A recent Business Insider investigation revealed a staggering truth: America now has 1,240 data centers, nearly four times the number from 2010, collectively poised to consume more electricity than Poland and guzzle billions of gallons of water daily. Yet the most revealing detail wasn't the scale—it was the decision-makers. These trillion-dollar infrastructure bets that reshape entire communities are being negotiated between "the private-jet-flying, expensive-suit-wearing representatives of transnational technology oligopolies and, like, Dave from the local planning commission."
We have a fundamental mismatch between the scale of technological infrastructure and the governance structures managing it. This isn't just an AI problem—it's a preview of far bigger challenges ahead.
The Infrastructure Blind Spot in Responsible AI
The responsible AI conversation has focused intensely on model safety, bias mitigation, and algorithmic transparency. These are crucial concerns. But whilst we debate the ethics of AI outputs, we've largely ignored the physical infrastructure that makes AI possible—and the communities bearing its costs.
Business Insider's analysis exposes the governance gap. Forty percent of US data centers sit in water-scarce regions, including Arizona's Maricopa County, an "extremely water-stressed area" hosting 48 facilities. The pollution from powering these centers will cost between $5.7 billion and $9.2 billion annually in public health impacts. Local electricity bills rise for everyone when data centers arrive, yet taxpayers subsidise their construction through tax breaks worth over $1 million per job created.
This is responsible scaling in reverse—externalising environmental and social costs whilst privatising technological benefits. We're building the physical foundation of the AI economy through unsustainable resource extraction and democratic deficit.
The companies involved know these costs exist. They've made pledges about water neutrality and renewable energy investments. But these commitments come after communities have already absorbed the initial impacts, and often rely on offset mechanisms rather than fundamental changes to resource-intensive development patterns.
The Scale Mismatch Problem
The current system treats massive infrastructure decisions as local land-use questions. Counties approve permits for facilities that will consume city-level amounts of electricity and water because that's how American development has always worked. But AI infrastructure operates at a completely different scale than traditional industrial development.
Companies choose to build massive centralized facilities partly because it's simpler from a governance perspective. One giant data center means one permit process, one set of tax negotiations, one community to manage. But this regulatory convenience often leads to less efficient infrastructure choices. Distributed smaller facilities could reduce cooling costs, improve resilience, and spread environmental impacts more evenly - but they require navigating dozens of local jurisdictions instead of one.
Our governance structures inadvertently incentivize technological choices that concentrate both power consumption and community impacts, rather than encouraging more sustainable distributed approaches.
Infrastructure Governance That Actually Scales
Responsible infrastructure scaling requires governance frameworks that match the scope and impact of these systems. This means:
Regional resource planning that accounts for cumulative infrastructure impacts across multiple jurisdictions. A single data center might fit within local water allocations, but dozens of them create basin-wide resource stress that no individual planning commission can address.
Transparent cost accounting that makes infrastructure resource requirements public information rather than corporate secrets. Business Insider had to reverse-engineer power consumption from emissions permits because usage data isn't systematically disclosed.
Democratic participation structures that give affected communities meaningful input into infrastructure decisions rather than after-the-fact notification. The current system essentially lets technology companies choose their regulatory counterparts by selecting friendly jurisdictions.
Benefit-sharing mechanisms that ensure communities hosting AI infrastructure capture appropriate value rather than just absorbing costs. If these facilities are generating tremendous wealth from local resources, local populations should share in that prosperity.
These aren't radical interventions—they're basic governance principles applied to infrastructure that happens to power artificial intelligence.
The Quantum Horizon: Getting It Right This Time
While we're still wrestling with AI infrastructure governance, quantum computing looms on the horizon with the potential to make today's challenges look modest. Quantum systems require extreme cooling, electromagnetic shielding, and specialised facilities that could reshape infrastructure requirements entirely.
The computational capabilities quantum systems promise—from breaking current encryption to revolutionising drug discovery to transforming climate modelling—could drive infrastructure demands that dwarf today's data centers. If quantum computers can solve certain problems exponentially faster than classical computers, the economic incentives to build quantum infrastructure could trigger another resource-intensive boom.
But we have an opportunity that we missed with AI: the chance to develop responsible infrastructure governance frameworks before the technology creates compelling economic pressures. We can learn from the current AI infrastructure crisis to build governance muscle for quantum transitions.
The quantum revolution will likely follow a similar pattern to AI if we don't intervene: initial development in research settings, breakthrough demonstrations that attract massive investment, rapid scaling driven by competitive pressures, and infrastructure build-out that outpaces governance development. We're seeing this exact sequence play out with AI infrastructure right now.
The difference is that quantum computing's potential applications—from financial modelling to national security to scientific research—could create even stronger incentives for rapid, resource-intensive deployment. If quantum computers can provide decisive advantages in strategic domains, the pressure to build quantum infrastructure quickly may overwhelm careful resource planning.
Quantum infrastructure governance will need to address not just resource consumption but also geographic concentration of strategic capabilities, international technology competition, and the societal implications of quantum-enabled AI systems that could emerge from these facilities.
Building Governance Muscle Now
The AI infrastructure crisis isn't just a problem to solve—it's a training ground for responsible technological governance. The frameworks we develop for managing data center impacts, resource allocation, and democratic participation will determine how well we handle quantum infrastructure and whatever comes after.
This means shifting from reactive to proactive infrastructure governance. Instead of approving facilities project by project and hoping offset mechanisms address cumulative impacts, we need integrated planning that considers technological infrastructure as critical social infrastructure requiring careful stewardship.
The alternative is repeating our current pattern at ever-larger scales: technological capabilities that transform economic and social systems whilst the governance structures managing their infrastructure remain stuck in frameworks designed for individual industrial projects.
The quantum revolution will test whether we've learned to govern transformative technology responsibly or whether we'll continue externalising the costs of technological progress onto communities with the least power to resist.
Right now, we're practicing with AI infrastructure. The final exam will be quantum computing. And the communities hosting these facilities—whether they're asked for permission or not—will be grading our work.
Everyone Has Something Right
Technology companies are right that AI infrastructure enables remarkable capabilities that could benefit society enormously. Local communities are right that they shouldn't bear disproportionate costs for private technological development. Environmental advocates are right that resource-intensive infrastructure must account for climate impacts. Economic development officials are right that technological infrastructure can drive regional prosperity.
No one has everything right about how to balance these competing concerns. But responsible scaling beyond the model requires governance frameworks sophisticated enough to address all of them simultaneously rather than optimising for any single priority.
The infrastructure decisions we make today will determine whether technological revolutions enhance human flourishing or simply concentrate power and externalise costs. Getting this right for AI infrastructure isn't just about data centers—it's about proving we can govern transformative technology responsibly when the stakes get even higher.