The Local Election That Could Reshape AI
Washington debates artificial intelligence, but governors, county boards and electricity regulators may decide whether America can build the infrastructure that powers it.
The overlooked midterm contest
The most economically consequential races in America’s November 3 midterm elections may not be for Congress. State and local contests could determine whether the United States continues rapidly constructing the data centres, power plants and transmission infrastructure required for artificial intelligence.
Congress controls national spending and regulation, but governors, state legislatures, county boards and planning commissions control many of the practical decisions that allow projects to proceed: land-use approval, environmental reviews, electricity-rate structures, tax incentives and connections to the power grid.
That authority turns previously obscure local elections into strategic contests over America’s technological future.
AI meets local resistance
The AI investment boom is increasingly colliding with concerns about electricity prices, water consumption, noise, land use and limited local employment. Communities may receive substantial tax revenue from data centres, but residents fear that households could eventually bear part of the cost of new power generation and transmission.
The political risk became visible in Loudoun County, Virginia. County officials announced a preliminary agreement under which Amazon would abandon proposed data-centre development on a former George Washington University campus. Amazon also offered land for community facilities and $5 million for the county’s housing trust fund.
The decision does not signal the end of American AI investment. It demonstrates that local opposition can redirect hundreds of millions of dollars even when Washington supports accelerated construction.
Electricity is becoming the constraint
The Department of Energy says data centres consumed approximately 4.4% of US electricity in 2023 and could consume between 6.7% and 12% by 2028. The Energy Information Administration projects that data-centre servers will remain a major driver of long-term electricity-demand growth.
The central issue is therefore no longer computing power alone. It is whether electricity networks can expand quickly enough—and who will pay.
Politicians who promise stricter permitting, environmental safeguards or protection from higher utility bills may gain support in heavily affected communities. Conversely, jurisdictions that offer predictable approval processes and credible community benefits could attract investment diverted from resistant regions.
WARYATV assessment
The AI race is becoming a competition between political systems as much as technology companies. China can coordinate industrial land, electricity and infrastructure through centralized authority. The United States depends upon negotiations among federal agencies, states, utilities, counties and residents.
That decentralized structure provides public accountability, but it can also produce delay and uncertainty. A widespread local backlash could slow data-centre development, raise construction costs and redirect investment among American states—or toward foreign jurisdictions offering abundant electricity.
Yet some resistance may strengthen the industry by forcing developers to finance grid improvements, conserve water and deliver measurable local benefits. The decisive divide is not simply Democratic versus Republican. It is between projects that impose costs on communities and projects that persuade those communities they are receiving a fair return.
America’s AI future will not be secured by investment announcements alone. It will depend on whether technology companies can obtain something increasingly scarce: durable public consent.
