
When Wendyliz Martinez traveled from Pennsylvania to the Bronx to stay with her mom during the pandemic, she was looking for a way to balance childcare with her online graduate studies at Penn State.
Instead, she found that she was often interrupted by glitches during Zoom meetings, stuck refreshing the same websites and crowded into a corner of her mom’s one-bedroom apartment in the hope of getting a stronger Wi-Fi signal.
Martinez could’ve chalked up her mom’s spotty cable internet to a number of issues: a faulty router, the Wi-Fi setup or a slow speed tier, perhaps. But she suspected something else after her mom switched to a new router, only to find the speeds still lagging.
If old equipment wasn’t to blame and Martinez’s mom was getting the fastest speeds she could afford in an area without other internet provider options, were other households in the neighborhood experiencing the same thing? Martinez decided to turn that question into a research project.
“I was one of the fellows for Library Futures that year, and I was still feeling the impact of the COVID pandemic,” Martinez said, referring to a nonprofit organization dedicated to advocacy for libraries. “I was interested in exploring that for the fellowship.”
Using US Census Bureau data, Martinez spent months mapping the state of internet connectivity in the Bronx.
“I was interested in the mapping portion because I’m a visual person,” Martinez said. “I know my own kind of experience, and I have done research on Black studies, so I’m aware of historical implications and all that. But how does this actually look on a map?”
What she discovered confirmed her suspicion: Low-income neighborhoods in New York City had less access to high-speed internet than wealthier neighborhoods nearby.
In fact, the Bronx was one of the city’s worst-connected boroughs.
But was the correlation between low-income neighborhoods and lack of adequate internet access an indication of deeper systemic issues?
A lot of researchers seemed to think so.
The divide between households with access to high-speed internet and households without access is a nationwide issue. The sudden reliance on a decent internet connection during the pandemic revealed the extent of that divide, especially in education.
Following the pandemic, big policy changes, like the Affordable Connectivity Program, BEAD and myriad city and state-driven initiatives, seemed to pave the way for improved internet access.
In 2025, despite the abundance of studies pointing to the broadband divide in recent years and the progress from broadband divide initiatives, low-income households and communities of color are still behind when it comes to high-speed internet.
Americans need a decent internet connection for everything from applying for jobs to buying groceries, getting healthcare and staying plugged into their communities. Households with high-speed internet may take it for granted that within seconds, an entire world is at their fingertips, one not so easily accessed for homes where fast internet is still a luxury.
What is digital redlining?
In 2017, Bill Callahan, co-founder of the National Digital Inclusion Alliance and Connect Your Community, published a study that revealed that AT&T withheld fiber deployment in low-income neighborhoods throughout Cleveland.
The report called the practice “digital redlining.”
The term echoes historic redlining when banks denied mortgages to middle-income and upper-income Black families during the 1940s with lasting effects over decades Digital redlining occurs when telecom internet service providers deploy upgrades to networks in wealthier parts of cities but leave low-income neighborhoods behind.
As a result, low-income communities are often stuck on the middling speeds of legacy DSL networks — paying roughly the same monthly costs as wealthier neighborhoods across town that have fiber internet from the same ISP.
Unlike redlining, however, digital redlining isn’t an illegal practice. Secondly, while both digital redlining and mortgage redlining disproportionately affect people of color, digital redlining is perpetuated based on income or return on investment first.
In other words, since higher-income neighborhoods will assumedly sign up for upgraded fiber internet more quickly than low-income households, ISPs can earn back the money spent from deploying fiber internet to that area faster. Consequently, low-income communities are often the last to get faster internet, time and time again.
The discovery that some neighborhoods historically redlined by banks in Oakland, California, and Cleveland overlap with the same neighborhoods digitally redlined today indicates that these patterns of exclusion continually overlook communities of color.
“Probably the starkest example [of digital redlining] that I can think of is the Hope Village neighborhood in Detroit,” Sean Gonsalves, a reporter at the Institute for Local Self-Reliance, told me. “This is a low-income neighborhood in Detroit, and in the height of the pandemic — this is an area that AT&T serves — there was a six-week internet outage.”
Since the internet is not considered a utility akin to water and power, universal service is not required, and ISPs build networks in markets that yield the highest profits.
Whether ISPs intended to discriminate when making these upgrades to their networks is inconsequential, although determining “intention” makes for tricky legal territory.
“With an internet service provider…it really matters that the intent is not relevant,” said Angela Siefer, executive director of the National Digital Inclusion Alliance. “They may not have meant to discriminate, but the end result is that someone doesn’t have access to the same tech support or the same marketing and sales folks.”
Subsequent reports by the NDIA pointed to instances of digital redlining in Dayton, Ohio, Detroit and Toledo, Ohio, among other cities.
Studies from other researchers followed suit with similar findings across the country, including in Los Angeles and Philadelphia.
In 2022, The Markup and the Associated Press published a study with evidence of digital redlining across 38 cities.
How infrastructure was meant to fix digital redlining
Following the pandemic, the federal government unveiled new legislation to make the internet more accessible and affordable.
