Wires to Nowhere

Illuminating History's Most Obscure Corners | Issue #27 | March 2026

America has a habit. Every few decades, it bets the house on a single transformative technology. Builds like there's no tomorrow. Then tomorrow arrives. This is the story of what happens next — and why you're already paying for it.

Check your electricity bill.

If you live in the United States, it's higher than it was five years ago. Probably by a lot.

The average American household paid around 13 cents per kilowatt-hour in 2019. By late 2025, that number had crept to 19 cents. A 46% jump. In the USA. Quietly. Without a big news story. Without a congressional hearing on your behalf.

You probably blamed inflation. Maybe fossil fuels. Maybe the grid aging out.

All partially true. But there's another reason — one that almost nobody is talking about in plain language.

The tech industry is building a new kind of railroad.

And you, without knowing it, are helping to fund it.

In 2025, data center operators submitted requests to connect more than 700 gigawatts of new capacity to the American electricity grid.

The entire United States consumed 477 gigawatts in all of 2023.

Let that land.

The tech industry asked for more new grid connections in a single year than the whole country currently uses.

Most of those projects will never be built.

But here's the trap: the mere act of requesting that power triggers real spending. Grid operators have to model for it. Utilities have to staff for it. Infrastructure gets planned and sometimes built speculatively.

Those costs go somewhere.

They go to you.

Utilities requested more than $29 billion in rate increases in the first half of 2025 alone — double the same period in 2024. And the distribution of that pain is grotesque in its lopsidedness.

Residential electricity prices rose 25% between 2020 and 2024.

Industrial electricity prices — the rates that data centers pay — actually fell over the same period.

The companies driving the surge in demand pay less per kilowatt-hour than you do.

In Georgia, voters figured this out before most pundits did. Residential electricity prices had climbed 41% in four years. In November 2025, they voted out two incumbent utility commissioners in a landslide. No policy paper required. Just a bill on the counter every month.

This is the hidden tax of the AI boom. And it has happened before.

To understand where we are, you need to go back to 1873. Further, actually.

Go back to the 1840s and 1850s, when the railroad was the internet of its age — the technology that everyone could see would reshape everything.

The great trunk lines connecting the East Coast were genuine marvels. Freight that once took weeks moved in days. Whole cities appeared where rail lines crossed.

Capital flooded in. Government land grants flowed to railroad companies like water. Investors poured in from Europe. Companies issued bonds on top of bonds. A bond, unlike a stock, is a loan — you hand a company your money and they promise to pay you back at a fixed interest rate over a fixed period.

In the railroad era they were issued by the thousands, many by companies with no realistic prospect of ever making good on them.

Between 1868 and 1873 alone, 33,000 miles of new track were laid across the United States. To put that in perspective: the distance from New York to Los Angeles is about 2,800 miles. America built the equivalent of eleven transcontinental railroads in five years.

It was also wildly irrational.

Many of the lines being built had no plausible customer base for decades. They were financed on the assumption of future demand that existed mostly in the imaginations of their promoters — who, in many cases, also knew exactly what they were doing.

The most spectacular scandal, the Crédit Mobilier affair, revealed that the men building the transcontinental railroad were essentially paying themselves out of government subsidies — invoicing the project at double cost, pocketing the difference, and distributing shares to congressmen to keep questions quiet.

The Panic of 1873 ended the party. Banks collapsed. Average wages fell by nearly a quarter. The country entered double-digit unemployment that lasted over a decade.

The era came to be known as the Great Depression — until 1929 reset everyone's sense of scale.

Then, because the lessons weren't learned, it happened again.

The Panic of 1893 sent one quarter of all U.S. railroad mileage into receivership — meaning the companies could no longer pay their debts, and a court-appointed administrator stepped in to manage their assets on behalf of creditors, while original investors lost everything.

Over 500 banks failed alongside the railroads. Unemployment in Michigan hit 43%.

The railroads themselves survived.

The tracks were still there. The locomotives still ran. The infrastructure, built at enormous and often fraudulent cost, became the circulatory system of American commerce for the next 70 years. The investors were wiped out. The nation got a railroad network for the price of their losses.

That's the deal America keeps making with itself.

The question worth asking — and the one the triumphalist version of this story always skips — is whether it is actually a good deal.

It wasn't. Not entirely, anyway.

The network was overbuilt not just financially but physically. Too many competing lines serving the same routes. Too many branch lines to nowhere. Too much duplicated infrastructure that would never generate the traffic to justify its existence. The financial collapse didn't cure that. It just transferred ownership.

And then a new technology arrived and made much of it irrelevant anyway.

