News

After the Midterms, AI’s Next Bottleneck May Be Political

• By John K. Waters
• 11/05/2026

For an AI company planning its next expansion, the hardest question may no longer be which chips to buy. It may be who will let it build the place to put them.

That is the stark reality emerging from the November 3 U.S. midterm elections, which handed Democrats majorities in both the House and Senate. The result gives the party control of congressional oversight when the new Congress convenes in January, while a Republican administration remains in the White House.

Although some vote counting continued past November 5, control of both chambers was clear. Early reports highlighted widespread Democratic gains in traditional strongholds, driven by voter anxiety over the cost of living and international conflicts.

For the AI sector, the implications extend from executives facing intense committee hearings to infrastructure projects encountering tougher regional negotiations. These distinct pressures operate on entirely different timelines, and neither indicates that enterprise AI adoption is grinding to a halt.

An AI Mandate, or an Anti-Trump Vote?
“Voters just sent a clear message that they want rules for AI and data centers," Americans for Responsible Innovation co-founder Brad Carson told Axios.

Adam Kovacevich, founder and CEO of the center-left technology industry coalition Chamber of Progress, disputed that reading. “Let’s not confuse an anti-administration wave with an anti-tech mandate,” he said in a post-election statement.

The disagreement matters because an electoral victory does not supply a ready-made AI policy. A vote against an incumbent can encompass dissatisfaction with prices or domestic economic issues without endorsing a specific approach to model safety.

Alejandra Montoya-Boyer, vice president of The Leadership Conference’s Center for Civil Rights and Technology, interpreted the results as a direct demand for protection from unchecked AI. “The midterms have been the first true AI election in the US,” she told Ars Technica. “We saw voters turn up at the polls in record numbers and vote for leaders who are willing to protect them and their rights from unchecked AI."

Republican technology consultant Evan Swarztrauber also saw political costs in his party’s association with unconstrained technology companies. However, he identified broader macroeconomic factors and foreign policy tensions as the primary drivers of voter sentiment, according to POLITICO.

A pre-election analysis from the Brookings Institution offers a framework for why these explanations overlap. Center for Technology Innovation Director Nicol Turner Lee and Senior Fellow Darrell M. West noted that data centers have essentially become a convenient proxy for major social debates regarding the future. Local opposition can therefore carry broader concerns about who benefits from AI, meaning not every successful candidate shares the exact same policy agenda.

Hearings Before Statutes
The clearest operational shift is institutional: Democrats will now set committee agendas and direct investigations. This oversight is a far more immediate prospect than passing a comprehensive federal AI statute.

“Now that they have the gavels, we should see a heightened committee investigatory process, and state Attorneys General will certainly be paying attention,” Stephen Cobb, a Foley & Lardner partner who leads its State Attorneys General practice, told CIO Dive.

Varun Krovi, executive director of the Center for AI Safety Action Fund, similarly expects a series of blockbuster hearings featuring high-profile tech executives, according to The Washington Post.

Former congressional staff official Michael Hettinger sees a similar trajectory at the House Oversight Committee. “You're going to see a lot of activity from Garcia and his team on the subpoena front,” he told Government Executive, referring to incoming Democratic leadership under Rep. Robert Garcia.

Importantly, these questions will reach far beyond model builders. “There will be an effort to do oversight and investigations into aspects of AI that are touching people’s lives on an everyday basis,” noted Bharat Ramamurti, a former economic adviser to President Joe Biden and Sen. Elizabeth Warren. This broadens the stakes for enterprise buyers: any organization deploying public-facing AI now shares this regulatory exposure, even if it has never trained a model itself.

Still, public scrutiny does not automatically equal effective policy. As California Representative Sam Liccardo told The Verge, “I think what is imperative for Democrats is to come to the table with solutions, not with slogans.” His warning exposes a central tension within the new majority over whether the scrutiny produces workable rules or primarily public confrontation.

Furthermore, dynamic control of Congress does not instantly eliminate gridlock. Because the executive branch retains a presidential veto, the election fundamentally changes who can demand answers, not which answers become law.

The Infrastructure Fight Has a Different Map
Congressional control is only one piece of the puzzle. Democratic victories in key gubernatorial races strengthened local executives who campaigned on restricting rapid data-center development. These contests carry immediate consequences beyond Washington, as local zoning and grid disputes do not require a federal mandate to halt a project.

Libby Cantrill, head of public policy at PIMCO, emphasized this distinction, telling Financial Investigator that near-term risks to the AI infrastructure buildout over the next two years sit primarily in state and local politics rather than Washington, D.C.

This localized friction fundamentally alters how enterprise customers must evaluate suppliers. Meta executive Dina Powell McCormick highlighted this shift at the WSJ Tech conference, pointing out that initiating entirely new data center build-outs in the United States will be significantly more challenging than expanding established, pre-approved sites.

Technology analyst Ben Thompson gave this landscape a sharper formulation in his Stratechery analysis: “The final constraint is political, and that may be the most imposing constraint of all.” Where others identify the levels of government to watch, Thompson questions an underlying business assumption: that technical and financial resources will eventually be enough to deliver the required capacity on their own.

A Slower Boom Could Produce Different Winners
“Tech needs to adjust and focus on a slower pace of this AI rollout,” Wedbush Securities managing director Dan Ives told Fortune.

Crucially, a slower infrastructure expansion will not affect every supplier equally. This shifts the competitive landscape from an industry-wide slowdown to a battle over existing moats. A legacy hyperscaler with facilities already operating faces a very different reality than an emerging challenger whose expansion depends entirely on obtaining new local approvals.

Commenting on the potential of a localized construction pause, a Financial Times analysis put the competitive advantage bluntly: “Clearly, it's good for those who have lots of capacity already.” Restrictions originally intended to restrain the tech sector may inadvertently make existing digital real estate vastly more valuable.

Economist Mohamed El-Erian offered a more cautious interpretation of this cooling market. The post-election shift “will let more of the air out of the bubble,” he noted in a POLITICO report. Yet while his critique targets inflated market valuations, it stops short of predicting a structural collapse in actual enterprise AI demand.

Offering a counterweight to lingering political anxiety, Goldman Sachs analyst Jenny Grimberg argues that infrastructure bottlenecks may be temporary. “We expect the data center issue to somewhat fade in importance ahead and any moratoria on data center development to be short-lived,” she stated in a client note via Investing.com.

For enterprise buyers, navigating this post-election landscape requires looking past the political noise. The true timeline of AI deployment will not be determined by sweeping congressional bills, but by the local political moats protecting the physical infrastructure beneath the cloud.