A very good read and true: "Technology changes. Human behaviour doesn't"
The Railway Bubble Explains AI Better Than Most People Think
The railway bubble is one of the best ways to understand whatâs happening with AI today.
The railway boom began with a real breakthrough.
Before railways, moving goods and people was slow, expensive, and limited. Goods travelled by horse carts, canals, and ships. People moved at a much slower pace. Markets were mostly local. A regionâs economic value depended heavily on how easily raw materials, workers, and finished goods reached other places.
Then railways arrived.
Distance began to shrink.
Trips once measured in days took hours. Coal, steel, food, and manufactured goods moved faster and cheaper. Cities expanded. Workers reached new places. Trade routes shifted. Entire regions became more valuable because they were connected to railway networks.
So the early excitement made sense.
Railways changed the world.
The first railway companies showed investors what was possible. Investors saw speed, scale, and profit. Governments saw national development. Promoters saw a chance to raise money. The public saw a future unlike the past.
Thatâs how the bubble began.
First came real innovation.
Then came the early winners.
Then came the story everyone wanted to believe.
âRailways are the future.â
And they were.
But the market slowly turned a true idea into a dangerous assumption.
âRailways are the future, so every railway company must be valuable.â
Thatâs where things went wrong.
Money rushed into almost every railway project. New companies appeared quickly. Promoters proposed routes everywhere. Investors bought railway stocks less because they understood each business, and more because railway stocks were rising.
The quality of the project mattered less.
The price mattered less.
The expected return mattered less.
Exposure to the future became the only thing people cared about.
That is classic bubble psychology.
A real technology creates real winners. Investors then stretch those early wins across the whole sector. Capital gets careless. Valuations break away from reality.
As railway mania grew, more routes were approved. Some made sense. They connected major cities, industrial hubs, ports, and high-demand areas.
Many others didnât.
Some lines copied routes already in place. Some connected weak markets. Some were built mainly because money was available. Some existed more for speculation than long-term profits.
The world needed railways.
It didnât need every railway, in every location, at every price.
That was the mistake.
The infrastructure had value. The capital allocation didnât.
Then reality arrived.
Railways were expensive to build. Land had to be bought. Tracks had to be laid. Bridges, tunnels, stations, and engines required huge sums of money. Costs rose. Many companies needed more funding than expected. Profits took longer to arrive. Passenger and freight demand wasnât strong enough to support every route.
Investors eventually understood the future was real, but the returns were not spread evenly.
The best railway lines survived.
The strongest operators became valuable.
Weak projects collapsed.
Share prices crashed.
Speculators were wiped out.
Money raised in a wave of excitement turned into losses.
Thatâs the main lesson.
Railways didnât fail.
The railway bubble failed.
The technology survived. The infrastructure stayed useful. Railways kept transforming economies for decades and became one of the foundations of industrial growth.
But investors who bought the wrong railway stocks at the wrong prices still lost money.
Thatâs how bubbles work.
They are not always built on fake ideas.
Many of the biggest bubbles begin with real technology, real innovation, and real productivity gains.
The bubble forms when investors confuse the future of the technology with the future returns of the stocks.
Railways changed the world.
Railway investors still lost money.
The internet changed the world.
Dot-com investors still lost money.
Now AI has taken the place of railways.
AI is real. It will change productivity. It will affect software, research, coding, automation, defence, healthcare, finance, content, and robotics.
But that doesnât mean every AI company deserves any valuation.
The tracks and stations of the railway era have been replaced by GPUs, data centers, power contracts, cloud spending, and private AI valuations.
The story is once again simple.
âAI is the future.â
And that statement is likely true.
But the dangerous assumption is back too.
âAI is the future, so every AI asset must be valuable.â
That is where investors need discipline.
The question is not whether AI is real.
The question is whether future AI revenue will justify the amount of money being poured into chips, data centers, energy infrastructure, cloud contracts, and private valuations today.
History doesnât repeat perfectly.
Human behaviour does.
Every generation believes its bubble is different because the technology is new.
The pattern usually stays the same.
Real innovation.
Early winners.
A story everyone repeats.
Capital rushing in.
Overbuilding.
Valuations getting stretched.
Reality returning.
A crash.
Survivors become infrastructure.
Speculators get wiped out.
Railways were real.
AI is real too.
But a real technology still turns into a financial bubble when capital loses discipline.
That is the lesson from railway mania.
The future might be right.
The price might still be wrong.