The AI boom: rational enthusiasm or the next dot-com bubble?

17 August 2026

By Malin Andersson, Johannes Breckenfelder, Stefano Corradin, Kalin Nikolov and Maria Antonietta Viola

Valuations on the US stock market, as measured by the CAPE ratio, are currently close to their historical peak.^[1] Euro area equity valuations have also risen, albeit to a lesser extent (Chart 1). Markets on both sides of the Atlantic reflect investors’ enthusiasm about artificial intelligence (AI) shaping the economy and driving profits. The extremely optimistic valuations raise questions: do today’s stock market prices reflect a rational bet on the transformative technology? Or are we seeing a remake of the dot-com bubble? We argue that economic research on past technological revolutions points to a worrisome conclusion: a correction of current stock market valuations is likely.

A sharp stock market correction would have severe consequences for the euro area, through two channels. One is euro area investors’ direct exposure to the Magnificent Seven stocks (hereafter Mag7) and the other is the degree of overexuberance in euro area stock markets themselves.^[2] This post explains why a correction should be expected even if current valuations are rational, why that matters not only to the shareholders who would take the direct hit, and what it implies for the euro area specifically.

Chart 1

US and euro area stock market valuations

a) S&P 500 Index (cyclically adjusted price-to-earnings ratio)	b) Euro area Market Equity Index (cyclically adjusted price-to-earnings ratio)
(Index)	(Index)

Why do technological revolutions so often end in a boom and bust?

The current excitement surrounding AI has many historical precedents. To name just a few: the railway boom of the 19th century, the expansion of electricity and radio in the 1920s and the surge of the internet, or the “dot-com era”, in the 1990s. In each case, a genuinely transformative technology attracted investment, and the stock market valuations of firms that adopted it rose strongly before falling sharply. Economic research offers two complementary explanations.

First, the rational view argues that high valuations can be justified by extreme uncertainty about a new technology’s productivity.^[3] Why has Nvidia’s share price risen 20-fold since 2022? Because investors rationally perceived that the company would become the next Google – with a highly uncertain and potentially large upside. In the worst case in such a scenario, investors lose their investment. But in the best case, the gains are large and genuinely hard to bound. This “option value” increases the stock valuations of early adopters, causing their price-to-earnings ratios to rise sharply.

Even if the technology succeeds, stock prices may eventually fall. Why? The nature of uncertainty shifts from a “single sector” to the “entire” economy. Initially, the new technology is like a small-scale experiment. If it fails, it’s unfortunate for that company, but the rest of the economy is unaffected. The risk can be diversified away. As adoption spreads, the same uncertainty becomes economy wide. If something then goes wrong with that technology, the whole economy suffers. This risk cannot be diversified, so investors demand a higher risk premium. However, this does not necessarily mean that profits will fall. Adoption itself is good news for cash flows, but the rising risk premium has the opposite effect and tends to prevail historically (Chart 2, blue line), unless profit growth is strong enough to compensate for that (Chart 2, yellow line). The exact timing is unknowable in advance. These boom-bust patterns are only identifiable with hindsight.

The second explanation for technology-driven boom and bust is the behavioural view. It holds that overconfident, overoptimistic investors bid up prices beyond fundamentals (see Chart 2, red line).^[4] When overconfidence fades, prices can fall even more sharply than in the rational scenario.

Both views imply a boom followed by a correction, or a pullback from wherever valuations have risen, at some point in the future. This does not mean that today’s prices represent a ceiling. If AI proves to be transformative enough, valuations could still be much higher in the future, even after a correction. As mentioned previously, it is impossible to know in advance where we stand on this path. Furthermore, the case for expecting a correction is not dependent on whether today’s prices are rational or irrational. We should be aware of that and prepare. The rest of this blog post considers two ways in which such a correction might affect the euro area – wealth effects for euro area investors in US stocks and contagion to euro area stock markets.

Chart 2

Stylised boom-bust scenarios

(Index)

How exposed is the euro area to a fall in Mag7 prices?

The dominance of the Mag7 on widely held global indices (e.g. MSCI World) carries significant risks for euro area investors. Panel a) of Chart 3 shows that most euro area exposures to the Mag7 are via investment funds – mutual funds and exchange traded funds (ETFs) – rather than direct holdings. Using data on the underlying mutual fund investors, we can identify which euro area investors carry the greatest exposure to US technology equities (Chart 3, panel b). Euro area households, which are increasingly channelling funds into low-cost ETFs, have around €440 billion of exposures to US technology equities without necessarily being aware of the associated concentration risk. Insurance companies and pension funds also hold significant exposures to the Mag7.

This fund-based structure is itself a transmission channel. A sharp correction can force funds to sell assets to meet redemptions – first, liquid holdings and then, if the correction persists, distressed assets – pushing valuations down further and triggering more redemptions. This is why a Mag7 correction is a question of financial stability for the euro area, rather than just a private one. The more severe scenario is not the equity correction on its own but a correction that coincides with broader market instability that policymakers cannot easily calm: unlike in the dot-com episode, today’s starting point leaves markedly less room to cut interest rates or use fiscal policy to cushion the fallout.

Chart 3

Euro area investors’ Mag7 exposures

a) Direct exposures	b) Indirect exposures
(EUR billions)	(EUR billions)

Is there overexuberance in euro area stock markets?

Beyond this exposure to the Mag7, is there also a risk of a home-grown correction? While euro area equity prices have risen substantially in recent years (Chart 1, panel b), valuation metrics such as price-to-earnings ratios remain considerably lower than in the United States (Chart 1, panel a). The euro area macroeconomic environment in the information and communication technology sector currently appears resilient compared with the times of the dot-com bubble. Productivity and markups in the sector are rising, the business climate in euro area digital services does not seem exuberant and euro area firms’ AI adoption is rising notably, only a few years after the launch of ChatGPT in 2022. The global AI boom has also sparked an ongoing and profound digital transformation across the euro area. The overall increase in digital investment in the euro area over the past decade was more than three times the cumulative growth in GDP over that period.

In sum, the euro area AI transformation is proceeding at a steady if unspectacular pace. On the one hand, euro area stock markets are dominated by “old economy” stocks, which show little of the AI excitement that we see in the US Mag7, reducing the risk of a future correction. On the other hand, euro area and US stock markets have historically been very highly correlated, meaning that a US correction will not leave the euro area unaffected.

Conclusion

Historical experience suggests that technological revolutions carry risks of a boom-bust cycle in asset prices, and this risk does not depend on today’s valuations being rational or irrational. The euro area’s smaller, less richly valued tech sector limits the risk of a home-grown crash. But this offers little reassurance: households, insurers and pension funds have significant exposures through global index trackers, and US equity stress has historically also had an impact on euro area stock markets. The effects of a US correction could extend beyond financial markets to euro area sentiment, financing conditions and hiring. A US AI fallout would not remain a US problem.

The views expressed in each blog entry are those of the author(s) and do not necessarily represent the views of the European Central Bank and the Eurosystem.

Check out The ECB Blog and subscribe for future posts.

For topics relating to banking supervision, why not have a look at The Supervision Blog?