As late as 1946 — more than 30 years after Henry Ford began mass-producing the Model T — trains still handled two-thirds of America's commercial passenger traffic. But the car was already eating the railroad's lunch. And then the government handed the automobile a gift the railroads couldn't match.

The Interstate Highway Act of 1956 created 41,000 miles of federally funded roads. Trucks could haul freight on highways the government built and maintained. Railroads paid taxes on every mile of track they owned.

The competition was structurally unfair from the start, and the railroads — already weakened by decades of overbuilding and collapse — couldn't survive it.

By 1966, less than 2% of all intercity passengers in America were traveling by rail.

The great passenger railroads died one by one.

In 1971, the federal government created Amtrak to run what was left — a government-supported service operating, in most of the country, on infrastructure that would embarrass a mid-sized European nation. The Northeast Corridor still relies on bridges and tunnels built in the 1930s and signal systems of the same vintage.

So was it a good bet?

Freight rail matters. The network is genuinely important to the American economy today.

But the passenger network — the thing that was supposed to transform how a continent moved — was outcompeted, starved of investment, and eventually handed to the government as a ward of the state.

The bet paid off partially, unevenly, and only after two depressions, millions of destroyed investors, and a century of stagnation.

Not a ringing endorsement.

We've been talking about massive, society-reshaping bets on infrastructure.

The railroad was one. Here is another — less visible in the history books, far more instructive, and centered on one of the most extraordinary characters American business has ever produced.

Most people have never heard of Samuel Insull.

This is genuinely strange, because for most of the 1920s he was one of the most famous men in America — and what happened to him reshaped the country's financial architecture forever.

Insull was Thomas Edison's personal secretary. Not the inventor. The man who actually ran operations while the inventor invented.

Methodical, driven, possessed of a specific kind of ambition that only finds peace in ever-larger projects.

Before we get to what he built, it's worth pausing on what electricity looked like in the 1880s — because it was chaos.

When electricity first arrived as a commercial product, it was delivered by competing private companies running incompatible systems. Different voltages, different plug types, different wires running down the same streets. You couldn't take an appliance from one neighborhood and plug it in at another. A voltage is simply the measure of electrical pressure — the force pushing electricity through a wire, analogous to water pressure in a pipe.

Standardizing voltages meant agreeing that electricity would arrive everywhere at the same pressure, in the same form, so that any lightbulb or motor built for one customer would work in any other home.

Obvious in retrospect.

At the time it required decades of industry warfare and the gradual, forced consolidation of hundreds of incompatible local systems.

Insull moved to Chicago in 1892, took over a small electricity company, and immediately grasped something his competitors hadn't.

Drive down the cost through scale. Connect more customers. Build bigger generating plants. Lower the price until everyone could afford it.

Treat electricity not as a luxury but as a commodity.

In 1902, the United States used 6 billion kilowatt-hours of electricity per year. By 1929, that number was 118 billion.

Insull built a significant fraction of the infrastructure behind that transformation. His name was on the power stations. His grid lit up Chicago, then the Midwest, then 32 states. He was celebrated, decorated, treated as a visionary.

He was also, quietly, building something that would destroy him.

Fearing that outside investors might accumulate enough shares to push him out of his own companies, Insull turned to a device called the holding company as protection.

A holding company is a corporation that exists solely to own shares in other corporations. It produces nothing — no electricity, no goods, no services. It sits on top of operating companies, collecting their dividends and controlling their votes. Think of it as a shell corporation that owns other corporations.

Insull didn't build one. He built a pyramid of them.

Here is how it worked, in plain terms. Imagine you start a real electricity company — call it Company A — that generates real power for real customers. You own 51% of its shares, so you control it. Now you create Company B, a holding company, whose only asset is shares in Company A. You sell shares in Company B to the public, keeping 51% for yourself. The public's money flows in. Now you create Company C, which holds shares in Company B. And so on up the chain.

At each level, you raise money from the public by selling shares. At each level, you retain control with just 51%. By the time you have four or five levels, a very small personal investment controls an enormous amount of other people's money — because control at the top means control of everything below it.

By the time Insull's structure was complete, he had formed more than 95 holding companies and 255 operating companies. For every dollar he had originally put in, he now controlled $2,500 of public money.

Not his money. Money that ordinary people — teachers, shopkeepers, retirees across the Midwest — had invested because the electric company seemed like the safest thing in the world.

Middle West Utilities was the cornerstone of the pyramid — the main vehicle through which Insull controlled his sprawling empire of regional electricity and transit companies across 32 states. On paper, one of the most valuable corporations in America.

In the summer of 1929, Insull's securities were appreciating at $7,000 a minute.

