The sudden bipartisan surge against AI data centers looks like grassroots democracy in action. Communities worried about power consumption, environmental impact, and land use competing with housing are raising legitimate concerns. But there is a risk that this visible, tangible fight has become the perfect smokescreen for something far larger: the concentration of wealth and control that AI companies are actually pursuing across entire industries.
The math tells part of the story. U.S. companies are pouring roughly three-quarters of a trillion dollars into data center infrastructure this year alone. That sounds enormous until you realize the enterprise software market is about twice that size. And both pale compared to what these companies ultimately want. AI firms are eyeing industries wholesale. They have already made inroads in customer service and sales. On the horizon are bigger prizes: enterprise software development, creative design, management consulting, legal services, medicine, and education. The companies would far rather spend political capital fighting local data center battles than defending how they reshape entire professions.
Communities fighting data centers have real grievances. These facilities consume land, energy, and resources while creating remarkably few jobs. The fact that fiercest opposition comes from lower-income areas reflects a rightful anger at an unequal bargain. Local residents bear the environmental and infrastructure costs while tech companies and developers pocket the gains. Yet some of these campaigns are succeeding only against the weakest targets. An OpenAI and Oracle-backed facility in Michigan faced local rejection, only to proceed anyway after developers sued the town and forced a settlement. Well-funded projects with serious backing have consistently overcome opposition. The current administration has signaled willingness to override state objections and use federal lands to advance AI infrastructure.
There is also the possibility that the current data center boom is a temporary spike rather than a permanent feature of the landscape. Chinese labs are innovating ways to make frontier AI models smaller and cheaper to run. Open-weight models are being miniaturized by power users to run locally on personal computers. Both Apple and Google are investing in infrastructure to run AI directly on mobile phones. What looks like a long-term infrastructure need could resemble the fiber optic cable bubble from the early 2000s as demand shifts away from centralized computing.
The real threats to broad prosperity and democratic integrity lie elsewhere. Energy prices are being shaped by geopolitics, not data centers. The U.S. is abandoning the renewable energy sector to China while canceling climate commitments. Heating buildings accounts for 10 percent of global carbon emissions, dwarfing AI energy use and treatable through heat pump technology. Housing affordability has been decimated while federal subsidies have flatlined for decades in inflation-adjusted terms.
Wealth concentration inside AI companies represents the greater existential risk. These firms are using political spending to shape the very rules that govern them. In a recent New York congressional Democratic primary, political committees linked to Anthropic and OpenAI spent millions backing competing visions of AI safety. Anthropic favors heavy-handed regulation it can navigate as the ethics-conscious vendor. OpenAI backs a lighter-touch approach that keeps federal control and excludes state regulators. Both companies profit from the narrative that their products are so powerful that controlling them is civilization's paramount challenge. Both gain whether regulators tighten or loosen the reins.
What should happen instead is a political reorientation toward populist resistance to corporate concentration itself. When AI companies pump millions into races, the response should not be panic about superintelligence. When a small town debates a data center proposal, the conversation should include corporate influence in politics and structural solutions like public campaign financing and state regulation. Slowing data center development is a starting point, not the destination. States should regulate AI directly and reject irresponsible uses. Computation could be taxed so the public captures some profit and companies absorb environmental costs. A public AI ecosystem developed under public control with incentives for public benefit rather than private gain could compete with private alternatives.
The real danger is not the infrastructure buildings themselves but the oligarchic power consolidating behind them. AI companies are positioning themselves to dominate how professionals work, how information flows, and how decisions get made across society. Stopping a single data center matters for that community. Stopping the broader agenda requires confronting political power itself.
Author James Rodriguez: "Opposing data centers feels like action, but it may be exactly the distraction these companies want while they rewrite entire industries behind the scenes."
Comments