Voters Oppose AI & Data Centers—No Party Has an Edge

What the NYT/Siena Poll Reveals About Voter Attitudes Toward AI

Sixty-one percent. That single figure, buried in a joint New York Times and Siena University poll released this week, encapsulates a growing political reality that neither Republicans nor Democrats have found a way to manage: the American public has turned against AI infrastructure, and the opposition is broad, deep, and increasingly resistant to partisan spin.

The AI public opinion poll surveyed 1,503 likely voters and asked a direct question — do you support or oppose the construction of data centers to power artificial intelligence technology? The result was unambiguous. A substantial majority said no. Not a slim plurality hedging toward skepticism, but a commanding 61 percent expressing opposition. For an industry that has spent the last several years cultivating a public image built around economic promise and national competitiveness, that number represents a serious reckoning.

What makes the finding particularly striking is the methodological rigor behind it. The Times-Siena polling operation is among the most scrutinized survey programs in American journalism, with a track record that draws serious attention from campaigns, academics, and policy analysts. When an institution of that caliber produces a supermajority finding on a technology question, the political class takes notice — and according to reporting on the poll, politicians already are.

The survey arrives at a moment when AI infrastructure is expanding at an unprecedented pace. Tech companies have announced billions in data center investments across dozens of states, often framing those projects as job creators and symbols of American technological leadership. Voters, it appears, are not convinced.

Why Voters Are Turning Against AI Infrastructure

Data centers are not abstract. That is the central insight that often gets lost in coverage of AI backlash, which tends to frame opposition as ideological or rooted in fear of automation. The opposition measured in the Times-Siena poll is, in significant part, a response to physical, tangible impacts that communities experience when a large-scale computing facility arrives nearby.

Energy economists have documented the strain that hyperscale data centers place on regional power grids. A single large facility can draw anywhere from 50 to 200 megawatts of power — enough to supply tens of thousands of homes. When multiple facilities cluster in a region, as they have in Northern Virginia's data center corridor or in parts of Iowa and Texas, the cumulative grid load forces utility companies to accelerate infrastructure upgrades, costs that are often passed directly to ratepayers. Residents who have never heard the phrase "AI public opinion poll" know exactly what it means when their electricity bill climbs.

Water consumption adds another layer of grievance. Cooling systems for large data centers consume millions of gallons annually, a fact that carries particular political weight in drought-prone Western states. Local planning boards in communities from Arizona to Georgia have faced public hearings where residents showed up not to debate AI policy in the abstract, but to ask specific questions about aquifer depletion and cooling tower discharge.

Noise is a less-discussed but equally real irritant. Industrial cooling equipment — fans, chillers, and associated mechanical systems — generates continuous low-frequency sound that neighboring residents describe as a persistent hum. Zoning disputes over data center siting frequently cite noise ordinances as the legal mechanism, but the underlying driver is quality-of-life degradation that voters register viscerally rather than philosophically.

Research from political scientists studying the history of infrastructure opposition suggests that this pattern is not entirely novel, but the speed of it is unusual. New technologies typically enjoy a prolonged honeymoon period during which public enthusiasm outpaces criticism. The compressed timeline of AI opposition — from widespread consumer enthusiasm in 2022 and 2023 to 61 percent infrastructure opposition by mid-2026 — indicates that the industry's growth rate has outpaced its capacity to build public trust.

Neither Party Owns the AI Backlash Issue

Here is where the political story becomes genuinely complicated, and where strategists in both parties are, by most accounts, struggling.

AI skepticism does not sort cleanly along partisan lines. Conservative rural communities object to data center siting on property rights grounds, noise and water impacts, and concerns about foreign-owned facilities near military installations. Progressive urban and suburban voters raise overlapping concerns about energy consumption, environmental justice, and corporate accountability. These are distinct ideological registers producing the same policy outcome: opposition.

For Republicans, the tension is acute. The party has generally positioned itself as favorable to deregulation and private-sector technology investment. Criticizing data center construction puts GOP candidates at odds with major donors in the tech sector and with a rhetorical tradition of opposing federal interference in business development. At the same time, rural constituents — a core Republican voting bloc — are frequently the ones bearing the most direct infrastructure burdens.

Democrats face their own contradiction. The party's climate wing has grown increasingly vocal about the energy intensity of AI systems, noting that data centers powered by fossil fuels undercut decarbonization commitments. But the party's labor and economic development constituencies tend to welcome large capital investments and the construction jobs they generate. Threading that needle in a single campaign message has proven difficult.

The Times-Siena poll suggests that neither party has successfully claimed ownership of this issue, which in electoral terms means neither party is positioned to convert widespread voter skepticism into durable partisan advantage. An issue where 61 percent of likely voters are on one side but no party has clearly aligned with them is, by definition, an opportunity that remains unclaimed.

The Political Stakes of AI and Data Center Policy Heading Into Midterms

Unclaimed opportunities do not stay unclaimed indefinitely. With midterm elections approaching, down-ballot candidates in competitive districts have strong incentives to pick up a populist issue that polls at supermajority opposition levels. The challenge is that data center policy is intensely local — driven by county zoning boards, state utility commissions, and regional water authorities — while AI policy at the federal level remains largely abstract and unlegislated.

That mismatch creates a specific electoral dynamic. Candidates running in areas where data center construction is actively underway or proposed have concrete, visible issues to campaign on. Candidates in districts without active siting disputes have less traction. The result is likely to be a fragmented political response — dozens of localized battles over specific projects rather than a coherent national policy debate.

Political scientists who study technology governance have noted that this fragmentation can actually slow policy development. When opposition is distributed across hundreds of local fights rather than consolidated into a national movement, industry has more opportunities to manage outcomes through community benefit agreements, zoning variances, and regulatory appeals — processes that are far less visible than federal legislation.

The poll's finding that politicians are already responding to this sentiment suggests the dynamic may be shifting, however slowly. Elected officials who see 61 percent opposition in credible survey data tend to adjust their positioning, even when they lack a fully formed alternative policy.

What This Means for the Future of AI Expansion in the U.S.

For the technology industry, the Times-Siena numbers represent a warning that deserves to be read carefully rather than managed through better public relations.

Industries that have faced comparable supermajority opposition in public polling — offshore drilling after major spills, nuclear power after safety incidents — have found that the political environment does not automatically recover. Opposition tends to harden when physical impacts continue and community grievances go unaddressed.

The AI sector's infrastructure ambitions are enormous by any historical measure. The capital expenditure commitments announced by major technology companies over the past two years imply a continuation of aggressive data center buildout well into the next decade. Executing that buildout in an environment where 61 percent of likely voters oppose it will require either a fundamental shift in how those facilities are built and sited, meaningful changes to how their energy and water costs are allocated, or a political strategy capable of moving public opinion — none of which the industry has yet demonstrated at scale.

The AI public opinion poll released this week does not predict a legislative outcome. It does not tell us which party will successfully convert voter anxiety into electoral gains. What it tells us — with the precision of 1,503 surveyed likely voters and the credibility of two institutions with a combined century of public affairs research — is that the politics of artificial intelligence infrastructure have entered a new phase. The public has formed a view. The parties have not yet formed a coherent response. That gap, for now, belongs to nobody.

Source: The Verge