Three-quarters of the American public have heard of datacenters, but they haven't quite made their minds up yet about whether they approve of them or not. That's the headline from a new Pew Research Center survey. But dig deeper and you find something more interesting: the public's intuitions about datacenters contain a kernel of truth, even if their overall assessment misses the real point.
The survey involved 8,512 adult Americans conducted in January 2026. Three-quarters of Americans say they've heard or read a lot (25%) or a little (50%) about data centers. The findings paint a picture of genuine uncertainty. Far more say data centers are mostly bad than good for the environment (39% vs. 4%), home energy costs (38% vs. 6%) and the quality of life for those nearby (30% vs. 6%). Yet when asked about local economic benefits, the picture flips. More adults say data centers are mostly good than bad for local jobs (25% vs. 15%) and local tax revenue (23% vs. 12%).
This contradiction is worth examining. Public scepticism about environmental and energy impacts reflects legitimate concerns backed by evidence. Data centres have drawn new attention in recent years as the global race to develop artificial intelligence intensifies; tech companies have invested hundreds of billions of dollars to construct buildings filled with computer equipment that can help train and run AI models. The energy demands are real and growing.
But here's where the picture gets complicated. The public's optimism about job creation and tax revenue doesn't square with actual performance. Good Jobs First, a non-profit body focused on corporate and government accountability, found that datacenter developers are benefiting massively from local subsidies in the majority of US states, and those states that do calculate their returns find they are losing money on the deals. In other words, taxpayers are bankrolling facilities that often don't deliver promised returns.
The job story is particularly telling. While hyperscalers like to claim their facilities bring jobs to a region, almost all of the work pertains to building the site. Once operational, many server halls employ around 30 to 50 permanent positions, while larger facilities can employ up to 200 people. The temporary construction work disappears once the facility opens. The permanent jobs don't materialise in the numbers promised.
Respondents who were Democrat-leaning view the impact of datacenters more negatively than Republicans. However, few Democrats or Republicans actually believe that server halls are mostly good for the environment, home energy costs and people's quality of life. Across the political spectrum, the pessimism is real. Younger adults are more sceptical than older ones.
There's an important detail the survey reveals about knowledge itself. Those who say they have heard a lot about datacenters are more likely to hold negative views across all five areas surveyed. That suggests that familiarity actually does breed contempt when people learn the details. The industry has an image problem it knows it has.
The tension here reflects a genuine policy trade-off. Datacenters do generate some tax revenue and provide some permanent employment. They also drive demand for electricity and water at a moment when both resources are increasingly constrained. Communities see the disruption and energy use up close. The promised jobs never quite materialise at the scale claimed. The subsidies flow before anyone can verify the benefits.
The American public isn't wrong to be sceptical. They're just looking at incomplete information. What they've grasped — that something doesn't add up — is correct. The real scandal isn't datacenters themselves but the way public money has been used to subsidise them without rigorous accountability for results. That's a problem fiscal responsibility demands addressing.