The 2026 AP-NORC/EPIC Energy Survey is an annual survey of U.S. adults conducted by the Associated Press-NORC Center for Public Affairs Research with funding from the Energy Policy Institute at the University of Chicago (EPIC). Staff from NORC at the University of Chicago, the Associated Press and EPIC collaborated on the study. Results were released on September 16, 2026 as a nine-page set of factsheets, Findings from a 2026 Survey of Adults Age 18 and Older, a three-page press release datelined Chicago, and a 34-page topline with full question wording and methodology. The survey's long-running items cover climate change, extreme weather, energy policy and electric vehicles; the 2026 edition added questions on the environmental impact of artificial intelligence and on data centers in the communities where they are built.
The release's summary finding is that concern about AI's environmental consequences rose over the year, most among younger adults, and that most Americans are concerned about data centers' effects on local electricity prices, water supply and power outages while supporting policies that would regulate them.
Method
Interviews were conducted July 6–24, 2026 with 3,424 adults age 18 and older in the 50 states and the District of Columbia. The sample combines two sources. The probability source is NORC's AmeriSpeak panel, recruited from the NORC National Sample Frame and covering approximately 97% of the U.S. household population; 1,430 panel members completed the survey (1,402 by web, 28 by telephone), with a weighted household panel recruitment rate of 24.9%, a survey completion rate of 47.6% and a weighted cumulative response rate of 11.9%. The non-probability source was 1,994 web interviews supplied by Forthright on quotas for age, race and ethnicity, gender and education; the methodology notes that standard sampling error and response rates cannot be calculated for a non-probability panel. Seventy interviews were removed before weighting for nonresponse to at least half the questions or for completion in under one-third of the median time.
The non-probability sample was incorporated using NORC's TrueNorth calibration, which the methodology describes as a tree-based supervised learning approach that adjusts the non-probability weights toward population distributions on characteristics correlated with the survey variables. Both samples were then raked on age, gender, census region, race and ethnicity, education and 2024 vote choice, using the 2026 Current Population Survey and official 2024 turnout and vote results. The overall margin of sampling error is ±2.2 percentage points at the 95% confidence level, including the design effect. Interviews were conducted in English. In the weighted sample, 40% identify with or lean toward the Democratic Party, 36% the Republican Party, and 23% are independents or identify with no party and do not lean toward either.
AI and the environment
Asked "How concerned are you about the environmental impacts of each of the following industries?", 53% were extremely or very concerned about artificial intelligence, up from 41% in September 2025. Concern about the other three industries asked about was similar to 2025.
| Industry (CCAI1) | 2026 extremely/very | 2026 somewhat | 2026 not very/not at all | 2025 extremely/very |
|---|---|---|---|---|
| Artificial intelligence | 53% | 25% | 22% | 41% |
| Meat production | 32% | 32% | 35% | 29% |
| Cryptocurrencies such as Bitcoin | 28% | 27% | 44% | 29% |
| Air travel | 28% | 34% | 37% | 23% |
Democrats were more likely than Republicans to be extremely or very concerned about the environmental impact of AI (65% against 42%), cryptocurrency (40% against 18%), meat production (42% against 21%) and air travel (38% against 17%). Democratic concern about AI rose from 49% in 2025. The increase was largest among adults aged 18–29, who moved from levels in line with other age groups in 2025 to the most concerned in 2026 (38% to 60%). The press release describes AI as "the only industry to see a large increase in concern over the past year."
Expected effects of AI
The survey asked: "In the next 10 years, do you think Artificial Intelligence, or AI, will do more to help or more to hurt each of the following, or will it not make much difference?"
| Domain (CCAI2) | Year | More to hurt | More to help | Not much difference | Not sure |
|---|---|---|---|---|---|
| Society as a whole | 2026 | 52% | 21% | 11% | 16% |
| 2025 | 44% | 22% | 13% | 20% | |
| The U.S. economy | 2026 | 43% | 22% | 15% | 20% |
| 2025 | 36% | 22% | 18% | 24% | |
| The environment | 2026 | 47% | 15% | 19% | 18% |
| 2025 | 33% | 18% | 25% | 24% | |
| You personally | 2026 | 29% | 22% | 30% | 18% |
| 2025 | 26% | 22% | 31% | 20% |
Democrats were more likely than Republicans to say AI will do more to hurt than help the environment (57% against 40%), the U.S. economy (49% against 40%) and themselves personally (34% against 26%), and society as a whole (57% against 50%), where half of Republicans shared the view. Democrats' pessimism rose from 2025 on the environment (37% to 57%), society (46% to 57%) and the economy (38% to 49%). Among adults 18–29, the share expecting AI to do more to hurt the environment rose from 39% to 57%.
