For years, artificial intelligence existed for most people as the substance of science fiction. A rogue computer or robot in a movie. And in real life, perhaps a chatbot, or a distant conversation unfolding somewhere between Silicon Valley boardrooms, research labs, and investor calls.
But now, across the United States and around the world, a new wave of AI factories, hyperscale data centers, semiconductor facilities, and energy-intensive infrastructure projects are beginning to reshape the landscapes around them. Rural towns, secondary cities, manufacturing corridors, and energy-abundant regions are increasingly finding themselves at the center of a global race to build the next era of intelligent systems.
And with that expansion has come resistance.
For decades, much of the technology sector operated within its own ecosystem while assuming the public would eventually adapt downstream. But now the abstraction layer is disappearing.
And, while most people, including policymakers, still don’t understand what artificial intelligence is, when they say they don’t want “it”, or even on the other hand, when choosing to utilize or champion AI-powered and native tools, they do understand when their utility bills rise. They understand when landscapes change. They understand when outside institutions arrive promising progress while offering little participation in shaping what that progress ultimately means for their communities.
People are asking difficult questions about water usage, land acquisition, utility strain, environmental impact, tax incentives, energy consumption, and who ultimately benefits from the transformation unfolding around them—or more accurately, to them.
Much of the technology industry has framed this resistance as fear, ignorance, or obstructionism. But the growing public unease surrounding AI infrastructure did not emerge overnight, nor is it simply pushback against innovation itself.
And as governments and corporations race to compete in what is increasingly framed as an existential technological contest, particularly against China, an uncomfortable question is beginning to emerge beneath the surface:
What does responsible and sustainable progress actually look like in the age of AI?
The Trust Gap Did Not Begin With AI
The growing public resistance surrounding AI infrastructure reaches back decades.
For much of the modern technology era, the industry operated inside increasingly concentrated ecosystems of capital, talent, influence, and cultural power. Founders spoke to investors. Engineers spoke to engineers. Silicon Valley evolved into its own language, value system, and economic engine. Innovation became both philosophy and justification. Speed became virtue. Scale became power.
At the same time, many communities outside the technology sector experienced a very different reality. Manufacturing jobs disappeared, and entire regions were reshaped by offshoring, automation, and economic consolidation. The digital economy created extraordinary wealth, but much of that wealth accumulated unevenly and often far from the communities absorbing the disruption.
In 1990, the United States produced roughly 37% of the world’s semiconductors. Today, it accounts for closer to 12% of global production.1
America led the semiconductor industry from its inception. Yet, the very jobs that could have brought large portions of the public into the innovation economy, making technological progress feel tangible, participatory, and economically beneficial, were steadily moved offshore. The center of gravity shifted abroad.
The result is a paradoxical moment in American technological history.
The country is now racing to rebuild strategic manufacturing capacity, AI infrastructure, and industrial resilience under immense geopolitical pressure while asking communities to support a technological transition that, for many people, feels increasingly inaccessible, centralized, and in many cases completely resistant to meaningful public participation.
Historically, major industrial transformations were far more visible to the public. During the Industrial Revolution, people could physically see factories rise around them, understand the labor being created, and directly participate in the economic systems transforming society, even amid enormous hardship and inequality. The family could take pride in riding in the Cadillac Dad helped build. Steelworkers could see their labor rise into bridges and skylines. Boeing workers could watch planes assembled in their communities connect the world.
Even earlier phases of the internet era often felt participatory to the public. People could build websites, launch businesses, create content, develop apps, and directly engage with the emerging digital economy.
The infrastructure supporting the internet also remained largely invisible to most people. Few citizens needed to think about server farms, energy load, semiconductor supply chains, or industrial-scale compute infrastructure in their daily lives. The digital economy arrived primarily through personal devices, software, platforms, and connectivity that often required little more visible infrastructure to the average household than a cable line, modem, router, or Ethernet port.
Now the infrastructure powering the AI era is becoming impossible to ignore. It feels fundamentally different: more industrial, more centralized, more capital-intensive, more intrusive, and increasingly shaped by a relatively small number of companies and institutions.
AI data centers require enormous amounts of land, water, energy, permitting, and public cooperation. Entire communities are being asked to absorb the physical demands of a technological transformation that many feel they had little role in shaping and don’t understand.
