Earlier this week, the number one most-read story in the Financial Times was headlined, “Nvidia shares fall on signs Google gaining upper hand in AI.” It turns out that Google’s AI software/product is doing about as well as its competitors but isn’t using Nvidia’s “must have” AI super-chips, shocking the market.
But even with that, we ain’t seen nothing yet: get ready for the Moores’ Law shock and a few others that seem increasingly inevitable.
For months now, the tech world has been drunk on the language of inevitability. AI will transform everything. AI demand for electricity will double America’s power consumption. AI data centers will be the new steel mills, the new auto plants, the new engines of prosperity.
Wall Street has inflated this into a bubble so big that it’s hard to see where the hype ends and the real economy begins.
But if history teaches us anything, it’s that bubbles don’t pop harmlessly; they burst outward. And when they do, the people who had nothing to do with inflating them are usually the ones who end up paying the biggest part of the price.
We’ve seen this movie before. When the dot-com bubble collapsed in 2000, it wasn’t just stock traders who felt the pain. Entire cities that had boomed on tech spending suddenly cratered. Construction, retail, restaurants, and transit systems all took the hit.
It wiped out retirement savings, yes, but it also wiped out jobs for working people who never owned a share of Pets.com. Economists later found that the recession following the dot-com crash fell hardest on lower-income workers who had been pulled into booming metro economies that vanished overnight.
Then came the housing and derivatives bubble of 2008. Once again, the story was sold as something contained within the financial sector. Exotic mortgage-backed securities. Credit default swaps. CDOs squared.
Most Americans had no idea what any of that meant, but when that bubble burst, they didn’t need a glossary. They felt it in lost jobs, lost homes, gutted neighborhoods, and a social safety net that was suddenly overwhelmed.
George W. Bush made sure that his Wall Street donor executives walked away with billion-dollar golden parachutes instead of going to jail, while average Americans endured a decade of lower wages, higher rents, and shredded public services.
When banksters crashed the S&L banking system during Ronald Reagan’s presidency in the 1980s, he prosecuted more than 3,000 of them and sent more than 1,000 to prison. But Republicans stopped respecting the rule of law in a big way around the time five corrupt Republicans on the Supreme Court handed the White House to Bush in 2000.
And then, of course, five corrupt Republicans on the Court legalized political bribery with Citizens United in 2010, so now the banksters own DC and have bought a lot of deregulation recently. As a result, the risks now are even greater than they were in 1999 or 2008.
Bubbles don’t stay on Wall Street: they metastasize through the entire economy. And the AI bubble today is no different; in some ways, it’s even more deeply wired into the lives of people who may never touch ChatGPT, Midjourney, or a Nvidia chip.
Right now, electric utilities from Arizona to Georgia are spending billions building power plants and upgrading transmission lines because they’ve been told AI demand will explode for decades.
Utilities aren’t like normal businesses; when they make a bad bet, they don’t absorb the loss. Instead, they pass it on to you. And that means that if the AI bubble bursts — if the rosy forecasts don’t materialize — average Americans could end up paying higher electric bills for a generation to cover infrastructure built for a demand that never arrived.
This has happened before. After the 1990s gas-capacity boom fizzled, ratepayers in multiple states spent decades covering the cost of underused plants. When nuclear projects went over budget or were abandoned altogether, like the VC Summer project in South Carolina, consumers were forced to cough up billions to foot the bill while executives walked away with multi-million-dollar bonuses.
The same fate is now looming over families already struggling with high utility bills. AI operators may walk away from canceled projects, but working class people can’t walk away from the debt that built the substations, transformers, cooling systems, and pipelines intended to serve the data centers. Giant utilities and their morbidly rich executives will squeeze it out of you, me, grandma, and the kids.
Low-income households will feel this first. When utility rates climb, affluent families may grumble, but low-income families get their power shut off. Thousands of Americans die every year in heat waves because they can’t afford air conditioning, and climate change is making the situation worse. Piling the cost of a Wall Street-driven bubble onto electric bills isn’t just unfair: it’s dangerous.
States are also handing out massive tax incentives and subsidies to lure AI data centers. These deals often rely on long-term economic projections that look a lot like the rosy promises made in the lead-up to every other bubble in the last century.
If AI expansion stalls, those states are left with reduced tax revenue, higher infrastructure costs, and no way to fill the gap except by cutting public services or raising taxes on people who can least afford it.
This is exactly what happened when manufacturing plants promised by past booms never materialized or closed early: schools went underfunded, transit systems decayed, and local governments fell deeper into debt. I grew up in Michigan and saw this first-hand as the auto boom collapsed under the weight of Reagan’s free-trade neoliberalism.
And then there’s the jobs picture. The AI boom has unleashed a construction frenzy of data centers, substations, power plants, cooling towers, and fiber lines. These are good jobs for electricians, pipefitters, carpenters, welders, and truck drivers. If the bubble bursts, however, those jobs will vanish overnight.
