 estimate overlaps four of five years, so a small change between them is not much of a signal. A change that holds across several releases is.
Meanwhile the numbers on the table are a fair summary: a city where more than nine in ten households are connected, and where the remaining tenth is large enough, in a city of more than eight and a half million people, to fill a good-sized city of its own.