Silicon Valley's AI leadership class has spent the last decade building on the assumption that democratic consent is optional. Stanford ethicist Rob Reich recounts a dinner where technologists casually dismissed democracy as too slow for their vision of a tech-optimized nation-state — preferring a 'beneficent technocrat' at the helm. That assumption is now being stress-tested by the American public. The numbers are stark. A Marquette University Law School survey finds 73% of Americans believe datacenter costs outweigh their benefits, and 70% think AI is bad for society. Pew Research tracks the concerned-versus-excited ratio climbing from 37% concerned in 2021 to 52% today, with even 55% of young adults — historically the most tech-enthusiastic cohort — turning skeptical. This is not a partisan divide: Democrats and Republicans are equally uneasy. Americans are also more mistrustful of AI than people in most other countries surveyed. The backlash has teeth. Villages, counties, cities, and states — including New York — are imposing moratoria and outright bans on new datacenter construction. The revolt is not merely NIMBY opposition to local development; it channels a deeper fury at a civilization-altering project pursued without democratic buy-in. As Johns Hopkins political scientist Henry Farrell frames it, the anti-datacenter push 'speaks to the frustration of people who don't feel they are making the choices in their lives.' The industry's structural insulation remains formidable, however. Trump remains firmly aligned with the datacenter buildout. The EPA is planning to eliminate the rule requiring states to solicit public input on air pollution permits for datacenters and their power plants — a move that could keep communities in the dark until construction begins. The bipartisan consensus that the US must stay ahead of China in AI capabilities gives the industry a geopolitical trump card that overrides voter sentiment. As Boston University historian Quinn Slobodian observes: 'If there is one sector in the US that is still globally dominant and profitable, it is tech. They can be confident that even if there is a change of leadership, the next person will have to dance with them anyway.' The article draws a revealing historical parallel. Previous technological revolutions — mechanization, industrialization — also proceeded without democratic consent. But those transitions ultimately delivered broad economic uplift. The automation wave of recent decades has not. It has flushed workers without college degrees into low-productivity service jobs at near-minimum wage. Americans have earned their skepticism about promises that AI will deliver shared prosperity, especially when those promises come from billionaires who dream of seasteading enclaves outside any legal jurisdiction. The deeper structural story is about the relationship between capitalism and democracy. The neoliberal project — independent central banks, WTO agreements, international treaties — was designed to insulate economic policy from democratic pressure, but it still acknowledged democracy's legitimacy. Silicon Valley's leadership class does not. Slobodian characterizes their worldview bluntly: 'The messiness of democracy has no function or utility. It is only bad code.' Peter Thiel has argued explicitly that democracy and liberty are incompatible. What emerges is a collision with no clear resolution mechanism. The public is organizing locally with real effect, but the federal apparatus and geopolitical logic are aligned with the industry. AI ethicist Rumman Chowdhury identifies the industry's master narrative — that AI systems are becoming too powerful for even their creators to control — as a deliberate stratagem to preempt legal accountability. The question is not whether democracy can stop the AI buildout, but whether the democratic system can impose meaningful conditions on it before the buildout renders the question moot.