Data centers
The data-center questions opened with a definition: "data centers are buildings that house computer servers and equipment used to process information online. They support a wide range of internet services, including artificial intelligence (AI), online banking, and cloud storage." Respondents were then asked how concerned they were "about the impacts of data centers in the communities where they are built."
| Local impact (DC1) | Extremely/very concerned | Somewhat | Not very/not at all | Not sure |
|---|---|---|---|---|
| Electricity prices | 63% | 21% | 9% | 7% |
| Water supply | 57% | 21% | 13% | 8% |
| Power outages | 53% | 25% | 13% | 9% |
| Noise | 39% | 24% | 25% | 11% |
The press release gives the shares at least somewhat concerned as 84% for electricity prices, 79% for water supply and 78% for power outages; the rounded components for water supply sum to 78%. Democrats were more likely than Republicans to be extremely or very concerned about electricity prices (70% against 58%), water supply (67% against 49%), power outages (61% against 46%) and noise (45% against 33%). Residents of the Northeast were less concerned than residents of the South about electricity prices (58% against 66%) and water supply (50% against 60%), and residents of the West less concerned than residents of the South about electricity prices (60% against 66%) and power outages (50% against 56%). Concern did not differ among urban, suburban and rural residents, which the factsheets note although more data centers are being planned in rural areas, citing a Pew Research Center analysis of April 13, 2026.
Expected benefits were more mixed. About half of adults said data centers are at least somewhat beneficial in the communities where they are built for technological advancement (55%) and job creation (49%), and fewer for improved infrastructure such as expanded grids and additional generation (44%) and tax revenue (43%). The extremely-or-very-beneficial shares were 24%, 20%, 18% and 16%. Republicans were more positive than Democrats on technological advancement (60% against 54%), job creation (55% against 46%) and infrastructure (51% against 41%); education and income made no significant difference.
| Data-center policy (DC3) | Overall support | Strongly support | Oppose | Democrats | Republicans |
|---|---|---|---|---|---|
| Requiring companies that build data centers to pay for the electric grid upgrades needed to support them | 73% | 56% | 7% | 79% | 76% |
| Requiring data centers to use clean energy sources, such as wind or solar, to supply their electricity needs | 65% | 44% | 9% | 80% | 56% |
| Limiting the number of new data centers that can be built | 62% | 41% | 8% | 70% | 57% |
| Restricting the construction of new natural gas power plants to supply electricity for data centers | 41% | 23% | 20% | 50% | 33% |
Adults 45 and older were more supportive than younger adults of the grid-upgrade requirement (80% against 66%) and the clean-energy requirement (68% against 63%). College graduates were more supportive than adults without a degree of the grid-upgrade requirement (83% against 68%), the clean-energy requirement (73% against 61%) and restricting new gas plants (47% against 37%). Independents are not broken out in the factsheet tables.
Climate, energy and vehicles
The survey's trend items were largely stable. Seventy-four percent said climate change is happening (the series has ranged from 71% to 78% since 2016), and among them 54% said it is caused entirely or mostly by human activity. Democrats were more likely than Republicans to say climate change is happening (92% against 55%) and mostly human-caused (68% against 34%). Ninety-four percent reported experiencing extreme weather in their community in the past year, most often unusually hot days (80%) or poor air quality (58%). Thirty-four percent had experienced an increase in homeowners or renters insurance premiums or deductibles as a result of extreme weather, ranging from 22% in the Northeast to 39% in the West, and 24% had considered moving to avoid extreme weather.
Energy policy (53%) and climate change policy (47%) were rated extremely or very important by fewer adults than the economy (81%) and health care (81%), and the factsheets describe a 49-point partisan divide on climate policy (73% of Democrats against 23% of Republicans in the rounded table). Support was highest for protecting conservation lands and wildlife (81%), enforcing strict clean air and water regulations (69%), incentives for new renewable energy projects (63%) and taxing companies on the carbon they emit (62%), and lowest for rolling back environmental regulations (23%) and auctioning more public land for oil drilling (19%); expanding U.S. oil and gas production drew 40%. Support for a monthly household carbon fee fell from 38% at $1 to 17% at $100. EPIC director Michael Greenstone said the average respondent was willing to pay about $29 a month to combat climate change, "the lowest amount since we started asking this question a decade ago, once adjusted for inflation."
