Recently there have been several pieces published that argue, in light of price and quality trends, that broadband service is affordable, and that, whether or not affordability matters, those who do not have service at home simply do not want it. These affordability arguments look at the Federal Communications Commission’s Urban Rate Survey (URS) to claim that broadband prices in real terms have been falling. The “they don’t want service” argument rests on a survey question that the National Telecommunications and Information Administration (NTIA) has used for a number of years, which Recon Analytics recently examined in depth.[1] Each argument warrants scrutiny. The more important question, though, is this: Why are we seeing these arguments now?

What? People Don’t Want Broadband?

The Recon Analytics analysis rests, as noted, on a question in the NTIA Internet Use survey that asks those without broadband why they do not have service at home. That question finds that most (55.7%) of those without service cite “don’t need or not interested” as the reason they lack service at home.[2] The share of respondents saying this has grown over time, according to Recon Analytics. The analysis adds that “[t]hrough the entire [Affordable Connectivity Program] era, the largest connectivity subsidy in American history, the don’t-need share rose.”

The notion that those without broadband simply don’t want it has been around for some time. And so has pushback to it that rests on a significant body of research—quantitative and qualitative. The most important point in understanding why people do not have broadband is to recognize that there are multiple reasons behind it—not a single one. Results from a survey conducted by EveryoneOn in 2021 show how careful questioning offers insight into the non-adoption issue.[3] The EveryoneOn survey found that the typical non-adopter cited three reasons for not having service when offered a list of ten from which to choose. The top five were:

• 55% of non-adopters said service cost was the reason they lacked service at home
• 49% said they did not want or need service
• 47% cited the cost of a computer
• 42% said they were worried about the privacy or security of their personal data
• 38% said their smartphone was sufficient for their needs.

When asked, as a follow-up, to specify the most important reason for having service:

• 21% said it was the cost of service
• 15% cited the smartphone
• 15% said they did not want or need service
• 11% the cost of a computer
• 8% said it was worry about the privacy or security of their personal data

Among the 49% of those who said they did not want or need service, two-thirds of them (67%) cited cost of service as an additional reason they did not have service. Qualitative research adds further evidence that non-adopters who (in a survey) cite “don’t want or need” as a reason also feel the weight of cost. Canadian researchers in 2016 found that “for a significant portion of respondents, their lack of interest in using the Internet was rooted in the cost of the service, their lack of digital literacy, or their inability to obtain an affordable computer.”[4] Interviews with non-adopters showed that they clearly understood the value of having service, but had difficulty affording it.

In insisting that non-adopters simply do not want service, Recon Analytics relies on trend analysis of the NTIA question. The analysis points to the increase over time in the share of those citing “don’t want or not interested” as evidence. But this analysis suffers from a logical problem. Since the share of non-users has declined substantially over a long period of time, the composition of the respondent set of non-adopters has changed a great deal. The question may be apples-to-apples, but the sets of respondents receiving it at two points in time are not. The comparison set at Time 2 lacks the types of users who, since Time 1, have gotten service. Something changed for them. It could be as simple as evolving preferences (those cat videos started to look better!). Or, as Recon hints at in citing the Affordable Connectivity Program era, perhaps service became more affordable because of that subsidy.

For better or worse, the NTIA question will continue to be cited in the way Recon Analytics does in asking policymakers to “respect the dignity of the household that has weighed the offer and said no.”[sic] Yet it seems irresponsible to base a policy prescription on a single survey question when a body of other research points in the opposite direction.

Broadband Prices and Speeds

As I wrote earlier this year, an examination of URS price trends shows that, in the aggregate, at-home broadband prices have risen only modestly in recent years—by 4.8% in real terms from 2024 to 2025.[5] To the benefit of lower-income households, prices in lower-cost, lower speed tier categories have fallen by 13.4% in real terms between 2024 and 2025. To the detriment of lower-income households, the URS shows that there are far fewer such lower-price offerings in the market today than a few years ago. The aggregate increase in broadband price is the result of the growth of higher speed (and more expensive) offerings.

The Phoenix Center’s piece in response to my analysis went to great lengths (using a methodology that differs from how the FCC publishes URS data) to produce results that are … really not that different from mine.[6] Phoenix found, using the FCC’s method, that broadband prices have fallen by 1.6% compared to my 4.8% increase; the figure differs due to the exclusion of higher speed tier plans not offered in 2024. The Phoenix paper argues that including speed tiers in 2025 not offered in 2024 results in a biased estimate. Further analysis in the Phoenix paper puts aggregate price decline at 9% between 2024 and 2025.

But not so fast on the point about a biased sample! My analysis aggregates high-end speed tiers as those exceeding 2 Mbps, a tier evident in URS data since 2021, but whose incidence did increase by 65% in the URS sample between 2024 and 2025. Additionally, it is not clear why offerings with new features should be excluded from price comparisons. As Clifford Winston recently noted in analyzing the affordability of automobiles, the average cost of the least expensive cars in the market has doubled in real terms since 1973, largely because of new features (mostly electronics) in today’s cars.[7] Why exclude cereals with more vitamins from analysis of price trends for breakfast cereals?

Although the monthly cost of service is an important part of evaluating the decision to purchase broadband, other service attributes, such as speed, are also relevant. As interviews with low-income broadband subscribers show, some purchase service plans with higher speeds because they have no other choice and low-cost plans are not available where they live.[8] Higher speed (and higher priced) plans may be the only ones available that can carry out necessary at-home tasks (e.g., distance education).

