At a time of profound polarisation, when the fracturing of media ecosystems has been so complete that any agreement on threats to our common life seems impossible, an issue has emerged capable of bringing voters of every political persuasion together.

Over the last year, popular opposition to the construction of the data centres in the United States has risen dramatically. According to a Gallup poll from May 2026, more than seventy per cent of Americans are opposed to the construction of a data centre in their area. In June, a Heatmap poll got the same result, and found that over fifty per cent of respondents said they favoured a ban on new data centres altogether. Meanwhile, fifteen states are considering, have introduced or already passed legislation temporarily banning data centre development.

Something similar is happening in the UK and Australia, where government eagerness to attract vast investment from AI companies has encountered strident opposition from local communities and electorates. The Australian Greens have called for a moratorium on data centres, while Pauline Hanson’s One Nation Party is urging stronger government regulation on AI. The South Australian state government has tried to get out ahead of electoral anxieties by announcing a royal commission into AI.

The proximate reasons for this opposition are straightforward enough. They primarily have to do with the effect of data centres on the environment, water and energy usage, utility costs, the amenity of adjacent communities and noise pollution. Lurking behind these concrete concerns is the more visceral alarm over the sheer speed of advances in AI over the last three years, along with the gargantuan sums of money frontier labs have spent in their pursuit of ever more “compute” and sources of data for training their models, as well as the out-sized influence that this new generation of tech oligarchs have had on government decision making.

Behind it all, there are the deep-seated worries about what this will mean for human beings: whether workers will be sacrificed on the altar of the quest for artificial general intelligence; whether human labour will come to play a subordinate role to the efficient operations of machines (which is to say, whether we will be reduced to “reverse centaurs”, as Cory Doctorow puts it); whether work itself will lose its meaning.

But does the way this popular opposition to data centres has coalesced into a broad-based if inchoate movement mark a productive development in modern democratic politics — say, the resurgence of “class” as a potent political force — or is this simply an extension of the pathologies that have eaten away at democratic politics for the better part of the last decade-and-a-half, including: what Pierre Rosanvallon calls the “negative sovereignty” of aggregated disaffection (which can unite around a common “no”, but cannot come together for a shared political project); populist antipathy toward monied and political elites alike; pervasive distrust that veers into unfalsifiable conspiracy theories.

There are plenty of good reasons to be politically and ethically suspicious of AI, about the resources it demands, the benefits it promises and the distorting effects it has on the economy. But perhaps the most pressing question of all is whether opposition to data centres can turn into a way of making AI democratically accountable, and whether the limits erected by popular consultation and consent can translate into a deeper form of political empowerment — an affirmation of the importance of cooperation and deliberation for the flourishing of our communities?

Guest: Craig Calhoun is University Professor of Social Sciences at Arizona State University and Visiting Professor of Sociology at Princeton University. He is in Australia to deliver the 2026 Thesis Eleven Annual Lecture on the decline of Western democracy, in Melbourne on 25 August 2026.