The technology lasts. The investors often don’t.

Railways, fibre optics and now data centres: every great infrastructure boom has changed the economy. Far fewer have rewarded those who financed them.

Railways

Before the first transcontinental line was completed in 1869, crossing the United States took months. Afterwards it took about a week. Goods, grain and coal moved at a speed that reshaped the economy, and cities grew around the tracks.

Investors understood the opportunity, and capital poured in. By the late nineteenth century railway securities made up the largest part of the American stock market. Competing lines were built between the same cities, financed largely with debt. The Panic of 1893 followed: more than 15,000 businesses and around 500 banks failed, and many of the largest railway companies went into receivership.

The tracks did not disappear. They were reorganised, consolidated and run by new owners who had acquired them at a fraction of what they cost to build. The country kept the benefit. Many of the original investors did not.

Fibre optics

A century later the pattern repeated. In the late 1990s, telecommunications companies laid fibre across continents and oceans in expectation of explosive demand for internet traffic. Demand did arrive, but more slowly than the financing assumed. By the early 2000s only a small fraction of the fibre in the ground was in use, and some of the most prominent carriers had filed for bankruptcy. The networks were bought cheaply and became the backbone of the internet that followed.

The pattern

Three things recur. The technology is real and transformative. Capital arrives faster than revenue, often financed with debt. And the returns accrue mainly to the wider economy and to those who buy the assets after the first owners have failed. The question for an investor is rarely whether the technology will matter. It is what price is being paid, how the investment is financed, and who bears the loss if revenues arrive late.

Data centres

The current build-out of computing capacity for artificial intelligence is among the largest private investment programmes in history. Some features echo the past: very large capital commitments ahead of proven profits, and a growing share financed through long-term credit. One feature is new. Much of the equipment, notably processors, has a useful life of only a few years, while some of the financing runs for decades.

None of this means the technology will disappoint. It means that the investment case and the technology case are different questions, and that history argues for keeping them apart.

How we think about it

We look at what a family owns rather than what it believes in: the price paid relative to cash flows, the strength of the balance sheet, and the concentration of exposure to a single theme. In recent years broad equity indices have become more concentrated in a small number of large technology companies than at any time in decades, so even a diversified portfolio may carry more of this theme than it appears.

Sources: historical accounts of the Panic of 1893 and of the 2000 to 2002 telecommunications downturn. This article is provided for information purposes only. It does not constitute investment advice, an offer, or a recommendation to buy or sell any financial instrument or to adopt any strategy. The views expressed are those of the author and of GYRA SA at the date of publication and may change without notice. Past performance is not indicative of future results. Data sourced from third parties is believed to be reliable but has not been independently verified by GYRA SA.