Democratic Wins Heighten Scrutiny of Data Center Expansion

Democrats notched their largest congressional majorities in decades in this week's midterms, and local resistance to data center construction appears to have played a measurable role in several key races and ballot measures. The party's gains extended well beyond traditional battlegrounds, with voters in suburban and rural districts rejecting both Republican incumbents and industry-backed projects tied to the artificial intelligence boom.

In New Jersey's 2nd congressional district, longtime Republican Representative Jeff Van Drew lost by three points to Democratic challenger Zack Mullock despite winning the seat by double digits two years earlier. The district had shifted rightward in recent cycles, yet the race tightened after the Democratic Congressional Campaign Committee spent $1.5 million highlighting local opposition to a large data center planned for the area. Residents cited noise, water use and strain on the electric grid as primary concerns. Similar dynamics surfaced in other states, where more than 40 ballot propositions aimed at curbing data center development passed or came close on Election Day.

Industry executives gathered at a Data Center Dynamics conference in Virginia during the voting expressed surprise at the scale of the backlash. Developers described much of the criticism as misinformation, though several acknowledged that earlier projects had sometimes overlooked community impacts. The timing underscored a growing disconnect: while companies forecast record construction to meet AI demand, voters in affected areas used the midterms to impose new limits.

The political fallout reaches into the incoming Trump administration as well. Interior Secretary Doug Burgum, tasked with accelerating AI infrastructure on public lands, holds between $18 million and $64 million in AI-related investments through his former firm, according to disclosures analyzed by public lands advocates. Burgum has framed the buildout as a national security imperative comparable to the Manhattan Project, yet the financial overlap has drawn scrutiny from both environmental groups and some Republican lawmakers wary of federal land decisions.

Companies are already adjusting strategies. Equinix, one of the largest data center operators, signaled it will prioritize facilities under 100 megawatts to avoid new state-level restrictions that often target larger projects. The firm has raised its annual capital spending to $7 billion amid surging demand, but executives noted that roughly half of projected U.S. hyperscale growth sits in states where newly elected Democratic governors may impose stricter permitting rules. Most 2027 projects remain committed, though delays and higher costs are expected starting in 2028.

The pattern reflects a classic tension between national technological goals and local governance. Data centers require enormous amounts of power and land, and their clustering in certain regions has produced concentrated costs that voters proved willing to punish at the ballot box. Democrats gained ground in places where these projects intersected with existing concerns over energy prices and environmental quality. Republicans, for their part, face internal divisions as the administration pushes aggressive expansion while some state-level officials confront the same community pushback.

How Congress and state legislatures reconcile these pressures will shape both the pace of AI deployment and the durability of the current electoral map. Early signs suggest the industry will adapt by siting smaller facilities and seeking more cooperative jurisdictions, yet the midterm results indicate that local consent can no longer be assumed.