In 1840s Britain, investors believed railways would transform the economy. They were right. The trouble was that technological progress did not prevent financial speculation. The frenzy known as Railway Mania sent railway shares soaring before a collapse began in 1845. By 1847, Britain was engulfed in a broader financial crisis, compounded by agricultural shocks and tightening credit. The railways continued to expand, but many investors lost fortunes along the way.
Nearly two centuries later, Gabriel Galípolo, president of Brazil’s central bank, sees a historical parallel. Speaking at a Bradesco investment forum on Wednesday (November 11), he observed that the scale of spending on artificial intelligence infrastructure, measured as a share of GDP, may have few precedents outside the railway boom. The technological enthusiasm is accompanied by extraordinary demand for capital, even as questions remain about whether the investments will generate sufficient returns.
Galípolo’s warning goes beyond the possibility of a stock market bubble. Technology companies that previously financed expansion with their own resources are increasingly turning to debt. Some are moving from positive to negative cash flow while committing billions of dollars to infrastructure. The distinction matters: a company can persuade shareholders to finance losses with promises of future growth, but it cannot service debt on a compelling narrative alone.
There is also a broader macroeconomic problem. The United States, Germany and Japan, economies traditionally associated with supplying liquidity to global markets, are facing growing financing needs of their own. Governments are issuing more debt just as companies compete for the capital required to build the infrastructure of the AI economy. That competition could put upward pressure on interest rates and raise borrowing costs, particularly for emerging markets such as Brazil, where real interest rates are already exceptionally high.
The irony is that artificial intelligence may indeed transform global productivity while leaving substantial financial losses in its wake. Something similar happened to railway investors in the 19th century: the technology succeeded, but the capital behind it was not always rewarded. Galípolo is not predicting an AI crash. He is pointing to an old rule of capitalism: technological revolutions can change the course of history, but they do not suspend the obligation to repay debt.
Leave a Reply