The layoffs won’t hit coders at Google; they’ll hit working people who relied on the stability of a years-long construction pipeline. When jobs like that collapse, they drag down entire communities: restaurants, small shops, repair businesses, daycares, and clinics. It’s the same domino effect we saw after both the dot-com collapse and the 2008 financial crash.
Meanwhile, pension funds — especially public pensions — are heavily invested in tech stocks and the infrastructure financing now increasingly tied to the AI boom. Teachers, firefighters, public employees, and retirees who depend on those pensions could watch helplessly as their future security evaporates through no fault of their own.
After Bush’s 2008 housing crash, public pension systems all over the country were left with massive unfunded liabilities that led to service cuts, higher contribution rates, and reduced benefits. The same pattern is already being written into the AI bubble, brick by brick, stock by stock.
Worst of all, an AI crash could hit at a moment when tens of millions of Americans have no margin left. Inflation has been punishing, and Trump’s incoherent tariff policies have made him and his kids rich (as they use tariffs to extort foreign governments to give them billions in cash, build Trump resorts, and even gift Trump a jet plane), but they’re relentlessly jacking inflation on the rest of us.
- Housing costs have become predatory as Republican-aligned Wall Street vultures swoop in and buy up entire city blocks of single-family homes to convert into rentals.
- Medical debt from for-profit hospitals and insurance companies is pushing families into bankruptcy because Republicans refuse to even allow a discussion of single-payer healthcare like the rest of the developed world has.
- Student loans, which Republicans sued to prevent Biden from forgiving, are again grinding down young workers.
- Meanwhile, billionaires are gambling with the $4 trillion tax cut Donald Trump gave them, and that loose money is jacking the stock market like in 1929 after the Republican Harding/Coolidge/Hoover tax cuts (from 91 percent on the morbidly rich down to 25 percent).
In an economy already stretched to the breaking point, the shockwaves from a tech market collapse could intensify already obscene levels of inequality in ways we haven’t seen since the Republican Great Depression.
The wealthy will weather it. They always do. They’ll diversify, hedge, shift assets, pick up distressed real estate at a discount, and wait for the next upswing. Most will even profit from it, buying up homes, businesses, stocks, and other assets for pennies on the dollar.
America’s billionaires saw their greatest gains during the dot-com bust and the housing crash. “Cash is king” was the saying in the 1930s, as well as after the dot-com and housing crashes. And they’re muttering the same today with breathless anticipation.
But low-income and working-class Americans — the people least responsible for the bubble — will face higher electric bills, job losses, crumbling schools, gutted pensions, and reduced public services. They’ll pay for the gambles made by the same financiers and speculators who made out like bandits in 2000 and 2008.
In 1965, Intel co-founder Gordon Moore postulated that every two years the number of transistors in an integrated circuit would double. Not only was his “Moore’s Law” right, but as the power of digital hardware increases, costs also reliably drop by a similar factor.
Most of today’s frenzy of data center and power plant construction is based on the current state of the AI chip-and-software art, but Moore’s Law dictates that over time — and not a lot of time, probably just a matter of months or years at the most — the size and power needs of these AI data centers will decrease exponentially.
Already, entire buildings filled with computers are on the verge of being replaced by a single “wafer” disc that will carry the computing power of billions of today’s red-hot chips. The Wall Street Journal broke the news just three weeks ago with an article titled and subtitled:
“The Microchip Era Is About to End. The future is in wafers. Data centers will be the size of a box, not vast energy-hogging structures.”
Author George Gilder noted we’re not just on the verge of the breakthrough; it’s nearly in production:
“Cerebras of Palo Alto, Calif., used the concept in its WSE-3 wafer-scale engine. The WSE-3 boasts some four trillion transistors — 14 times as many as Nvidia’s Blackwell chip — with 7,000 times the memory bandwidth. Cerebras inscribed the memory directly on to the wafer rather than relegating it to distant chips and chiplets in high-bandwidth memory mazes. The company stacked up its wafer-scale engines 16-fold, thereby reducing a data center to a small box with 64 trillion transistors.”
A data center — a 700,000 square-foot building drawing gigiwatts of electricity and chugging millions of gallons of water for cooling — replaced by “a small box” that could be powered by rooftop solar and a good battery bank.
Ya think that may have an impact on our economy?
Places booming today as AI and other data facilities and the power plants and transmission lines to feed them are being built — if Moore’s Law applies to data centers as it has proven to apply to all things digital from computers to cell phones to TVs and satellites — will soon look like Flint, Michigan when Reagan’s free trade policies began to seriously bite in the 1990s.
If there’s a lesson from history, it’s that bubbles only appear harmless until they burst. And the bursting always lands hardest on those who never benefited from the boom in the first place.
In a future article, I’ll examine the political consequences of this possibility (hint: read Andrew Ross Sorkin’s new book about the 1929 crash and its impact on America, the world, and how it realigned our nation’s politics) but for now, get ready. This could get real ugly, real fast.