That context matters.
Public resistance surrounding AI factories and hyperscale infrastructure projects is being shaped by accumulated distrust, economic displacement, institutional fatigue, and growing public skepticism toward industries that have spent decades operating at an extraordinary scale while remaining culturally and structurally distant from much of the broader public.
The False Binary: Build or Be Left Behind
The urgency surrounding artificial intelligence has increasingly produced a narrow and deeply binary public narrative.
Build faster. Approve faster. Scale faster. Compete with China. Attract hyperscalers. Secure infrastructure investment. Win the future.
And beneath much of the messaging sits an increasingly implicit warning: regions, cities, and states that hesitate risk being left behind economically, technologically, and geopolitically.
For many public officials, the pressure is immense.
Doing the Math
AI infrastructure projects often arrive carrying the promise of jobs, investment, national relevance, and participation in what many believe will become one of the largest economic transformations in modern history. Few leaders want to be remembered as the ones who turned opportunity away.
But the economic equation surrounding AI infrastructure is often far more complicated than the narrative being presented to both policymakers and the public.
Tax incentive packages designed to attract hyperscale data centers can result in billions in forgone public revenue annually—sales tax exemptions on hardware, property tax abatements, and infrastructure subsidies. Studies have repeatedly questioned whether the long-term public return justifies the scale.
The imbalance becomes even more pronounced when considering employment. Despite enormous physical footprints and infrastructure demands, hyperscale AI data centers employ relatively few permanent workers once construction ends.
When corporate property and sales taxes are significantly reduced, municipalities face growing pressure to fund schools, transportation, emergency services, and infrastructure. In some cases, the burden shifts to small businesses and homeowners through higher taxes, utility costs, or budget strain.
And most taxpayers have little visibility into any of it. The complexity and opacity of incentive agreements make it difficult for communities to evaluate long-term fiscal impact—or hold developers and policymakers accountable.
Then there is the question of energy.
AI servers are ravenous—for power, for cooling, for the transmission lines, substations, transformers, and water systems required to keep them alive. Utilities are expanding at scale to meet the demand. Residents are footing the bill.
In high-growth corridors, some residents are already seeing electricity costs rival monthly mortgage payments. In PJM Interconnection, the largest electricity market in the country, wholesale electricity prices jumped 76% year-over-year, from roughly $78 to $137 per MWh, driven directly by data center demand. A federal watchdog warned that those price impacts are not reversible.2
The concerns don’t stop at economics.
In Morrow County, Oregon, investigative reporting linked data center expansion to allegations of industrial wastewater practices, groundwater contamination, and clusters of rare cancers, miscarriages, and kidney-related illnesses. Amazon later paid a $20.5 million settlement related to wastewater discharge violations tied to operations in the region.3 The echoes of DuPont are hard to ignore.
Not every public concern will prove fully justified. But communities are increasingly being asked to absorb industrial-scale infrastructure whose long-term environmental, economic, and societal consequences are still unfolding in real time—with limited recourse and limited transparency.
History is instructive here. When populations feel excluded from major economic transformations, resistance rarely stays contained to public comment periods or town hall meetings. It escalates into lawsuits, permitting delays, political upheaval, and in more extreme cases, civic unrest that carries its own economic toll. The 1992 Los Angeles riots caused an estimated $1 billion in property damage across roughly 1,100 buildings.
Although that riot was coupled with the horrific beating of Rodney King by Los Angeles law enforcement, it and several other historic riots serve as reminders that human frustration tied to perceived inequality and institutional disregard has always had a price. In a worst-case scenario, it may be the very municipalities that championed the projects, approved the incentives, absorbed the infrastructure, and staked their credibility on the promise of progress that could end up paying for it.
Weighing the Options
The current AI infrastructure race is often framed as a binary: approve hyperscale expansion as quickly as possible, or risk economic irrelevance.
But technological progress has never depended on a single implementation model.
A region seeking to participate meaningfully in the AI economy may benefit from asking broader and more foundational questions.
Does long-term resilience come primarily from hosting large data centers alone?
Could a more diversified approach ultimately create stronger local economies, broader public participation, and more durable public trust?
