For most of the digital age, data centers were invisible infrastructure — windowless buildings on the edge of town that quietly stored emails, ran payroll software, and later, streamed movies. Few people thought about them, and almost no one protested them. That has changed. In 2026, data centers have become one of the most contentious local and political issues in the United States, showing up in zoning board meetings, statehouses, and even midterm campaign ads. To understand why a warehouse full of computers has become a flashpoint, it helps to understand where these facilities came from — and how a decades-long story of quiet, incremental growth turned into an AI-driven building boom that is straining power grids, water supplies, and public patience all at once.
The Origins: Room-Sized Computers (1940s–1960s)
The data center’s ancestor wasn’t a facility — it was a single machine. Early computers like ENIAC (1945) and UNIVAC (1951) were so large, power-hungry, and finicky about temperature and humidity that they effectively required their own dedicated rooms, staffed by technicians who monitored vacuum tubes and cooling systems around the clock. There was no separation yet between “computer” and “computer room” — the machine and its environment were one and the same.
By the 1960s, as businesses and government agencies adopted mainframe computers from companies like IBM, the modern concept of a computer room began to take shape: raised floors to hide cabling and route cold air, dedicated air conditioning, and physical security to protect an asset that often represented a company’s entire technological investment.
The Corporate Data Center Takes Shape (1970s–1990s)
As computing spread from a handful of mainframes to distributed minicomputers and then personal computers, organizations needed a place to consolidate their growing IT operations. The term “data center” entered common usage during this period, describing an internal facility — typically in a company’s own basement or a nearby office building — that housed servers, storage, and networking equipment for that single organization.
The 1990s brought two forces that reshaped the field. First, the rise of client-server computing and enterprise software (email systems, databases, ERP platforms) meant nearly every mid-sized business needed serious server infrastructure. Second, and more dramatically, the commercial internet arrived. The dot-com boom created explosive demand for web hosting, and a new industry — the colocation provider — emerged to meet it. Companies like Equinix and Digital Realty built large, specialized facilities where multiple businesses could rent secure space, power, and cooling for their own servers, rather than building out costly infrastructure themselves.
The Cloud Era and the Rise of the Hyperscaler (2000s–2010s)
The dot-com crash of 2000–2001 temporarily emptied a lot of that new data center capacity, but it also set the stage for the next transformation. Companies that survived — Amazon, Google, Microsoft — had built enormous internal computing infrastructure to run their own services. Amazon’s 2006 launch of Amazon Web Services turned that spare capacity into a business: instead of every company building its own data center, they could simply rent computing power over the internet.
This kicked off the “hyperscale” era. Rather than the modest, general-purpose facilities of the colocation age, companies like Google, Amazon, Microsoft, and later Meta began constructing massive, purpose-built campuses — some spanning millions of square feet — designed to run their own cloud, search, social media, and streaming operations at global scale. These facilities pioneered new efficiency techniques (like Google’s advanced cooling systems and outside-air cooling in cold climates) and increasingly ran on long-term renewable energy contracts, as tech companies raced to make public sustainability commitments.
By the late 2010s, data centers had become deeply embedded, if largely unnoticed, physical infrastructure. Northern Virginia’s “Data Center Alley” outside Washington, D.C. emerged as the world’s largest concentration of these facilities, drawn by proximity to internet backbone infrastructure, favorable tax policy, and reliable power. Cloud computing was mainstream, but the public conversation about it remained mostly about privacy and cybersecurity — not electricity or water.
The Inflection Point: Artificial Intelligence (2020s)
Everything changed with the generative AI boom that began in earnest around 2022–2023. Training and running large AI models requires enormously more computing power — and therefore electricity — than the web-hosting and cloud-storage workloads that came before it. A new category of facility emerged: the AI-optimized data center, packed with specialized chips (GPUs) that draw far more power per rack and generate far more heat than traditional servers, often requiring liquid cooling instead of simple air conditioning.
Nvidia CEO Jensen Huang has captured this shift by describing these new facilities as “AI factories” rather than simple storage sites — places where, in his words, computing itself is being manufactured at industrial scale. The numbers reflect that shift: there are now more than 4,000 data centers operating in the United States, concentrated heavily in Virginia, Texas, and California, with roughly 3,000 more under construction or in planning. Developers have announced roughly 190 gigawatts of future data center capacity since 2024 — an enormous figure that dwarfs anything the industry built in its first seven decades combined.
Why the Debate Has Exploded in 2026
This is the context for today’s backlash. A facility type that operated for decades with minimal public scrutiny is now colliding with three hard limits: electricity supply, water supply, and community tolerance.