Then October came.

The Depression hit Insull from both directions simultaneously. Industrial electricity use collapsed as factories shuttered. Households cut their usage to survive. Operating cash flows shrank across the pyramid.

But here was the structural trap the holding company model had created: each layer owed dividends upward to the layer above, all the way to Insull at the top. The whole structure depended on a continuous flow of cash rising from the operating companies at the base. When that flow slowed, the pyramid didn't adapt. It starved. And because there was debt at every layer — borrowed against expected future earnings — the interest payments didn't stop just because the revenue did.

Could he have saved it? Almost certainly not, once the Depression hit. The structure was built for growth and couldn't run in reverse. A simpler company — one real utility, one balance sheet — might have weathered the storm as many utilities did.

Insull's pyramid was too tall, too leveraged, too dependent on conditions that no longer existed.

He tried everything anyway. He sold his personal estate. He cashed in his life insurance policy. He put up his own shares as collateral for new loans — pledging his control of the empire to borrow money to keep it alive.

Middle West Utilities stock, once trading at $570 a share, fell to $1.25.

The final blow was almost absurdly small in scale. A Chicago printing firm to which Middle West Utilities owed a bill of $3,000 — against a $3 billion empire — filed to put the company into receivership.

That single unpaid invoice became the pin that popped the pyramid. Cross-default clauses — contractual tripwires specifying that one company's failure automatically triggers failure in related companies — began firing.

The bankruptcy of the Chicago Rapid Transit Company dragged 65 affiliated companies down within months.

Insull fled to France, then Greece, which had no extradition treaty with the United States. He was eventually arrested in Turkey, extradited to Chicago, tried, and acquitted.

He died in 1938, collapsing alone on the platform of a Paris Métro station. He had once been worth $150 million. His pockets contained the equivalent of about $20.

Over a million ordinary investors lost their savings. And the lights stayed on.

This is the part that tends to get left out of the story — and it matters for understanding the full bet.

The grid Insull built was real.

His operating companies — ComEd in Chicago, the regional utilities across the Midwest — survived his collapse, were reorganized under new management, and continued expanding the American electrical infrastructure for decades.

The electrification of rural America, the standardized national grid, the cheap reliable power that made American manufacturing dominant for most of the 20th century — much of that foundation was laid during the boom years Insull personified.

The collapse also, involuntarily, built the rules that protected everyone afterward.

Three laws emerged directly from the wreckage. The first forced companies to tell investors the truth about what they were actually selling. The second banned the kind of multi-layered holding company pyramid Insull had built, requiring electricity companies to actually be in the business of selling electricity — not financial engineering. The third brought power to rural areas that private capital had never bothered to reach.

Amazon, Microsoft, Google, and Meta collectively spent over $200 billion on capital expenditures in 2024. A 62% year-on-year increase. Almost all of it: data centers for AI.

In the first nine months of 2025, AI infrastructure investment accounted for more than a third of all U.S. GDP growth.

One third.

The historical echoes are not subtle. In July 2024, a voltage fluctuation in northern Virginia — the data center capital of the world — caused 60 data centers to simultaneously disconnect from the grid. The resulting 1,500-megawatt surge nearly cascaded into something much worse. The system held. Just.

Here is what's different this time. The railroad bet was a bet on permanent physical assets. Steel track laid in 1870 was still carrying freight in 1950. The investors were destroyed, but the infrastructure endured.

Data centers don't work that way. The servers inside have a useful life of three to five years. Today's leading-edge AI chips will be obsolete before a typical infrastructure bond matures.

But serving data centers requires permanent grid infrastructure — transmission lines, substations, transformers — built right now on the assumption that the data centers will keep drawing the power they've promised. If the boom consolidates faster than expected, the servers become scrap. The grid upgrades remain. And ordinary people keep paying for them on their electricity bills for decades.

America's deal has always been the same: private capital goes berserk, builds too much too fast, collapses, socializes the losses, and wakes up a decade later with infrastructure it couldn't have built any other way. The railroads left behind steel that lasted a century. Insull left behind a grid that still powers America.

What will data centers leave behind when the servers are five years old and the boom has moved on?

Right now, the answer looks like higher electricity bills and warehouses full of concrete, cooling fluid, and obsolete chips.

You've already opened the first invoice.

______________

Sources: U.S. Energy Information Administration; Novel Investor; American Heritage Magazine; Chicago History Museum; EBSCO Research Starters; Living New Deal; Encyclopedia.com; American Association of Railroads; utility industry reporting, 2025.

What bizarre historical phenomenon should we investigate next? Drop us a line at [email protected].

Unsubscribe | View in browser | Past issues