Could investment also be directed toward semiconductor manufacturing, energy modernization, workforce training, cooling innovation, grid infrastructure, research partnerships, technical education pipelines, sustainable industrial development, and locally integrated technology ecosystems?
These are not consolation prizes for communities that missed the hyperscale wave. They are, in many cases, the more durable path.
A city or town evaluating a major AI infrastructure project is not simply reviewing a construction deal. It is making decisions that could reshape local identity, public resources, infrastructure priorities, and quality of life for decades. That deserves more than a fast yes driven by fear of being left behind.
The regions that navigate this transition most successfully will be those that asked harder questions early.
What China Actually Built
Much of the public conversation surrounding artificial intelligence has increasingly reduced the relationship between the United States and China to the language of a technological arms race.
Who will build the most powerful models first?
Who will dominate compute?
Who will control the future of artificial intelligence?
But the more instructive question is not who is moving fastest, but rather who is building most deliberately.
What China spent decades building was not simply technological capability, but coordinated industrial capacity on an extraordinary scale. The country concentrated large portions of the world’s manufacturing across critical industries while simultaneously investing heavily in transportation infrastructure, energy systems, industrial corridors, semiconductor ambitions, supply chain integration, research institutions, workforce development, and state-supported technology ecosystems tied together through long-horizon strategic planning.
The result was not simply economic scale. It was industrial coordination.
Entire regional ecosystems were intentionally developed around specialized production capabilities and evolved together rather than independently.
And while many Western technology conversations focused heavily on software, applications, platforms, and consumer growth, China continued investing aggressively in the physical foundations required to support long-term industrial and technological power.
That distinction matters enormously in the AI era because artificial intelligence does not operate independently from the physical world.
The lesson is not to copy China. But there is something worth understanding: China did not win ground in AI by moving fast and hoping infrastructure would follow. It built the infrastructure first, strategically, and at every layer of the stack.
AI capability emerges from deeply integrated infrastructure ecosystems, not isolated breakthroughs.
And that reality complicates the increasingly simplistic narrative that any nation can compete primarily by accelerating approvals, concentrating more hyperscale infrastructure, or framing public resistance as an obstacle to national progress.
Long-term technological leadership requires more than urgency. It requires industrial strategy, infrastructure resilience, public trust, workforce development, energy planning, manufacturing capacity, and social cohesion capable of sustaining transformation over decades rather than election cycles or quarterly earnings reports.
If the answer to China’s coordinated industrial strategy is simply “approve everything faster,” then the U.S. may win the speed contest while quietly losing something harder to rebuild.
Beyond Resistance: Building a Smarter AI Transition
The conversation surrounding AI infrastructure has spent considerable energy on the wrong question. Meanwhile, the reality is that artificial intelligence is becoming deeply embedded in the future of economies, industries, national competitiveness, healthcare, education, defense, logistics, energy systems, and everyday life. The infrastructure supporting that transition will continue expanding.
But expansion alone is not a strategy. And speed alone is not resilience.
The deeper challenge facing policymakers, companies, utilities, and communities is determining how to participate in the AI economy in ways that are economically durable, environmentally sustainable, publicly legitimate, and capable of creating meaningful pathways for broader societal participation.
Not every opportunity within the AI ecosystem carries the same infrastructure burden, environmental impact, workforce potential, or long-term economic return.
The table below does not argue against AI infrastructure. It argues for clarity. Communities evaluating their role in the AI transition deserve an honest accounting of what each pathway actually demands and what it is likely to return—not just in capital, but in jobs, public trust, environmental stewardship, and long-term resilience.
| Opportunity Area | Public Acceptance | Long-Term Economic Value | Workforce Participation | Infrastructure Burden† | Environmental Impact† | Strategic Importance |
|---|---|---|---|---|---|---|
| Hyperscale Data Centers | Low-Mixed | Medium | Low | Very High | Very High | High |
| Semiconductor Manufacturing | Medium | Very High | High | High | High | Very High |
| Grid Modernization | High | High | High | Medium | Medium-Positive | Very High |
| Workforce Training & Technical Education | Very High | High | Very High | Low | Low | High |
| AI Research Partnerships | Medium-High | High | Medium | Low | Low | High |
| Cooling & Energy Innovation | High | High | Medium | Medium | Positive | High |
| Local AI Startup & Business Ecosystems | High | High | Very High | Low | Low | Medium-High |
| Renewable Energy Integration | High | High | Medium | Medium | Positive | Very High |
A few patterns emerge immediately.