The 2021 Infrastructure Investment and Jobs Act allocated $90 billion to improve internet access, creating the Broadband Equity Access and Deployment Program, or BEAD, and the Affordable Connectivity Program, or ACP. Both programs were instrumental in heralding a new era of improved internet connectivity.
BEAD is the largest federal investment in broadband to date, with $42.45 billion in funding set aside for deploying internet. The ACP provided a significant $30 to $75 monthly discount on internet bills for qualifying households.
However, the IIJA’s clauses on eliminating digital discrimination were possibly the most significant federal structure for stopping digital redlining, a task expressly delegated to the Federal Communications Commission.
In response to the IIJA’s section on digital discrimination, the FCC established the Digital Discrimination Task Force in 2022, but the venture was short-lived and dissolved in 2025 as an instance of “Diversity, Equity and Inclusion” (DEI). The organizational framework, meant to promote fair treatment and participation, came under fire as a discriminatory program in early January. All DEI-related web pages have since been scrubbed from the Commission’s website.
Similar legislative efforts face uncertain futures too.
BEAD is on track to approve proposals and dole out its funding in the coming years. Although structural changes to BEAD have not yet happened, the newly appointed program leader, Arielle Roth, described the previous administration’s preference for fiber internet as part of a “woke social agenda,” echoing other GOP leaders seeking to change the program’s preference for fiber and potentially migrate to a Starlink satellite internet-favored approach instead.
The ACP, which ran out of funds in May 2024, has not been renewed — leaving the 23 million Americans who depended on it without a funding alternative and often resorting to cutting food and health costs from their budgets to afford monthly internet bills.
The legal battles surrounding equitable internet access have swung back and forth over the years in a similar fashion. But as our world becomes increasingly dependent on the internet, accessibility remains a widespread problem.
What’s the current state of digital redlining?
One-third of Americans have access to only one or no internet provider, and six ISPs cover 98% of the internet market. Like Martinez and her mom, the lack of ISP options means you’re either stuck with whatever speeds you can get and that ISPs can increase your bill after a year — and you can’t do anything about it.
The trade group USTelecom found that broadband prices decreased while the costs of overall consumer goods and services increased since 2015. Even so, the average cost of the internet is still high, especially in more populous states: After initial promotional discounts expire, the average monthly fee is $98, according to US News.
Those expensive costs and the lack of ISP options in low-income areas have created a fragmented landscape of internet connectivity. According to a Pew Research Center study, households with lower incomes are less likely to have an internet subscription than households with higher incomes.
As part of the Digital Equity Act, each state has its own Digital Equity Plan with various strategies and proposals for bridging the broadband divide. Each state has offered slightly different plans, though each emphasizes community-centric solutions. As such, the responses from cities grappling with digital redlining have all been unique.
According to data from Connect Your Community, Detroit and Cleveland were the worst-connected US cities with over 100,000 households in 2023.
Although Detroit and Cleveland have made significant progress in internet connectivity in recent years, gains may stall without continued federal support. Callahan credits that success to the effectiveness of federal programs like the now-ended Emergency Broadband Benefit and the ACP, as well as locally led efforts in the cities.
“Cleveland has had more people working on this problem in the community for the last 25 years than in almost any city in the country,” Callahan said.
Cleveland helped establish an unlicensed fixed wireless network, DigitalC, which costs $18 monthly for speeds “guaranteed” at a minimum of 100 megabits per second, the new baseline for broadband that the FCC raised last year from 25Mbps. DigitalC received $2.76 million as a part of its four-year contract with the city, though the Cleveland City Council expressed concern over DigitalC’s failure to meet its subscriber goal for 2024.
Following Callahan’s report in 2017, multiple Cleveland residents filed claims of digital redlining by AT&T with the FCC. As of early 2018, those claims were resolved and dismissed. AT&T responded directly to the allegations at the time, stating, “We do not redline,” and highlighted the company’s diversity and inclusion efforts. The telecom outlined to CNET what it’s done in the years since.
“In the Cleveland area, we invested more than $550 million in our network infrastructure from 2019 to 2023 to connect more people to greater possibility,” AT&T said to CNET in a statement.
The statement pointed to AT&T’s partnership with JobsOhio to build out a fiber network through East Cleveland, affecting 14,000 residents and slated to be finished by the end of 2025.
Still, data from Connect Your Community as recent as June 2024 suggests that most nonwhite households in Cleveland rely on asymmetric DSL internet services, which use copper telephone lines for data with speeds averaging lower than 25Mbps. For context, that’s probably not enough speed to stream high-quality video if there’s more than one internet user in the house.
Cleveland Census Tracts by Median Household Income
ALL AT&T BSLs | FIBER | VDSL (25-100Mbps) | ADSL (<25Mbps) | |
---|---|---|---|---|
Under $35,000 | 36,338 | 13,479 | 3,497 | 19,362 |
$35,000 – $49,999 | 47,916 | 22,515 | 11,373 | 14,028 |
$50,000 – $74,999 | 25,528 | 16,184 | 5,071 | 4,272 |
$75,000 and up | 6,237 | 4,934 | 1,153 | 150 |