In recent years, Artificial Intelligence has become far more than an emerging technology. It has turned into an all‑encompassing promise: of limitless growth, unprecedented productivity, and new economic models capable of reshaping entire industries. And it is precisely when a promise becomes totalizing that markets begin to lose clarity.

More and more analysts, institutional investors, and financial authorities are now asking the same question—this time with less enthusiasm than before: is AI truly priced for what it is today, or for what we hope it will become tomorrow?

The distinction is not theoretical. It is the thin line between innovation and financial instability.

According to several analyses from the Financial Times, a significant share of today’s valuations for AI‑related companies already embeds extremely optimistic growth assumptions—often unsupported by proportional cash flows.

This does not mean AI lacks value. It means the market is paying today for earnings that may materialize many years from now—or not at all. It’s a familiar mechanism in financial history: when an asset’s price reflects perfect expectations, even a small deviation from the narrative can trigger a sharp correction.

Unlike previous technological waves, Artificial Intelligence carries a financially dangerous feature: it is extraordinarily expensive to sustain. As reported by Reuters, the AI race has pushed major tech giants into massive investments in data centers, specialized chips, and energy infrastructure—costs running into tens or even hundreds of billions of dollars.

The critical point is that these costs are immediate, while returns are delayed and uncertain. If AI adoption slows, or if business models fail to scale as expected, many companies could end up with cost structures far heavier than their real revenues. In such cases, markets don’t punish gradually. They cut.

A particularly telling warning comes from those building AI itself. Demis Hassabis, CEO of DeepMind, recently described the current investment climate as “bubble‑like,” noting that many startups are raising enormous sums without mature products or viable business models. Bill Gates has also urged caution, warning that AI hype risks destroying value for firms unable to turn innovation into sustainable profits.

When even the pioneers start hitting the brakes, the message is clear: the risk is no longer theoretical.

Another factor heightening the fragility of the situation is the concentration of value. A large portion of global stock‑market gains is now tied to a small cluster of companies heavily exposed to AI. As Business Insider has reported, a significant correction in AI stocks could cascade through the entire financial system, amplifying volatility and striking even investors with no direct exposure to the sector.

This is not an “isolated” bubble. It is a potential driver of systemic instability.

The comparison with the dot‑com bubble is unavoidable. Between the late 1990s and 2000, euphoria over the emerging digital world turned Wall Street into a virtual gold rush. Adding “.com” to a company name was enough to send its shares soaring, often without revenues or functioning business models. In March 2000 the spell broke: the NASDAQ collapsed by 80% over two years, taking hundreds of startups down with it.

The dot‑com crash remains one of the strongest warnings in modern finance: even real revolutions—like the internet—do not justify valuations detached from economic reality.

The International Monetary Fund has recently noted that today’s AI investment surge exhibits similar dynamics: early exuberance, rapid valuation growth, and an inevitable phase of selection and consolidation. The concern is not a total collapse, but a sudden, disorderly correction capable of eroding trust and capital at high speed.

That is the central paradox of Artificial Intelligence: the technology is real, but the pricing often isn’t. Part of the market treats AI as if it were immune to the basic rules of economics—costs, margins, demand, sustainability.

History teaches that when a technology is seen as “inevitable” from a financial point of view, the risk of excess becomes maximal.

AI is not going away. But the way it is currently funded, narrated, and priced is sending signals markets can’t afford to ignore.

The real danger is not that AI will fail.

The real danger is that expectations will fail first—triggering a correction that could leave the market more unstable, more nervous, and less rational.

As with all great speculative waves, in the end one distinction will remain:
those who create real value—and those who survive only on promises.