Americans are growing more skeptical of data centers and fewer report owning cryptocurrency, while a separate international study finds widespread optimism about artificial intelligence. The findings arrive as New York restricts permitting for certain large data centers, develops community investment guidelines and supports AI research involving Cornell University.
A Pew Research Center survey finds increasingly negative views of data centers’ environmental and community effects. Meanwhile, Gallup’s cryptocurrency survey shows ownership falling from its 2025 high, and Gallup’s research with Microsoft finds that expectations for AI often exceed trust in its accuracy.
The studies examine different populations and different technologies. Together, they document several distinct concerns confronting the digital economy: what people expect technology to accomplish, whether they trust it and what they believe its infrastructure will cost their communities.
For the Finger Lakes, those questions have concrete connections to Cornell’s research and Greenidge Generation’s cryptocurrency and power generation facility in Dresden. Statewide, they overlap with decisions about electricity demand, environmental permitting and the benefits communities should negotiate before hosting large computing facilities.
More Americans see disadvantages to data centers
Pew’s survey, released Sept. 22, shows a substantial shift since January in Americans’ assessments of data centers.
The share saying data centers are mostly bad for the environment increased from 39% to 54%. Half now say they are mostly bad for home energy costs, up from 38%, while 49% say they are mostly bad for the quality of life of nearby residents, up from 30%.
Only 4% describe data centers as mostly good in each of those three areas.
Those findings measure public perceptions. They do not establish how much a particular facility changes electricity bills, water use or neighborhood conditions.
Views of economic benefits are more divided. Twenty-four percent say data centers are mostly bad for local jobs, up from 15% in January, while 22% say they are mostly good. On local tax revenue, 21% say mostly bad, up from 12%, and 22% say mostly good.
Many respondents remain uncertain: 23% are unsure about the employment effects and 32% are unsure about the tax revenue effects.
The growing skepticism also extends to whether Americans would welcome a facility nearby. Sixty percent say they would be not too comfortable or not at all comfortable with a new data center in their community. Nineteen percent would be somewhat comfortable, and 7% would be extremely or very comfortable. Another 12% have not heard of data centers.
Discomfort crosses party and geographic lines. It reaches 68% among Democrats and Democratic-leaning independents and 54% among Republicans and Republican-leaning independents.
By community type, 63% of suburban residents, 60% of rural residents and 57% of urban residents express discomfort.
The rural findings are relevant to the kinds of siting questions smaller communities can face, although they are national results, not estimates for the Finger Lakes. Among rural Americans, the share describing data centers as mostly bad for nearby residents’ quality of life rose from 28% to 45%. Negative views of environmental effects increased from 32% to 50%, and negative views of home energy costs rose from 32% to 46%.
Public awareness increased during the same period. Thirty-five percent now say they have heard a lot about data centers, compared with 25% in January. The share that has heard nothing fell from 25% to 12%.
New York pauses certain permits and outlines community benefits
New York has already begun addressing several of the concerns Pew measured.
Gov. Kathy Hochul’s July 14 executive order directs the Department of Environmental Conservation to hold certain discretionary permit applications for data center construction or expansion while the state completes an environmental review. The pause applies to applications that DEC had not determined were complete before the order.
The order covers qualifying facilities capable of consuming at least 50 megawatts of power. It excludes facilities primarily used for manufacturing, research, education or medical care, including qualifying academic research facilities and Empire AI. The permitting provision does not apply to local government approvals.
The Department of Public Service must assess potential effects involving energy demand, water use and quality, air quality, noise and disadvantaged communities through a process that includes public comment and a hearing.
On Sept. 15, the state released a voluntary community investment framework to help municipalities negotiate with developers.
Its recommended benchmark is $1 million in community investment for each megawatt of utility demand. A 50-megawatt project, for example, would have a recommended investment level of $50 million. That is negotiating guidance, rather than an automatic statewide fee.
The framework identifies potential investments in infrastructure, housing, schools, child care and workforce training. It also recommends commitments addressing water consumption, noise, lighting and local hiring, along with annual reporting on employment, utility use and community contributions.
The guidance does not replace environmental review or permitting. It gives local governments a way to evaluate promised benefits alongside a project’s demands on public services and infrastructure.
Global AI optimism comes with limits on trust
The international Gallup-Microsoft study presents a different picture from the increasingly negative U.S. views of data centers.
In 34 of the first 37 countries and territories surveyed, more AI-aware respondents reported at least one positive emotion about AI than at least one negative emotion. The exceptions were the United States, Egypt and the State of Palestine.
Positive and negative feelings can coexist. A respondent could be curious about AI while also feeling worried.
In 25 of the 37 countries, a majority of respondents who had heard of AI expected it to make their daily lives better. Optimists outnumbered pessimists on that question in 32 countries.
Expectations for broader national benefits were somewhat less positive. Respondents expecting AI to mostly help their country outnumbered those expecting harm in 29 countries, but formed a majority in only 18.
Among AI-aware respondents, 93% in China and 91% in Vietnam expected AI to mostly help their country. The U.S. figure was 36%, among the lowest in the study.
Optimism frequently exceeded firsthand experience. In Nigeria, 68% of AI-aware respondents expected AI to help their country, while 20% of the overall surveyed population had ever used it.
Across the 37 countries, the median share that had heard of AI was 81%, and the median share that had ever used it was 43%. Those are country medians, not estimates that 43% of the world’s population has used AI.
