The Infrastructure Trap

There is a specific kind of ruin that only appears after a boom.

Steel rises. Roads are widened. Power lines are extended. Politicians cut ribbons. Investors applaud discipline and ambition. Analysts revise their forecasts upward. Local governments compete for the project. Contractors get paid. Engineers solve hard problems. Everyone points to the physical evidence and says: this is the future becoming real.

History is filled with monuments to yesterday's inevitability. Canal networks built on the assumption that water would remain the dominant artery of commerce. Telegraph lines stretched across continents just before newer communication systems changed the economics. Fiber networks laid during the dot-com frenzy faster than demand could absorb them. Printing plants expanded in the years before digital advertising gutted the newspaper business. Retail empires constructed around foot traffic that moved online. Office towers financed on the belief that white-collar work would always centralize in the same way.

“The AI boom has entered the stage where belief becomes concrete.”

None of those investments were irrational in the moment they were made.

The trap does not require stupidity. It requires extrapolation. It requires taking a real trend and building physical infrastructure as if the trend will continue in a straight line long enough for the assets to earn back their cost.

When the trend bends, the infrastructure does not bend with it.

A strategy document can be rewritten. A data center cannot be unpoured from the ground.