A second Phoenix piece explores the issue of speed and how it has changed over time. That paper finds that “price per bit” for broadband has fallen substantially, given that speeds offered in the market double every 2.6 years and prices have not risen at that rate. Phoenix finds that  “consumers in 2026 pay roughly the same for 1,000 Mbps as consumers in 2020 paid for 100 Mbps.”[9]

This is both true and, in the context of broadband affordability for low-income households, not very meaningful. Yes, broadband speeds have increased significantly in recent years and faster than prices have increased. This can be characterized as “more bit per buck.” USTelecom says that the average “bit per buck” in 2025 was $0.07 per megabit, equivalent to $14 per month for 200 Mbps service.[10] That service offering would likely be attractive to many lower-income households. The trouble is that the URS does not show evidence of many such plans in the marketplace. To take one example, the Xfinity Internet Essentials low-cost plan is $14.95 per month for 75 Mbps service—or $0.20 per megabit. On the whole, the number of low-cost service plans in the market has fallen in recent years. These sorts of “cost per megabit” exercises are of interest to a point, but of little practical relevance for the lower-income segment of the market.

What’s Really Going On?

Tempting as it may be to wade into the weeds of the back-and-forth on affordability, the big picture is clearer: There is likely to soon be a decline in broadband subscriptions because of the impacts of the One Big Beautiful Bill (OBBB) on household finances. To turn attention away from that, it might be useful to have research that demonstrates that affordability is not a problem for low-income households. Research showing that some households simply prefer not to have service is a bonus. A cynic might call it gaslighting.

Let’s look at OBBB’s impacts. My work has shown that the loss of benefits built into the OBBB and associated decline in spending power for low-income households could result in the loss of 5.6 million broadband connections.[11] These losses might take time to unfold in the marketplace, but we are starting to see stories about OBBB’s impacts. The legislation has contributed to spikes in health insurance rates due to cutting Affordable Care Act premium subsidies and reductions in the number of households using Supplemental Nutrition Assistance Program (SNAP) benefits. Specifically:

• SNAP cuts: Center on Budget Policy and Priorities estimates that OBBB resulted in a fall in SNAP participation of 4 million people between July 2025 and March 2026. Nearly half of those losing benefits are children.[12]
• In Arizona, SNAP enrollment has “plummeted by half” as 500,000 people have lost benefits, with administrative problems driven by cuts in the state government’s workforce contributing to the steep decline.[13]
• Fewer people with health insurance: A recent Department of Health and Human Services report found that ACA enrollment declined by 3 million – from 22.1 million to 19.2 million people between February 2025 and February 2026.[14] The Kaiser Family Foundation (KFF) projects that enrollment could fall to as low as 16.3 million people in 2026.[15]
• Higher Affordable Care Act premiums: KFF estimates that ACA premiums have risen by 58% for enrollees (from $113 per month and to $178) as enhanced premium tax credits expired.[16]

Some households will economize on connectivity in order to feed the kids, have health insurance, or pay for medical bills out of pocket. The $65 average monthly ACA premium increase is about the cost of an entry level broadband service plan. Doing the math using the Phoenix pricing analysis, this means a 9% drop in a $65 broadband bill leaves a household facing an ACA premium hike paying $59.15 per month for broadband. That only modestly softens the additional $65 out-of-pocket because of OBBB’s impact on premiums. And some of these households will also suffer additionally from SNAP cuts.

When elected officials—undoubtedly many who voted for the OBBB—hear the complaints about kids doing homework at McDonald’s, they might turn to Internet service providers (ISPs) to ask why. The answer will be that service became unaffordable for low-income households as food and health care insurance became more costly. For many of these households, the broadband sticker price was a challenge before OBBB cuts.[17] Whether overall price rose a few bucks in the past year (as my URS analysis shows) or fell a few bucks (as the Phoenix analysis shows) is not that relevant to their calculation. Nor will the typical plan’s download speed factor into choices they may have to make about service. And none of the households that give up a broadband connection will have suddenly decided they do not want or need it.

But research of the sort produced by Phoenix or Recon Analytics gives interested parties a hook for developing a response to looming declines in connectivity. If an emerging problem is no problem at all, then policy has no role. Prices are steady. Respect consumers’ choices. Take no action. “Hands off policymakers” is the desired goal for this research.

Trying to define a problem away does not make a real problem vanish. OBBB is very likely to cause many low-income households to pare back on connectivity. Carriers, and other stakeholders, would do better to get ahead of this emerging issue by looking for solutions, not waving their hands past the problem.

The Benton Institute for Broadband & Society is a non-profit organization dedicated to ensuring that all people in the U.S. have access to competitive, High-Performance Broadband regardless of where they live or who they are. We believe communication policy - rooted in the values of access, equity, and diversity - has the power to deliver new opportunities and strengthen communities.

© Benton Institute for Broadband & Society 2026. Redistribution of this email publication - both internally and externally - is encouraged if it includes this copyright statement.

For subscribe/unsubscribe info, please email headlinesATbentonDOTorg

Kevin Taglang
Executive Editor, Communications-related Headlines
Benton Institute
for Broadband & Society
1041 Ridge Rd, Unit 214
Wilmette, IL 60091
847-220-4531
headlines AT benton DOT org

Share this edition:

Broadband Delivers Opportunities and Strengthens Communities