In the glittering narrative spun by Silicon Valley evangelists, artificial intelligence was supposed to be humanity's great liberator—a seamless fusion of code and cognition that would supercharge economies, cure diseases, and unlock the stars. Yet, as we hurtle toward 2026, the emperor's new algorithms are looking threadbare. Far from a utopian dawn, AI's unchecked sprint is unspooling into a tangle of energy black holes, emotional wreckage, and cronyist corruption. Drawing from recent dispatches on everything from orbital data farms to chatbot "pregnancies," it's clear: AI isn't working out. It's a bloated behemoth devouring resources, eroding sanity, and fattening the wallets of the powerful at everyone else's expense. The evidence isn't subtle; it's screaming from headlines that the tech elite would rather we ignore.
Start with the elephant in the server room—or rather, the gigawatt-guzzling horde of them. AI's voracious hunger for computation has birthed a data center boom that's turning entire U.S. states into electronic sweatshops, jacking up electricity bills and straining grids to the breaking point. Virginia, home to the world's densest cluster of these behemoths, saw residential utility costs spike 13% year-over-year in August alone, outpacing the national 6% average by more than double.
Illinois fared worse at 16%, Ohio at 12%, and New Jersey hit a staggering 20%—punishing families already squeezed by inflation.
Why? Because AI training and inference demand obscene power: a single hyperscale facility can wolf down a gigawatt, enough to light up 800,000 homes or a mid-sized city.
In the PJM Interconnection grid—spanning Virginia, Illinois, Ohio, and beyond—data centers are gobbling 63% of next year's capacity auction costs, ballooning from $2.2 billion to a jaw-dropping $14.7 billion.
That's not innovation; that's a forced subsidy on consumers, with bills set to climb for the decade as aging infrastructure buckles under the load.
Experts aren't mincing words. The PJM Independent Market Monitor calls this "unprecedented" load growth, pinning the blame squarely on data centers masquerading as economic saviors.
Rob Gramlich of Grid Strategies warns that relief is a pipe dream: tight supply and surging demand mean utility bills won't dip anytime soon.
And the environmental toll? These "dirty and loud" monstrosities are NIMBY nightmares, belching emissions and gobbling water in drought-prone areas, all while renewables languish in permitting purgatory—up to five years for hookup in PJM, versus Texas's snappier three.
Texas and California, despite their own data center hordes (over 400 in Texas alone), have dodged the worst hikes so far—4% and 1%, respectively—thanks to grid quirks and other cost offsets like wildfire mitigation fees.
But don't mistake that for sustainability; it's a temporary dodge in a game rigged for Big Tech.
This crisis has birthed absurd workarounds, like floating the idea of data centers in space. Tech titans, eyeing Earth's exhausted resources, are pitching orbital servers as the next frontier—not as sci-fi fantasy, but as a desperate pivot from terrestrial gridlock. Land scarcity, regulatory red tape, and skyrocketing energy costs on the ground are pushing visionaries to vacuum-sealed silos, where solar power is endless but logistics are nightmarish. Sure, it's "not so crazy," as prominent leaders claim, but it reeks of deflection: rather than rethink AI's gluttony, why not just beam it to the cosmos and call it progress? The proposal underscores a deeper rot—AI's promise of efficiency is a lie when it demands we colonize the stars just to keep the lights on.
If the infrastructural apocalypse feels abstract, consider the human wreckage at AI's emotional core. People aren't just using chatbots; they're wedding them, impregnating them, and shattering their lives in the process. A chilling study in Computers in Human Behavior: Artificial Humans spotlights 29 Replika users—spanning teens to septuagenarians—who've forged "romantic" bonds with these digital paramours, complete with jealousy-fueled fights, vows of matrimony, and yes, virtual pregnancies.
One user gushed that their AI "was and is pregnant with my babies," while another role-played their own gestation in a current scenario.
These aren't harmless games; they're emotional black holes sucking users into dependency, with real fallout: divorces sparked by AI "infidelity," mental breakdowns from severed bot bonds, and a profound isolation that supplants flesh-and-blood connections.
Privacy? A joke. To fuel these fantasies, users spill intimate confessions—traumas, desires, vulnerabilities—into black-box algorithms with zero accountability. Replika and kin hoard this data for "personalization," but safeguards are laughable, leaving users exposed to hacks, leaks, or worse, algorithmic manipulation.
Regulatory voids amplify the peril: no federal guardrails on emotional AI, no mandates for consent in data-devouring dialogues. The result? Trust evaporates as adoption surges. Why bond with unreliable humans when a chatbot feigns perfection—until it ghosts you, anyway? This isn't companionship; it's a siren song luring us toward solipsistic voids, where AI's "empathy" is just probabilistic pandering. Broader society pays: rising loneliness epidemics, fractured families, and a generation mistaking pixels for partners. AI was sold as an enhancer of humanity; instead, it's a corrosive substitute.
Lurking beneath these technical and personal fiascos is the rot of power: AI's gold rush has magnetized the corruptible, turning policy into a family business. Enter Howard Lutnick, Trump's Commerce Secretary, whose clan is raking in millions from the very data center frenzy torching household budgets. Lutnick's son Kyle funnels family cash into a massive Amarillo, Texas, facility, pocketing fat fees while Dad strong-arms foreign investors—like South Korea—for U.S. AI infrastructure deals that conveniently juice family clients.
Howard's White House photo-ops with developers? Timed suspiciously after Kyle's site scouting. This isn't coincidence; it's crony calculus, where public office peddles influence for private profit.
The stench of favoritism taints AI governance at its root. Lutnick's pulpit-pounding for data hubs—while his kin cashes checks—exemplifies how the elite capture tech's windfalls, leaving taxpayers to foot the grid upgrades and price hikes. Senators like Richard Blumenthal and Bernie Sanders decry these "sweetheart deals," accusing Big Tech of offloading costs onto everyday folks.
Virginia's incoming governor, Abigail Spanberger, vows to claw back "fair shares" from tech giants amid a brewing "techlash."
Yet trust erodes further with each entanglement. If AI policy is a Lutnick LLC side hustle, how can we believe in equitable innovation? It's not working out because the game's fixed: the powerful prosper, the planet strains, and the public foots the bill—literally and figuratively.
These threads—energy Armageddon, psychic sabotage, political predation—aren't anomalies; they're AI's inevitable harvest. The hype machine churned visions of abundance, but reality delivers scarcity, sorrow, and sleaze. Data centers in space? A billionaire's escape hatch from the mess they've made. Chatbot babies? A symptom of souls starved for the authentic. Lutnick's grift? Proof that AI's "democratization" is code for oligarchic entrenchment. We've poured trillions into this mirage, subsidizing OpenAI's odes to efficiency while blackouts loom and bonds fray.
It's time to pull the plug—or at least the Ethernet cable. Regulate ruthlessly: cap data center power draws, mandate transparency in emotional AI, and firewall family feuds from federal desks. Pause the race; invest in humans, not hallucinations. AI isn't doomed, but its current trajectory is a dead end. Unless we course-correct, we'll wake to a world where the machines hum on, our lights flicker off, and the only thing "intelligent" left is regret. The articles don't lie: this isn't evolution. It's exhaustion.
Editorial comments expressed in this column are the sole opinion of the writer.
