The race to build artificial intelligence infrastructure is colliding with local resistance, but blocking data centers may backfire in ways communities don't expect. Rather than preventing AI development, restrictions simply shift where compute gets built and who benefits from it, according to policy experts and industry leaders analyzing the infrastructure crunch driving the AI boom.

The fundamental tension is straightforward: AI systems require enormous amounts of electricity and cooling capacity. Communities across the United States and Europe are increasingly saying no to data center projects, citing concerns about power grid strain, water consumption, and environmental impact. But this resistance may be engineering the opposite outcome from what opponents intend.

Why Blocking Data Centers Concentrates AI Power?

If a county rejects a data center project, it doesn't disappear. The facility gets built in the next county, the next state, or on federal lands. If regional deployment slows significantly, hyperscalers (the massive cloud companies like Amazon, Microsoft, and Google) build on the ocean floor, in the Gulf, or eventually in orbit. Microsoft has already prototyped subsea data centers, and SpaceX is openly designing for compute in space.

The consequence of this geographic arbitrage is predictable: compute becomes scarce, and scarce resources flow to the highest bidder. That means the Department of Defense, major cloud hyperscalers, and the largest financial institutions get priority access. The middle of the economy, where innovation often happens, gets crowded out.

"Restrict data center construction and two things happen, both predictable. First, compute becomes scarce and scarce compute does not flow to the people who need it most. It flows to the people who can pay the most: the Department of Defense, the major cloud hyperscalers, the largest financial institutions on the planet," explained Stefan Bauschard, a policy analyst focused on AI infrastructure.

Stefan Bauschard, Policy Analyst

The second effect is equally damaging: restrictions raise prices. When supply tightens, costs increase. Hyperscalers absorb these costs and pass them along. Everyone else bears the burden. A small business trying to access frontier AI models, an independent researcher, a regional hospital, or a school district cannot outbid Lockheed Martin or the Pentagon.

What Does Intelligence Rationing Look Like?

The parallel to electricity is instructive. Imagine if in 1925, worried about coal pollution (a legitimate concern), the United States had decided to cap power plant construction until cleaner alternatives matured. Electricity would not have stopped. It would have been rationed to the wealthy and well-connected. Rural electrification, the single most economically equalizing infrastructure project of the 20th century, would never have happened. The middle class would not have been built.

AI is following the same trajectory, but the stakes are higher. Education is already a scarcity machine. Only so many students can attend Stanford or Exeter, and society has built an entire social hierarchy around that competition. The richest universities and most expensive private schools will pair their students with access to the most capable frontier AI models. Everyone else will access whatever the free tier allows when it is not compute-constrained.

The result is a widening intelligence gap. A nurse in rural Ohio, a teacher in West Texas, a high school senior who has never met anyone who went to college, and a small business owner in Detroit will not have equal access to AI tools that could amplify their capabilities. Outcomes diverge. Economic mobility collapses. The story society tells itself about merit becomes even thinner.

How Communities Can Negotiate Better Data Center Deals

• Secure a durable tax base: Data centers are among the few pieces of physical capital in the modern economy that cannot be relocated when tax rates change. A 200-megawatt facility with billions of dollars of fiber, substations, and cooling infrastructure sunk into the ground stays put. Communities can lock in a reliable revenue stream to fund schools, services, retraining programs, or partial universal basic income if AI disrupts employment.
• Negotiate infrastructure commitments: Rather than blocking projects, communities can demand that hyperscalers invest in local grid upgrades, renewable energy partnerships, and water management systems. These commitments improve the entire region's infrastructure, not just the data center.
• Establish workforce development agreements: Data centers create direct jobs, but communities can negotiate for training programs that prepare local workers for high-wage positions in facility operations, maintenance, and management.

The leverage local communities actually have is not the power to prevent construction. It is the power to shape it. To decide, while there is still time, what a host community receives in return.

Why Is Ireland Losing the Data Center Race?

Europe's data center industry has traditionally centered around Frankfurt, London, Amsterdam, Paris, and Dublin, known as the Flap-D markets. These cities attracted investment because they combined major internet exchange points, strong demand from finance and tech sectors, excellent connectivity, and stable regulatory environments.

But Ireland is falling behind despite having 14 Amazon Web Services data centers. The culprit is electricity cost. Ireland has the highest electricity prices in Europe by a significant margin, making it uncompetitive for energy-intensive AI workloads.

"The planning system needs to deliver, and better supporting infrastructure was required," stated Niamh Gallagher, country lead for Amazon Web Services.

Niamh Gallagher, Country Lead for Amazon Web Services

Amazon CEO Andy Jassy announced in February 2026 plans to raise overall investment to $200 billion for the year, doubling the 2025 forecast, with the majority earmarked for expanding AWS data center capacity to meet AI demand. But as of March 2026, Amazon announced its biggest European Union investment would be in Spain, not Ireland. The 33.7 billion euro investment in Spain is geared toward transforming Spain into a European hub for AI.

Competition for data center investment is intensifying in the Nordic region, where OpenAI, Microsoft, and Google are seeking locations for energy-hungry AI workloads. Norway, Sweden, Finland, Denmark, and Iceland offer a combination of low-cost energy, 100 percent renewable energy, a cool climate, and high operational reliability.

The Irish government has approved a landmark investment package of up to 18.9 billion euros in the national electricity grid and network for 2026 through 2030, aimed at reducing energy costs and ensuring supply certainty. The Large Energy Action Plan also targets support for energy-intensive industries including semiconductors, pharmaceuticals, and precision engineering. But industry executives at the Bord na Mona energy conference questioned whether these actions would arrive in time.

What Happens If Communities Continue to Block Data Centers?

The geopolitical incentives, capital flows, and national security logic all point toward continued data center buildout. The Pentagon is not going to wait on a county zoning board. If the United States slows domestic deployment through local resistance, compute infrastructure gets built overseas, on federal lands, or in orbit.

The result is a future where superintelligence arrives wearing a uniform or a corporate logo, with no third tier of academic, civic, or public infrastructure for everyone else. The alternative is to accept that data centers are getting built somewhere regardless of local opposition, and use that reality to negotiate the terms that benefit communities most.

Compute is the new electricity. That is no longer a metaphor; it is an operating fact of the economy. The question on the table is not whether the largest AI systems in human history get built. They are getting built. The only live questions are where, for whom, and whether enough compute exists to benefit everyone.