Commentary. The 2026 midterm elections delivered a message that Silicon Valley should not dismiss: communities want a say in how artificial intelligence is built, powered and regulated. In at least 10 states, voters considered restrictions on data-center development, while local moratoriums and zoning rules reportedly advanced in several communities.[1] The backlash is understandable. It is also not, by itself, a sufficient technology policy.
Our view is straightforward: governments should slow the rush to build, require companies to pay the public costs they create, and protect workers and citizens from demonstrable harms. But they should do so through transparent, enforceable standards—not indiscriminate bans that turn legitimate skepticism into economic self-sabotage.
The public is not imagining the problem
Artificial-intelligence systems depend on vast data centers that consume electricity, land and, in some locations, significant amounts of water. Residents who face higher utility bills, strained local infrastructure or industrial construction next to their homes are right to ask who benefits and who pays. The Atlantic reported that more than 70 percent of Americans opposed data-center construction in their area, while more than 80 percent believed technology was advancing too quickly.[1]
Those numbers should be read as a democratic warning, not as proof that every AI project is harmful. They reflect a loss of trust after years in which technology companies presented disruption as inevitable and regulation as an obstacle. Companies spent heavily during the election to oppose regulations and candidates they viewed as threatening, according to the same report.[1] That spending may have been legal, but it reinforced the impression that public consent was an afterthought.
Residents deserve meaningful disclosure before approval: projected electricity and water use, expected emissions, tax incentives, noise impacts, construction timelines and the number and quality of permanent jobs. Local authorities should also have the power to reject projects that cannot meet those conditions.
Why blanket bans would be a mistake
The counterargument deserves equal weight. Data centers can bring construction work, tax revenue and infrastructure investment. Artificial intelligence may improve medical research, accessibility tools, manufacturing and public services. A patchwork of permanent prohibitions could push investment to jurisdictions with weaker standards, while denying communities the chance to capture benefits on fair terms.
Nor is every concern uniquely an AI concern. Electricity prices, transmission bottlenecks and water scarcity often reflect broader planning failures. Blocking one facility does not solve an aging grid or guarantee that the next industrial project will be cleaner. Policymakers should avoid treating a politically fashionable target as a substitute for competent infrastructure policy.
That is why the emerging debate over federal permitting reform matters. Congress is considering changes to the system for approving power lines, clean-energy projects and AI data centers, with a procedural vote reportedly planned for November 9.[2] Faster approval can be valuable—but only if speed does not erase environmental review, public participation or accountability.
A workable bargain
A better policy would establish national minimum rules while leaving communities room to set stricter ones. Developers should fund grid upgrades proportional to their demand, disclose water use, meet efficiency benchmarks and secure power without shifting hidden costs to households. Tax incentives should be conditional on measurable local benefits, including quality jobs and infrastructure commitments.
Regulators should also distinguish between building a data center and deploying an AI system. The former raises physical and environmental questions; the latter raises issues of privacy, labor displacement, discrimination and safety. Treating both as one argument obscures the choices before lawmakers.
Finally, elected officials should resist both corporate fatalism and populist absolutism. The claim that AI development cannot be slowed is false. The claim that it can be halted without costs is equally false. Democratic governance means deciding which costs are acceptable, who bears them and what safeguards make innovation worth pursuing.
The midterms did not produce a mandate to abandon artificial intelligence. They produced a mandate to stop treating public consent as a technicality. The companies that accept that lesson may still earn the right to grow. Those that do not should expect more moratoriums, stricter rules and a deeper backlash.