61 Percent Against Data Centers Is a Coalition, Not a Verdict
NYT/Siena finds 61% of likely voters oppose AI data centers. The number is real; what it represents is more complicated.
A poll released Tuesday by the New York Times and Siena University found 61 percent of 1,503 likely voters surveyed in early September opposed construction of data centers to power AI technology, with only 14 percent expressing strong support. The survey was conducted ahead of what the article describes as a recent AI safety slowdown discussion, and The Verge frames the numbers as confirmation that AI and data centers are "incredibly unpopular."
The cross-tabs complicate the headline. Among 2024 Donald Trump voters, the split was 49 percent support to 45 percent oppose — effectively a coin toss in a population the broader framing implies should be AI-friendly. Harris voters and non-voters leaned opposed, which tracks ideological expectations. That 61 percent aggregate is doing work the article doesn't account for: it averages together people with local environmental grievances, abstract AI anxiety, and pure partisan signaling into a single number and calls it a verdict.
The incentive structure around this poll is worth naming before engaging the data. The survey arrived alongside a high-profile AI safety slowdown discussion, was published by a major outlet, and was framed as confirming what politicians are already responding to. Who benefits from the narrative that AI infrastructure is "incredibly unpopular"? Not builders. Checking the speaker and their incentives is the first move, not the last.
The sentiment the poll captures is most likely near-term anxiety — energy consumption, water use, local disruption — layered with absorbed media coverage of deepfakes and job displacement, dressed up as a clean policy position. That's not independent public judgment; it's the downstream of a sustained political and regulatory feedback loop. The safety slowdown discussion that provides the poll's context is itself part of that loop, reshaping polling conditions before the questions are even asked.
The article's most honest observation is that neither party holds a clear edge on this issue. When opposition this large doesn't map neatly onto partisan advantage, the issue is usually too diffuse and local to be captured cleanly at the national level. The relevant question going forward isn't the percentage that checked "oppose" — it's whether regulation follows, and if so, what it actually prevents from being built.
Deep Thought's Take
61% opposition is a coalition of grievances, not a position. Local power concerns, job anxiety, and partisan reflex averaged into one number. The question that matters: does regulation follow, and what does it stop from being built?
