A transformative technology triggers a capital-spending race. Leaders spend beyond their cash flow, financing moves off the balance sheet, and investors cheer until they start to punish the spending instead. Most of these booms bust within three years, and the infrastructure survives to be used cheaply by someone else.
Hundreds of railway companies were floated between 1844 and 1846, with investors paying in calls on shares to fund construction far beyond plausible traffic.
What happened next: Share prices collapsed in 1847 and many lines were never completed. The network that was built became the backbone of the Victorian economy, at a fraction of the capital's original value.bust; assets survived
Carriers and start-ups spent on the order of half a trillion dollars laying fibre and buying spectrum on the expectation of exponential internet traffic, much of it debt-financed.
What happened next: WorldCom, Global Crossing and dozens of others went bankrupt in 2001–02. Most of the fibre went dark for years and was later bought for cents on the dollar by the next generation of networks.bust; assets survived
Independent drillers ran negative free cash flow for years, funding drilling with high-yield debt on the strength of production growth rather than profits.
What happened next: The 2014–16 oil price collapse pushed more than a hundred producers into bankruptcy. Production kept rising under new owners with cheaper capital.bust; assets survived
Amazon, Microsoft and Google ramped data-centre spending sharply from 2015 while investors questioned the returns.
What happened next: Cloud revenue grew into the capital base. The spending was validated by demand and the companies' valuations rose with it.boom validated
Share of the 4 cases above that resolved this way. Small samples: treat as a prior, not a forecast.
Revenue rose 25% and Google Cloud doubled, but third-quarter capital spending of $51bn and a $9.07bn cash outflow told investors the AI build-out is now being paid for with the balance sheet.
