TL;DR — Key Takeaways
- Money buys influence, not consent. Political spending and corporate investment can open doors, but voters and communities still have the power to reject the industry’s plans.
- The AI infrastructure boom has a trust problem. Building bigger data centers is relatively straightforward compared with persuading communities to accept their long-term costs.
- Public acceptance is becoming a business requirement. AI developers that fail to address local concerns risk delays, rising costs and lost opportunities, regardless of their technical capabilities.
Elon Musk contributed more than $445 million to Republican candidates and allied political organizations ahead of this election, according to an OpenSecrets analysis of federal campaign-finance disclosures. He was the largest donor of the cycle. That is an extraordinary amount of money for one person to put behind a political outcome, and the outcome was a Democratic takeover of both chambers of Congress. Money bought an enormous campaign. It did not deliver the result.
Across communities voting on data center development, another limit on the industry’s power became visible. Residents were deciding whether the infrastructure behind the AI boom belonged in their towns and under what conditions. For those of us covering technology, these votes deserve attention alongside the congressional results. The industry’s ability to finance something and the public’s willingness to accommodate it are separate considerations. Too many business plans appear to have assumed the second would follow the first.
I would not credit AI with causing the entire Democratic wave. People voted on their finances, the administration and plenty of concerns unrelated to technology. But where data centers appeared directly on ballots, we have something more concrete than speculation about voter motivation. In Sunbury, Ohio, unofficial results showed restrictions winning 78.6% of the vote. In Granville, the figure was 82.4%, with all precincts reporting in both communities. Those are substantial majorities.
Consider Oregon, Ohio, where Cleveland.com reported that opponents of a data center restriction raised $335,000, compared with $4,600 reported by supporters. Vantage Data Centers supplied $250,000 of the opposition funding, with union groups contributing additional money. Voters approved the restriction anyway, according to local broadcaster WTVG. Developers and their allies had ample opportunity to make their case. Residents reached their own conclusion.
The concerns are understandable: Electricity costs, water availability, noise, land use and tax incentives. The consequences of an individual project need to be established, rather than assumed. Developers deserve a fair hearing when they explain jobs, tax revenue and infrastructure improvements. Residents deserve equally credible answers about obligations that could last for decades. Someone can use AI every day and still want to know why a proposed facility should receive tax relief or how its utility requirements will affect the surrounding community.
Calling that person afraid of progress is a good way to lose the next vote, too. I have spent much of my professional life covering technologies that promised to change how we work and live. I believe AI can deliver enormous benefits. That belief does not require accepting every proposed construction project, subsidy or promise of future prosperity. A company asking a community to commit resources should expect questions about what the community receives in return.
Texas makes this more than a collection of local disputes. Gina Hinojosa defeated Greg Abbott after a campaign that included opposition to energy-intensive data centers alongside affordability and public-school issues. Her published platform called for a moratorium pending legislative protections, greater community participation, transparency, an end to subsidies and safeguards against passing electricity and water costs to households. Those positions now carry the weight of an incoming governor’s agenda.
Her election does not automatically cancel permits or stop every construction site. The rules ultimately adopted, the authority behind them and their application to existing projects will determine the consequences. Nevertheless, developers have to reconsider assumptions about approvals, incentives and schedules. A delay can increase financing costs and complicate customer commitments even when a project eventually proceeds. A rendering of a giant campus and a signed customer agreement cannot resolve a dispute over where its power will come from.
Washington presents a different set of constraints. Democrats will control congressional committees, spending negotiations and the Senate’s confirmation process. Technology companies should prepare for closer examination of their relationships with government and the public costs of their expansion. Trump remains president, however. Anyone predicting an immediate, comprehensive Democratic rewrite of AI policy is getting ahead of the machinery required to enact it. More scrutiny and legislative stalemate can coexist.
There is already a substantive policy debate to cover. Senator Maria Cantwell’s October proposal calls for enforceable frontier-AI safety standards, independent audits, reporting of serious incidents and human oversight of consequential decisions. The administration’s March framework emphasizes a uniform national approach, innovation and faster infrastructure deployment. These are competing proposals about how AI should develop, rather than a simple choice between supporting the technology and opposing it.
The White House framework also explicitly says residential ratepayers should not bear data center costs. We should acknowledge that and judge the resulting policies against it. A promise to protect households becomes meaningful through utility arrangements, infrastructure funding and enforceable obligations. Technology leaders would be better served helping establish workable protections than treating every demand for accountability as an attack on American competitiveness.
Musk’s political investment belongs squarely in this discussion. He chose to put extraordinary resources behind Republican candidates and organizations. We do not need to pretend that alignment was ambiguous, or soften it with a ritual paragraph suggesting every technology executive made the same bet. Political access can be valuable to a business. It can also encourage a dangerous assumption that the people inside the room have settled matters for everyone outside it.
Losing an election does not erase Musk’s wealth, his companies’ influence or his ability to spend again. Nor does it prove that money no longer matters in politics. It demonstrates that even extraordinary spending cannot guarantee public acceptance. For executives accustomed to setting product roadmaps and investment priorities, elections impose an uncomfortable requirement: Other people get to decide, and they may have priorities that do not appear in the investor presentation.
For enterprise technology buyers, this belongs in procurement and planning conversations now. If a supplier promises substantial new capacity, ask how much depends on facilities awaiting approvals, power connections or infrastructure agreements. Examine what happens if delivery slips. AI teams should also revisit utilization, model selection and deployment choices. Smaller models or local inference will fit some workloads; others will require large cloud resources. Those decisions should follow actual requirements rather than an assumption that additional capacity will arrive whenever it is needed.
There is an uncomfortable competitive possibility here as well. Restricting new construction could increase the value of existing facilities and favor incumbents that already control power and compute. Smaller companies might face greater difficulty obtaining capacity. Slower expansion could impose useful capital discipline while also concentrating market power. Policymakers need to consider both outcomes, and technology executives cannot assume that every restriction will hurt every company equally.
The industry has practical work to do. Publish credible resource requirements. Explain who pays for grid improvements. Make community commitments enforceable. Distinguish temporary construction employment from permanent jobs. Give residents enough information to evaluate a proposal before asking them to accept it. Those obligations belong alongside the engineering and financing needed to make a project viable.
I want AI to succeed, and I want the people building useful technology to succeed with it. That requires a relationship with the public that survives an election and holds up when promises are tested. The industry keeps asking people to trust it with the future. After this election, it should spend more time explaining what people will get in return, and listening when they say the terms are unacceptable.