For years, much of the public conversation surrounding AI infrastructure has focused heavily on hyperscale data centers as the primary gateway into the future economy. Yet data centers represent only one layer of a much broader industrial and technological ecosystem. They carry the highest infrastructure burden and environmental footprint of any category examined, while delivering medium to long-term economic value and the lowest workforce participation of any option on the table. That combination should give any community pause before treating hyperscale hosting as the default definition of AI participation.
The more beneficial long-term opportunities for many regions may emerge from diversified participation in semiconductor manufacturing, energy modernization, workforce development, cooling innovation, grid infrastructure, technical education pipelines, AI research partnerships, and locally integrated business ecosystems capable of creating more durable economic participation over time.
In many cases, these investments may generate stronger public support precisely because communities can more clearly see where they fit within the transformation itself.
The lesson China learned through decades of industrial coordination, the U.S. is only beginning to absorb: the foundation matters more than the flagship. And the data bears this out in a striking way. Progress at the foundational level doesn’t just benefit communities, it transforms the data center row on this chart entirely.
China’s most significant environmental achievement in the data center sector illustrates this directly. Shanghai’s Lingang Underwater Data Center uses offshore wind and natural ocean cooling to slash cooling energy demands by over 90%4, achieving a Power Usage Effectiveness as low as 1.15—a figure that makes conventional land-based facilities look primitive by comparison. More than 95%5 of its electricity comes from offshore wind turbines. It uses zero freshwater. It occupies a fraction of the land. The environmental footprint that makes hyperscale data centers so politically and ecologically costly on land becomes a fundamentally different calculation when renewable energy integration, cooling innovation, and long-term infrastructure planning are treated as prerequisites rather than afterthoughts.
The frontier is extending further still. Orbital data centers operating in low Earth orbit are moving from theoretical to actively explored, promising natural thermal regulation, proximity to satellite infrastructure, and removal of land and water constraints entirely.
The implication is significant: communities that invest in the foundational layers don’t simply improve their own economic position. They help create the conditions under which data center development itself becomes more sustainable, more publicly defensible, and more strategically sound. The columns on the chart are not fixed. They are responsive to how deliberately the infrastructure beneath them is built.
The question is no longer whether to participate in the AI economy. It is how to participate on terms worth accepting.
The Future Needs Better Rooms
Artificial intelligence is often discussed as a technological transition.
In reality, it is also an infrastructural transition, an economic transition, an environmental transition, and a societal transition unfolding simultaneously.
The challenge is not simply building more compute. It’s building systems, institutions, and public conversations capable of holding the complexity surrounding what comes next.
For too long, the decisions shaping the future of technology have unfolded inside insulated ecosystems—boardrooms, investment circles, policy silos, and hyperscale negotiations conducted far from the communities expected to live with their consequences.
But societies function best when people feel meaningfully included in the future being constructed around them. The AI transition will require more than capital, energy, and infrastructure. It will require better rooms.
Rooms capable of holding disagreement without collapse. Rooms capable of balancing innovation with stewardship. Rooms where technologists, policymakers, utilities, labor leaders, environmental experts, educators, local communities, and citizens can participate in shaping what progress actually looks like together.
https://www.z2data.com/insights/who-are-the-top-us-compa,nies-in-the-semiconductor-industry
https://oregoncapitalchronicle.com/2026/03/31/amazon-to-pay-20-5-million-settlement-over-northeast-oregon-nitrate-pollution/
https://interestingengineering.com/innovation/china-offshore-wind-data-center
https://en.lingang.gov.cn/html/website/lg/English/CityofInnovation/InnovationProjects/1985965874467635201.html
Is your city ready to lead?
The ME on Main Street Civic AI Readiness Assessment helps city leaders, economic development directors, and civic planners evaluate where their community stands across six dimensions—AI literacy, infrastructure strategy, workforce alignment, environmental stewardship, public trust, and long-term positioning.
It’s free. It takes about 15 minutes. And it’s the first step toward building a participation strategy worth committing to.
