Voters rejected data center expansion as U.S. adversaries exploited American divisions. How much of the backlash was homegrown - and how much was amplified from abroad?

One of Tuesday’s biggest losers never really appeared on a ballot: America’s advantage in artificial intelligence.

Across communities voting on data center development, the verdict was overwhelmingly restrictive. Candidates promising tighter controls also won important offices. 

For Americans who see AI infrastructure as an economic opportunity and a strategic necessity, the results deserve more attention than another round of partisan celebration.

A country cannot lead the computing revolution while making it increasingly difficult to build the facilities that power it.

Ballotpedia counted 41 local measures on Election Day. Forty produced outcomes favoring restrictions. The measures differed: some imposed bans, others temporary moratoriums, while some rejected development proposals or expressed nonbinding preferences. 

This was no nationwide referendum on AI — not exactly. Nevertheless, the consistency of the results signals a formidable obstacle to expansion.

There are legitimate concerns. Residents deserve clear answers about electricity bills, water consumption, noise and tax incentives. Developers should cover the infrastructure costs their projects cause. Local officials should negotiate publicly and enforce their agreements.

But addressing those concerns is different from treating every proposed data center as a threat. Electricity demand can be met by expanding supply. Appropriate industrial sites can accommodate development. Standards can protect residents while allowing investment.

A blanket refusal to build carries costs of its own.

China has an obvious interest in those costs. Computing capacity supports AI research, commercial applications and national security capabilities. If American projects encounter mounting delays while competitors expand, our current strengths become harder to sustain.

That does not mean every canceled project constitutes a Chinese victory. It means an accumulation of avoidable restrictions could produce a strategic disadvantage that local voters were never asked to consider.

More troublingly, Chinese efforts to influence this debate are documented.

OpenAI identified an operation it called “Data Center Bandwagon,” involving accounts likely originating in China that generated comments and images about American electricity costs. Operators used fabricated American identities and circulated material alongside legitimate reporting.

OpenAI assessed that they likely worked for a private Chinese technology company serving provincial government clients. That is a significant reported connection, although it does not establish direct instructions from Beijing’s central leadership.

X separately reported approximately 200 accounts posting about American AI and energy policy within a suspected Chinese network of roughly 200,000 accounts. The larger number should not be mistaken for the size of the identified data center campaign.

The sinister element is the deception: foreign operators posing as Americans to manipulate an American argument.

Beijing’s propaganda machinery has also exploited Western dissatisfaction. A Global Times editorial argued that Western democracy was losing its appeal. That is a familiar authoritarian maneuver: present democratic frustration as evidence that citizens no longer want democratic government.

Americans voting incumbents out demonstrate their ability to demand accountability. Beijing’s preferred interpretation conveniently overlooks that distinction.

Iran, meanwhile, welcomed the election as a rejection of war. Parliamentary Speaker Mohammad Bagher Ghalibaf publicly endorsed that interpretation. Tehran can use the results to suggest that enduring American pressure eventually weakens Washington’s political resolve.

Iran’s messaging and China’s infrastructure interests are separate issues. Neither proves an overall geopolitical victory. But American leaders should recognize the opportunities both governments perceive.

There is another uncomfortable possibility: restrictions could strengthen the biggest technology companies.

Existing facilities become more valuable when new competitors cannot build. Wealthy incumbents can better absorb delays, financing expenses and community negotiations. Smaller challengers may disappear. Politicians promising to confront billionaire power could inadvertently protect it.

OpenAI says it will continue investing. Its new commitments include paying attributable energy and infrastructure costs, improving disclosure and involving communities earlier at future campuses it owns and operates. Those promises deserve enforceable agreements and careful scrutiny.

They also suggest a practical alternative to paralysis.

America can insist on responsible development while building more power, expanding computing capacity and preserving competition. Communities can negotiate meaningful benefits without surrendering every opportunity.

Was Tuesday a lucky break for America’s enemies, or something more sinister? The evidence supports a combination of genuine domestic opposition and documented foreign attempts to exploit it. The scale of their influence remains unproven.

The strategic danger is clearer. America could allow understandable local anxieties, partisan incentives and hostile amplification to obstruct infrastructure essential to its future.

Our adversaries do not need to manufacture every fear. They benefit if we let fear determine what America can build.

(Contributing writer, Brooke Bell)