A nineteenth-century cartoon, “The Unrestrained Demon,” portrayed a terrifying sight: a skeletal monster racing through a city street, its skull-like head fixed beneath a glowing bulb and its body tangled in overhead wires. A horse lies collapsed in the road. Men are thrown backward. A mother faints before her child. The city looks under attack.
The image did not caution about war, plague or natural disaster. It was warning about the dangers of electricity. The image captured a familiar instinct: when a powerful new technology arrives, many people first see danger, disorder and loss of control.
Today, the monster is the data center, and Gov. Kathy Hochul has cast herself in the role of hero, arriving just in time to rescue New York by enacting the nation’s first statewide moratorium on new hyperscale facilities.
To critics like Hochul, these facilities look like dangerous, alien intrusions on ordinary life: oversized warehouses consuming land, electricity and water for the benefit of large technology companies. But they are not speculative indulgences, according to work done by the Goldwater Institute. They are the physical infrastructure of the digital economy. They support the modern world: AI tools, financial transactions, medical records, supply chains, classrooms, defense systems, cloud backups and the everyday digital services Americans rely on.
The digital economy may feel weightless, but it rests on steel, concrete, fiber, servers, cooling systems and electric power. We can pretend otherwise only because the system usually works too well to notice.
That is why the current debate is so often backward. Critics tally the costs of AI infrastructure while ignoring the value it creates.
The recent rise of artificial intelligence makes that infrastructure even more important. AI is not merely another consumer product or workplace convenience. It is a general-purpose technology that can expand access to knowledge, reduce the cost of complex tasks and give individuals and small firms capabilities that once belonged only to large institutions. That kind of progress requires a physical foundation.
None of this means concerns about electricity, water, land use or local communities should be dismissed. They should be answered – with evidence, markets and clear rules, not moratoriums, retroactive zoning changes and political vetoes dressed up as planning.
The grid challenge is real. But the answer is not to stop data centers. It is to build more energy, improve transmission, incentivize large users pay for the upgrades they require and use prices that reflect scarcity when the grid is under strain.
Data centers have become a convenient scapegoat for rising electricity prices, but the likelier culprit is years of policy choices that made power more expensive and less abundant. Reliable generation was retired before adequate replacements were ready. New supply has been slowed by permitting delays, regulatory barriers and green-energy mandates. A data center may expose those weaknesses, but it did not create them.
Electricity prices are not acts of nature; they are shaped by policy choices, generation constraints and market design. Blaming data centers for the consequences of Green New Deal-style energy mandates will not make power cheaper or more reliable. It will only give politicians an excuse to avoid fixing the system they helped break.
The same is true of water. Modern data centers are increasingly designed to reduce water use through air cooling, closed-loop systems and reclaimed or non-potable sources. Water is not free to operators. It is a cost, and markets create powerful incentives to conserve it. Political rationing is a crude substitute for innovation.
A serious policy framework should begin with a simple presumption: America should welcome the infrastructure of the future unless there is evidence of concrete harm. Projects that meet objective standards should receive predictable approvals. Large-load users should pay their own way. Energy policy should favor abundance and reliability. Privacy and constitutional protections should follow Americans into the digital age.
The alternative is a new Luddism, one that treats every emerging technology as guilty until proven harmless and every infrastructure project as a threat until politics permits it to exist. That path will not stop AI, cloud computing or the digital economy. It will simply move the investment, jobs and infrastructure somewhere else.
Europe offers the warning. For years, much of the continent has approached technology through a politics of suspicion: regulate first, permit slowly, ration energy, elevate process over production and treat scale itself as a danger. The result has not been a more humane technological future. It has been slower growth, fewer global technology champions and a growing dependence on innovations built elsewhere.
America should not copy that mistake. A nation that lets fear dictate its infrastructure policy will still use the tools of the future. It just will not own them, build them or profit from them.
The future will be built. The question is whether America will build it here.
