The gatherings have begun to mirror one another: from rural Virginia to San Marcos, Texas, locals warn that data centers are “gobbling up” electricity and pushing utility bills higher. In Indiana, lawmakers consider moratoria to prevent hyperscale campuses from landing near cornfields. So far, at least fifteen counties have adopted such measures.
The rhetoric is strikingly uniform: Data centers are “vampires.” They are “stealing” energy. They power chatbots while ordinary ratepayers pay the bill.
In Tallahassee, Florida Gov. Ron DeSantis stated it plainly: “I don’t think there are many people who want higher energy bills simply so some chatbot can corrupt a 13-year-old online. That’s not what anyone signed up for.” The framing is revealing. The issue isn’t merely cost; it concerns moral worth. Between the lines rests a quiet assumption: electricity is scarce, and some uses deserve it less.
Two decades ago, faced with rising energy prices, much of rural America clung to a different reflex: Drill, baby, drill. The slogan was blunt, but the instinct was clear. When demand climbs, you increase supply. You build. Nearly a century earlier, as the world slid toward war, American leaders and industrialists made the same wager: that production could outrun panic.
Today, confronted with growing electricity demand, many of those same communities have abandoned that wager. Panic takes hold. Hit the brakes. Treat today’s maximum as tomorrow’s limit … and start deciding who must live within it.
How did that happen?
Bills rose. Citizens grew angry. And politicians picked the most expedient target: data centers and the companies building them, which are conspicuous, unpopular, and easier to blame than the grid itself.
Increasingly, even the companies themselves have learned to play along, apologizing, signing pledges, and performing public contrition for the energy their own infrastructure requires. The ritual satisfies the moment. It does nothing to produce a single additional megawatt.
It is easier to blame chatbots and Silicon Valley than to admit that decades of policy have constrained supply and made scarcity feel permanent.
Scarcity isn’t permanent. Capacity is handcuffed.
Electricity demand is rising again after more than a decade of relative stagnation. From 2008 to 2021, U.S. electricity demand grew at an average pace of about 0.1 percent per year. Then the pattern changed. In 2024, demand surged nearly 3 percent, adding 128 terawatt-hours, one of the strongest growth years this century. The trend has continued, with generation reaching a record 4.43 trillion kilowatt-hours in 2025, up another 2.8 percent. And this is only the beginning. Many forecasts now project U.S. electricity demand increasing by 20 to 25 percent by 2030, driven largely by data centers, artificial intelligence, and a broader push toward electrification. For years, utilities planned for flat growth. Suddenly, projections are moving upward.
That may feel like a crisis to some policymakers, but rising demand is not novel. It is the recurring feature of every transformative industry.
In 1973, when OPEC’s oil embargo sent prices soaring and gas lines stretched for blocks, the prevailing wisdom held that America had entered an enduring era of scarcity. President Richard Nixon imposed price controls. Congress established the Strategic Petroleum Reserve. Politicians preached conservation. Domestic production was treated as marginal.
Yet between 2008 and 2018, U.S. crude oil production more than doubled—from roughly 5 million barrels per day to more than 12 million—driven largely by private innovation in horizontal drilling and hydraulic fracturing. By 2019, the United States stood as the world’s largest producer of petroleum and natural gas. Scarcity had not been geological. It had been technological and regulatory.
The same panic greeted the commercial internet. In the late 1990s and early 2000s, bandwidth constraints were treated as looming bottlenecks. During the net neutrality battles of the 2010s, we were warned that streaming video would overwhelm networks without aggressive federal oversight.
Instead, fiber deployment accelerated. Wireless technology leaped from 3G to 4G to 5G. U.S. internet traffic increased more than twentyfold between 2010 and 2020. Netflix did not collapse the grid. Capacity expanded.
Go back further.
In the late 19th century, as whale oil declined and cities swelled, many watchers feared a lighting crisis. Electricity generation seemed speculative. Within a few decades, power plants dotted the country and electric light became common in American homes. What appeared to be a ceiling proved to be a transition.
