No Contest: Republicanism and Data Centers

In divisive times, issues where most Americans agree are rare. However, we appear to be coalescing on one: opposition to data centers in our backyards. A new Politico poll found that 41% of respondents would oppose a data center constructed within 3 miles of their residence, while 24% would support it. These results may not appear striking, but the shift is noteworthy; a poll with the same question, conducted in January, found that 37% of respondents said they would support a data center near their home while 28% expressed disapproval. What’s striking here is not the absolute result, but the swing. A shift of +9% to -17% approval in months. Even so, these polls appear to be low outliers. Reuters, in contrast, saw a majority express disapproval of data center construction within their communities and concern that AI will raise the price of electricity, while rejecting the idea that rapid data center construction is a good thing. A University of Pennsylvania poll found that most respondents, regardless of political affiliation, oppose the construction of data centers in their area. A Gallup poll showed that 71% of respondents opposed local data center construction, with 48% strongly opposing.

Why the opposition to data centers? Opponents think they offer a losing trade-off; data centers create burdens on local communities, and their detractors think the benefits are not worth these sacrifices. Data centers are the physical structures supporting much of the internet – any cloud-based or streaming service requires the use of data stored somewhere physically – however, the current data center rollout has been driven by AI firms. CEOs across industry have warned for years that AI will take your job. Industry insiders are warning that AI could kill us all soon. Data centers serve as a physical manifestation of such threats, prompting immediate disdain from many.

Proponents, meanwhile, argue that data centers create new jobs and raise tax revenue. However, most of these jobs are temporary construction work (and even these workers who benefit from the construction may have reasons to care about the long-term impacts of data centers on labor, as Evan Arnet recently explored). Data centers create far fewer permanent jobs, and these jobs may cost up to $2 million per job in terms of tax incentives given to data centers. Data centers in general were responsible for 4.4% of all energy consumption in the U.S. in 2024, and a single hyperscale data center can consume as much electricity as 50,000 households. In regions located near significant data center activity, electricity costs have increased by as much as 267% since 2020. And the noise and air pollution produced by data centers may threaten the health of nearby residents.

The temperature on this issue is rising, however, because residents are losing opportunities to voice their concerns about the construction of data centers. Public officials that negotiate deals permitting the construction of data centers have signed NDAs in some cases, preventing information about the projects from reaching the public. Data centers have avoided public commentary periods by exploiting zoning laws. The EPA has recently moved to eliminate the public comment period before issuing air pollution permits to new data centers, with the White House directing agencies to accelerate data center construction in general. When members of the public do have the opportunity to express disapproval of data centers, they often face reprisal. Citizens have been removed from, and arrested at, local council meetings for infractions such as going over their allotted time, threatening board members with recall elections, and clapping.

My goal in this column is to do some diagnostic work. The hope is that, if we can properly diagnose the problem, we can better understand what’s necessary to resolve it. Specifically, I argue that we can best understand the current issues surrounding data center construction through the lens of republican political theory, namely, the requirement of contestation. If this diagnosis is correct, then it seems that the solution ought to be focused on altering the distribution of power among the parties to this debate. As things currently stand, it appears to me that the tech companies behind the data center rollout are wielding domineering power to the extent that data centers are constructed without careful consideration of the public’s objections.

I have previously discussed the idea of republicanism in the context of LLMs and the wages of tipped workers. Nonetheless, a brief restatement is pertinent here. In Republicanism (1999), Philip Pettit presents a view of liberty grounded in freedom against domination. When one dominates you, they have the power to arbitrarily interfere with you. By this, Pettit means that the dominator is capable of interfering with you without considering your interests. Notice that the emphasis is on ability. Even if one does not actually interfere with you, when one could interfere with you in this way, they dominate you.

Consider, for instance, the situation of the American colonists prior to the revolution. After Parliament passed the Stamp Act of 1765 and the Townshend Acts, colonists who later went on to rebel felt these acts demonstrated the arbitrary power that Parliament had over them; the colonists had no formal representation in Parliament’s policy making process, and there were no procedural requirements for legislators to formally consider the interests of colonists. Some colonists went so far as to refer to themselves as slaves, living wholly at the whim of Parliament. Their view was not that these taxes themselves were unbearable, but rather that they demonstrated the power Parliament could wield against them in more extreme ways.

To eliminate the power of arbitrary interference, Pettit contends that there must be a mechanism by which those who are subject to the decisions of states and governments can contest them. Although we may develop procedures which force governing bodies to consider the interests of all affected by considered policies, there will inevitably be lapses or cases where some feel their interests were not adequately considered. Thus, affected groups must have the power and opportunity to challenge decisions as arbitrary if they are to have freedom from arbitrary interference. Forms of contestation that Pettit articulates include lawsuits and, pertinent to this case, public commentary periods for policies.

Why is this relevant to data centers? As noted earlier, the opportunities for residents to challenge their construction are being throttled. Large tech companies sign lawmakers to contracts which require withholding pertinent information from the public. They exploit legal loopholes to avoid commentary periods. The Federal government is pushing to wield its regulatory power to streamline this process even further. The measures designed to force regulators to consider the interests of the public, and to give the public the opportunity to contest decisions, are being eroded in the interests of tech firms. As a result, it is hard to see streamlined data center construction as anything but domination.

Of course, maybe the public’s concerns are misplaced. Perhaps there are serious national interests at stake in U.S. companies being at the cutting edge of AI, the downsides of data centers are overblown, and they do generate great windfalls for surrounding communities. But even if the public’s objections stem from ignorance or misinformation, this does not relieve public officials of responsibility for addressing these concerns. When considered in the context of large tech firms spending hundreds of millions a year lobbying Congress and backing candidates while building robust political influencing infrastructure, as their current/former owners are the richest people in the United States, it is hard to view this as anything but the scales being tipped to give their interests greater weight in decision-making than the interests of general citizenry.

The foundational ideal behind a republic is the notion that all citizens are equals, without hierarchy. The state, in recognition of citizens’ equality, is obligated to take seriously the interests and views of citizens in its process of developing policy. To the extent that the objections of citizens are set aside without consideration, even if it does promote the national good, a state fails to be a republic. Many rebuff critiques of undemocratic institutions such as the Electoral College and Supreme Court by noting that the U.S. is a republic rather than a democracy. But if the nation is not democratic, and it fails to meet the foundational standard of republicanism, then what exactly is it?