History does not say that AI is doomed to fail

The market has always lived with one habit it has never been able to get rid of — it loves the future more than the present. As soon as a technology appears that can change the rules of the game, capital begins to move faster than the technology itself can develop. This was the case with railroads, when investors priced in entire continents that had yet to be connected. This was the case with radio, which promised a new era of communication. This was the case with the internet, which in the late 1990s turned from an idea into mass hysteria. And, somewhat ironically, even simple things like ballpoint pens once triggered investment booms as if they were a breakthrough of the century.

Today, artificial intelligence has taken on this role, and the dynamics of the Nasdaq 100 index only reinforce the feeling of déjà vu. Over the past 10 years, its return has exceeded 640%, placing the current cycle among the strongest in history. It has surpassed eras such as the Japanese bubble of the 1980s, when the market believed in the endless growth of Japan’s economy, the “Roaring Twenties” in the U.S. with their faith in industrialization, as well as the post-war growth of the 1950s. Essentially, only the technological surge of the 1990s remains ahead — the one that ended with the painful collapse of the dot-com bubble.

But if you look deeper, it becomes clear that the issue is not a specific technology, but the very logic of innovation. Studies covering more than a century and a half of market development show a consistent pattern: in roughly three-quarters of cases, major inventions are accompanied by the formation of investment bubbles. This is not a system failure, but its natural state. When something truly new appears, the market does not know how to value it properly. It either underestimates the potential or, more often, overestimates the speed and scale of its realization.

It becomes especially complex when a technology does not just create a new product but triggers an entire chain of changes. In the case of artificial intelligence, this chain is multi-layered. First come chip manufacturers providing computational power. Then come models that use these resources. On top of the models, software is built. And that software begins to transform business processes, reduce costs, automate decisions, and ultimately reshape entire industries. The problem is that the market tries to price in this entire cascade at once, even though each stage develops at a different pace and with a different level of risk.

This is where the main trap for investors lies. During technological booms, there is an illusion that industry growth automatically means success for all its participants. But history suggests otherwise. During the dot-com era, investors were not wrong about the big picture — the internet did become the foundation of the modern economy. The mistake was in the details: they assumed that almost every company associated with the internet would secure a lasting place in the market. In reality, only a few survived, while most either disappeared or were absorbed by stronger players.

This lesson is especially relevant today. Artificial intelligence will most likely indeed become one of the key technologies of the 21st century. But that does not mean that every company using the word “AI” in its presentation will automatically become a long-term winner. On the contrary, it is precisely in such periods that the gap between real value and market valuation becomes the widest.

At the same time, it is important to understand that a bubble is not always a bad thing. In many cases, it accelerates technological development. Excess capital allows infrastructure to be built, research to be funded, and solutions to be created that, in a more “rational” environment, might develop much more slowly. Railroads, the internet, and mobile communications owe much of their scale to periods when money flowed faster than profits appeared.

In this sense, the current AI boom may play a similar role. Massive investments are already shaping the infrastructure of the future — from data centers to specialized chips and software platforms. The only question is who will ultimately turn this into a sustainable business and who will remain part of an “expensive experiment.”

For investors, the key skill in such periods is not to guess the peak or try to predict the moment of collapse, but to learn to distinguish levels of quality within a single trend. There are companies building the foundation of the new economy, and there are those simply trying to capitalize on the hype. On a chart, they may look the same, but in the long run, the difference becomes critical.

History does not say that the market will inevitably crash. It says that the market almost always first overestimates the speed of change and then underestimates its depth. And those who can live through both phases without excessive emotion usually end up winning.

In the end, the key question today is not whether there is a bubble. The question is whether you understand what exactly you are investing in: short-term hype or long-term economic transformation. Because the difference between the two usually becomes clear only in hindsight.