Eight percent of respondents lived in a household that owns or leases an electric vehicle and 40% said they were at least somewhat likely to buy one. Offered otherwise similar EVs, 63% would pay $500 more for a U.S.-made vehicle over a Chinese-made one, while at a $10,000 difference 54% would choose the cheaper Chinese-made vehicle; preferences for the cheaper Chinese vehicle rose from 2024. Fifty-one percent had cut back on driving because of high gas prices in the preceding months.
Statements on release
Jennifer Benz, director of the AP-NORC Center, said Americans "are concerned about the infrastructure that supports" AI and that "many Americans support policies intended to address the demands data centers place on energy and water resources." Greenstone said that "any durable climate policy must balance emissions reductions with affordability." The Associated Press reported the poll under the headline "There's one issue where Republicans and Democrats are unified before the midterms, a new poll finds" (Source: ap.org).
Figures that differ across documents
- Field dates. The press release, the topline and the Associated Press story give July 6–24, 2026; the AP-NORC project page gives July 7–24 (Source: apnorc.org).
- Water supply. The press release gives 79% at least somewhat concerned; the factsheet components (57% and 21%) sum to 78%.
- Sample description. The release calls the second source a non-probability sample; the Associated Press story describes it as interviews from "opt-in online panels."
Relation to other surveys
The survey is one of several 2026 instruments measuring local data-center sentiment. The Annenberg Public Policy Center's W38 wave recorded opposition to new data centers in the respondent's own area rising from 48.9% in February–March 2026 to 60.6% in June–July 2026 (Artificial Intelligence — Topline Report (APPC W38 National Survey)), and the Public First polling presented by ITIF recorded net support for a local data center falling from +9 in January 2026 to −17 in July 2026, with higher electricity bills the most-cited drawback (Public First / ITIF, \"The opposition to data centers\" (polling deck, August 2026)). Those surveys measure support or opposition to local construction; the AP-NORC/EPIC survey measures concern about specific impacts and support for specific regulatory requirements, so the figures are not directly comparable. The survey's partisan pattern, with Democrats more concerned about data centers and Republicans more positive about their benefits, is consistent with the partisan differences on energy production recorded by American AI Sentiment in 2026 (Gallup and SCSP).
Provenance
The factsheets and press release were captured on September 17, 2026 from apnorc.org, the AP-NORC Center's own domain (Sources: apnorc.org; apnorc.org). The topline was captured the same day from epic.uchicago.edu (Source: epic.uchicago.edu); some question stems and row labels were lost in parsing, including those of the issue-importance grid and the extreme-weather grid, so figures above are taken from labelled rows or from the factsheets. The two captures are held as separate raw files. The factsheets carry the partisan, age, education and regional splits, which the topline does not. The release compares its AI and data-center items against the 2025 AP-NORC/EPIC survey (fielded September 2–18, 2025, 3,154 adults), which is not held in the vault's raw sources.
Relationships
- supports: Public Opinion on AI — environmental concern, expected effects of AI, and data-center attitudes by party.
- supports: Data Center Siting / AI Power Politics, AI Data Centers — local concern about electricity prices, water and outages; support for developer-pays grid requirements.
- supports: AI Environmental Impact — concern about AI's environmental impact rose from 41% to 53% in a year.
- supports: Ratepayer Protection Act (H.R. 9340), Energy and Electric Power Sector — public support for requiring data-center developers to pay for grid upgrades.
- related: AI Labor Disruption — 43% expect AI to do more to hurt than help the U.S. economy.
- related: Artificial Intelligence — Topline Report (APPC W38 National Survey), Public First / ITIF, \"The opposition to data centers\" (polling deck, August 2026), American AI Sentiment in 2026 (Gallup and SCSP) — other 2026 surveys of data-center and energy attitudes.
- related: AP-NORC Center for Public Affairs Research, Energy Policy Institute at the University of Chicago (EPIC) — producer and funder.
- related: AI Energy Management Alliance (AEMA) — the industry coalition launched the same day, whose coverage cited the poll.