Adoption varied sharply. In Singapore, 79% had used AI and 46% used it daily. In Malawi, 3% had used it and 1% used it daily.
Trust was more limited than optimism. The median share of AI-aware respondents trusting AI completely or a lot to provide accurate information was 36%. Only four countries had majorities expressing that level of trust: Israel, Vietnam, China and Nigeria.
In 35 of the 37 countries, the share expecting AI to help their country exceeded the share expressing strong trust in its accuracy.
The United States stood out for worry. Seventy-four percent of AI-aware Americans said AI made them feel worried, the highest figure among the surveyed countries.
Frequent users were more trusting, but far from universally confident. In the United States, 45% of daily users trusted AI completely or a lot to provide accurate information, compared with 26% of weekly users, 15% of less-than-weekly users and 8% of people who had never used it.
The survey cannot determine whether using AI builds trust, whether people who already trust it use it more often, or whether both occur.
Cornell connects AI’s potential to regional research
The Finger Lakes has a direct connection to the applications that underpin hopes for AI.
According to Cornell’s AI research initiative, researchers are developing and using the technology in sustainable agriculture, nutrition, materials science and precision medicine.
Cornell is also a founding member of Empire AI, a consortium of 10 New York institutions supported by more than $400 million in public and private investment. The consortium created a shared computing facility in upstate New York to support research and development focused on public benefits.
These efforts illustrate why public expectations about AI and opinions about commercial data centers need to be examined separately. A person can see value in scientific applications while questioning a proposed facility’s electricity demand or neighborhood effects.
The surveys did not ask respondents to evaluate Cornell or Empire AI. Their findings therefore cannot establish public support for either initiative.
Cryptocurrency ownership falls as investors remain cautious
Gallup’s cryptocurrency findings concern financial participation rather than the use of AI or attitudes toward infrastructure.
Its June survey found that 9% of U.S. adults owned cryptocurrency, down from 14% in 2025.
Among investors — defined as people with at least $10,000 in investable assets — ownership fell from 17% to 11%. That remained above the 6% recorded in 2021.
Interest among nonowners did not substantially expand as ownership fell. Nineteen percent of investors were intrigued by cryptocurrency or planned to buy it, compared with 18% last year. Two-thirds, 66%, said they had no interest in buying it.
Most investors continued to regard cryptocurrency as risky: 63% called it very risky and 31% somewhat risky.
Ownership did not mean an absence of concern. Nearly half of current owners, 47%, described cryptocurrency as very risky. That compared with 37% of nonowners who were intrigued by it and 68% of investors who were uninterested.
Younger male investors remained the most likely age-and-gender group to own cryptocurrency. Among men ages 18 to 49, ownership was 24%, down from 33% in 2025.
Ownership was also higher among upper-income investors, at 15%, than among middle-income investors, at 7%, or lower-income investors, at 4%.
The results establish a decline in reported ownership. They do not show that cryptocurrency mining activity, electricity consumption or the number of mining facilities declined.
Dresden links cryptocurrency to energy and emissions decisions
In the Finger Lakes, cryptocurrency’s connection to physical infrastructure is visible at Greenidge Generation’s Dresden facility.
Greenidge describes itself in a company announcement filed with the Securities and Exchange Commission as a cryptocurrency data center and power generation company. It says the Dresden operation supplies electricity to the grid while using power generated at the facility for cryptocurrency mining.
Its air permitting dispute also shows why developments at an individual facility require separate scrutiny from national ownership trends.
In a Nov. 8, 2025, statement, DEC announced an agreement resolving pending litigation and establishing emissions reductions to be incorporated into a renewed air permit.
DEC said Greenidge would be required to reduce actual greenhouse gas emissions by at least 25% and potential emissions by 44% over the renewed permit’s term. The agency said the permit modification would undergo public review and comment.
Those figures describe requirements under the agreement, rather than proof that the reductions had already occurred.
Gallup’s ownership survey does not measure Greenidge’s operations or establish how changes in retail cryptocurrency ownership affect the Dresden facility.
What the surveys can tell New York
The three reports provide evidence about public attitudes and behavior, but their results cannot be combined into a single measure of technology support.
Pew surveyed 10,548 U.S. adults from July 20 through Aug. 9. Its methodology lists an overall margin of sampling error of plus or minus 1.3 percentage points.
Gallup’s cryptocurrency survey interviewed 2,043 U.S. adults online June 1-15, including 1,048 investors. The margins of sampling error were plus or minus 3 percentage points for all adults and 4 points for investors.
The Gallup-Microsoft AI findings cover the first 37 countries and territories in a project intended to reach 140. Interviews were conducted between April and July among people ages 15 and older, with country-specific methods and margins of error. Most findings about AI’s expected effects, trust and emotions apply specifically to respondents who were aware of it.
None of the reports provides a separate estimate for New York or the Finger Lakes.
The state’s next decisions will concern matters those surveys cannot settle: the environmental effects of particular projects, who pays for infrastructure and whether negotiated community benefits are delivered. New York’s permitting review and community investment guidance put those questions before regulators and local governments as AI research continues and cryptocurrency remains part of the region’s energy economy.
FingerLakes1.com is the region’s leading all-digital news publication. The company was founded in 1998 and has been keeping residents informed for more than two decades. Have a lead? Send it to news@fingerlakes1.com.