Electricity and utility bills. Data centers already account for an estimated 2 to 3 percent of total U.S. electricity consumption, and forecasts suggest that figure could reach 10 to 15 percent within a few years as AI-driven demand accelerates. The U.S. Energy Information Administration expects national electricity consumption to set new records in both 2026 and 2027, driven substantially by data center growth. Because the electric grid is a shared system, that new demand doesn’t stay contained to the tech companies causing it — utilities are requesting billions of dollars in rate increases to fund new generation and transmission, and residents in data-center-heavy regions have reported sharp jumps in their own power bills. Analysts note that overall power bills have risen roughly 40 percent since 2021, a trend with multiple causes, but one that data centers’ growing appetite for electricity has intensified and politicized.
Water use. Many large data centers rely on evaporative cooling systems that can consume hundreds of thousands to millions of gallons of water per day, drawing scrutiny in regions already facing drought or water-stressed municipal systems.
Local disruption and jobs. Communities have raised concerns about construction noise, the constant hum of industrial cooling equipment, diesel backup generators, and the visual impact of large, windowless buildings — often described by critics as out of scale with rural or suburban surroundings. Developers frequently promote job creation as a benefit, but critics counter that construction jobs are temporary (often lasting only a year or two), while a completed, highly automated facility may employ only a few dozen to a couple hundred permanent staff — not enough, critics argue, to offset the tax incentives many local governments offer to attract these projects.
The politics. These concerns have translated into remarkably fast-moving political backlash. Polling in 2026 has found that around 70 percent of Americans oppose construction of a new data center in their own area — a level of opposition that, according to one survey, now exceeds public resistance to building a new nuclear power plant nearby, a striking reversal given nuclear power’s long history as a lightning rod for local opposition. Data Center Watch, a research group tracking the issue, found that roughly 75 U.S. data center projects worth an estimated $130 billion were delayed or blocked by local opposition in just the first three months of 2026 alone — comparable to the total blocked for all of 2025. The number of active grassroots opposition groups more than doubled, from 396 at the end of 2025 to 833 by March 2026, spread across 49 states. By mid-2026, at least 69 local governments had enacted bans or moratoriums, and state legislatures introduced hundreds of bills addressing data center siting, taxation, and utility cost allocation. The issue has also become notably bipartisan: in Michigan, a Democratic Senate candidate has called for federal guardrails on new data center approvals, while in Ohio, a Republican gubernatorial candidate has pushed a “data centers first” pledge demanding new state oversight.
The Industry’s Response
Facing this pressure, data center developers and hyperscale operators have begun adjusting their strategies rather than simply pushing forward as before. A growing share of new capacity is being built with dedicated, “behind the meter” on-site power generation — often natural gas turbines or reciprocating engines from manufacturers like Caterpillar and Wärtsilä — specifically so that new facilities don’t add strain to the public grid or compete with residential customers for electricity. Some companies have also turned to acquiring renewable energy developers directly, or negotiating deals to help extend the life of existing power plants, including coal facilities in several states, in order to secure supply. Meanwhile, industry groups continue to argue that the economic activity, investment, and long-term tax revenue these projects generate outweigh the costs — while pointing out that the public conversation sometimes conflates energy issues that predate the AI boom, like aging grid infrastructure, with new impacts specific to data centers.
Where This Leaves the Country
The data center has traveled a long way from the room-sized machines of the 1940s to today’s gigawatt-scale AI campuses — but the current debate is really a debate about growth itself: whether the country can build the physical infrastructure that AI and cloud computing require fast enough, without pricing out or burdening the communities asked to host it. Unlike past waves of data center construction, which mostly proceeded quietly because the facilities were smaller and their impact was diffuse, the AI-driven scale of expansion has made trade-offs — over electricity, water, and land use — impossible to ignore. With the issue now surfacing in state legislatures, local ballot measures, and national elections, how that trade-off gets resolved is likely to shape not just where the next generation of data centers gets built, but how much AI computing the country is ultimately able to support.
About the Author
Get real, Lindsey! Anywhere one goes today, one sees people staring at the palm of their hand while frantically pawing…
As I said, your lack of discretion with your addiction is alarming. You certainly are addicted to that plastic crack…
The reason you would know that people stare at their cell phones 45-50 minutes out of every hour is that…
Data centers are going to increase electric bills for everyone because their high electricity demand leads to infrastructure costs that…
Very good! Ive been saying all along. Data Centers are not good or bad, Republican or Democrat. They are however…