10/07/2026
In 1999, all it took was adding “.com” to a company’s name to watch its share price take off. Companies with no revenue, sometimes with no finished product, raised tens of millions on the simple promise of the internet. Eighteen months later, the bubble burst: the Nasdaq index lost 78% of its value between 2000 and 2002. Yet, with hindsight, the internet did deliver on its promise. Amazon, which lost more than 90% of its value during the crash, is today one of the largest companies in the world.
This double lesson — that euphoria can be excessive, and that technology can nonetheless change the world — applies directly to what we are experiencing today with artificial intelligence.
Unlike some purely speculative market fads, generative AI already has concrete, measurable uses: automating administrative tasks, supporting medical decisions, optimising logistics, accelerating scientific research. This is not an abstract promise — it is a tool already deployed across millions of businesses. On this point, the parallel with the rise of the internet or of industrial electricity largely holds.
Some indicators nonetheless echo past excesses: technology company valuations that price in very ambitious growth scenarios, an unusual concentration of stock market performance in a handful of stocks, and a flow of capital toward anything carrying the “AI” label, sometimes with no direct link to a profitable business. These are exactly the mechanisms that, historically, have preceded periods of correction.
The history of great technological revolutions teaches one thing: the eventual winners are not always the companies that dominate the first wave of media attention. During the 19th-century railway boom, most railway companies went bankrupt — yet rail itself permanently transformed the economy. The same is likely true for AI: some of today’s most prominent players will not necessarily be the ones capturing the most value in ten years’ time.
Ignoring this revolution altogether would mean missing out on a genuine growth theme. But putting an excessive share of one’s wealth into it would mean betting on the outcome of a game whose winners nobody yet knows.
The right approach remains the same as for any promising theme: measured exposure, integrated into a diversified allocation across sectors, geographic regions and asset classes — rather than concentration on a handful of fashionable names.
When it comes to wealth management, it is not the most spectacular bets that build lasting performance, but discipline and diversification, maintained over time.