Even during World War II, when resources were truly constrained, the dominant national response was expansion. As historian Arthur Herman documents in his book Freedom’s Forge, Washington often misjudged timelines and misallocated inputs. Yet American industry outpaced bureaucratic hesitation. Between 1940 and 1944, U.S. aircraft production increased more than fifteenfold. Shipyards that had never built a Liberty vessel—ships built under an emergency shipbuilding program—launched them at astonishing speed. Industrial capacity multiplied at a rate that would seem improbable under today’s permitting regimes. Fixed-pie arguments existed then, too. Steel was scarce. Fuel was scarce. Rubber was scarce. But the louder voice in the room belonged to those who believed production could outrun constraint.
The pattern repeats: Demand surges, prices rise, panic follows, politicians promise control, production expands.
Which brings us back to electricity.
Across the United States today, more than 2,600 gigawatts of proposed generation capacity sit in interconnection queues—solar, gas, storage, nuclear—awaiting studies, approvals, and transmission upgrades. The entire existing U.S. generation capacity stands at just over 1,200 gigawatts.
In other words, the country has nearly twice its current capacity waiting in line, trapped in bureaucracy and permitting backlogs.
The fight over data centers, then, is not really about servers humming in anonymous warehouses. It concerns whether rising demand should trigger new capacity or a new referee to decide who gets to use it. Artificial intelligence sits at the center of this debate because it magnifies both load growth and cultural resentment. The image is easy to dislike: deepfakes, essay bots, algorithmic feeds engineered to keep teenagers scrolling at 2 a.m.
Some of those concerns are real. Large facilities can strain local infrastructure, raise questions about water use, and create friction in communities unaccustomed to industrial-scale development. But they are increasingly woven into a broader narrative that casts demand itself as the problem, rather than the system’s failure to build.
Yet the same computational infrastructure now treated as indulgent underwrites pharmaceutical research—where AlphaFold has predicted the structures of more than 200 million proteins, cutting drug-discovery timelines that once spanned years—as well as logistics optimization, grid-management software, advanced manufacturing simulations, and medical diagnostics.
Right now, AI systems are flagging tiny abnormalities on mammograms that even seasoned radiologists can miss. They are catching cancers earlier, reducing false alarms, giving patients options they would not have had a decade ago. Machine-learning models are narrowing thousands of drug candidates down to a handful worth testing, compressing timelines that once stretched for years. The difference is measured in earlier diagnoses and lives saved or extended.
Those breakthroughs are powered by electricity and by our ability to produce a lot more of it.
And artificial intelligence is not merely a consumer novelty. It is a strategic capability. Nations that can scale electricity and computing power will shape military logistics, advanced manufacturing, financial systems, and intelligence analysis.
Meanwhile, China is building generation, transmission, and data infrastructure in tandem, often with timelines measured in months, whereas ours stretch into years because of interconnection queues, permitting delays, and local opposition. There, computing power is treated less as a moral question and more as a strategic resource to be provisioned.
This global race for AI dominance through abundant, scalable power and computation is already underway. Yet much of our domestic debate fixates on scarcity as if the pie were fixed forever, which leads to moralized calls about who gets to use the limited resources we have. Once electricity becomes a hierarchy of virtue, you may not like where you land.
This reflex is bipartisan. On the right, suspicion centers on coastal tech elites siphoning power from “real Americans.” On the left, it gathers around corporate excess and environmental harm. When trust in builders erodes, it does not leave a vacuum: Gatekeepers step forward. A pause. A moratorium. A pledge to shield families from footing the bill for someone else’s servers. In truth, it amounts to giving someone else permission to decide which electrons are righteous and which are waste. Once you declare the pie cannot grow, a authority must divide it, seeking villains, assigning virtue, declaring some uses essential and others indulgent.
Those who once cried “Drill, baby, drill” but today seek to curb AI’s growth may think they are defending markets or protecting their households. They may not notice the pivot from “Build more!” to “Stop them!” That is the turning point.
One path expands capacity and yields economic and strategic leverage. The other erects a gate. This is a choice about the kind of country we are becoming—one that builds when demand rises, or one that decides who must do without.
