Five researchers from the European Central Bank published an official blog post on August 17 titled "The AI Boom: Rational Exuberance or the Next Internet Bubble?" The authors, Malin Andersson, Johannes Breckenfelder, Stefano Corradin, Kalin Nikolov, and Maria Antonietta Viola, open by noting that artificial intelligence has driven a "rapid surge" in technology stocks, with cyclically adjusted price-to-earnings ratios (CAPE) in U.S. equities approaching historical peaks, while euro area valuations have also risen, albeit to a lesser extent.
Their research into past technological revolutions yields what they describe as a "troubling conclusion": a correction in current equity valuations is highly likely. The blog clarifies that the views expressed belong to the authors and do not necessarily represent the ECB or the Eurosystem. They also write that this risk should be acknowledged and prepared for.
Two complementary mechanisms point to a boom followed by a pullback, without relying on the assertion that "this is definitely a bubble." The first is a rational perspective: high valuations reflect the option value and uncertainty surrounding AI's productivity prospects. Once a technology shifts from a sector-specific issue to a macro-economic variable, investors demand a higher risk premium; historical experience shows that rising premiums often outweigh profit growth, causing prices to fall. The second is a behavioral perspective: overconfident investors push prices away from fundamentals, and when enthusiasm fades, the decline is steeper than the rational path would suggest. The authors conclude: "Both views imply a correction after the boom, or a pullback from already elevated valuations."
Railways, electricity, radio, and the early internet are all used as reference points. They also note that if the technology ultimately proves transformative, prices after a correction could still remain well above today's levels. Therefore, the article is not a trading recommendation on any specific index level, but rather a generalization about how asset prices behave during technological revolutions: even if the fundamental narrative holds, a repricing of risk premiums alone could be sufficient to generate a significant adjustment.
For the euro area, this is a financial stability issue, not merely a matter of shareholder gains and losses. There are two transmission channels. The first involves direct and indirect holdings of the U.S. "Magnificent Seven" — Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla. Euro area households hold approximately €440 billion in U.S. technology stocks through funds, with many unaware of the concentration; insurers and pension funds have comparable exposure. Holdings are primarily in mutual funds and exchange-traded funds, and redemptions or forced selling could amplify the decline. The second channel is the high correlation between euro area equities and U.S. markets: Europe's AI transformation is described as solid but unspectacular, with indices still skewed toward the "old economy," yet a sharp U.S. stock drop would spill over through sentiment, financing conditions, and hiring budgets. The authors write that the U.S. AI shock "will not remain an American problem."
The more severe scenario is not merely a price correction, but one overlapping with broader market turmoil, at a time when policy space is insufficient to easily smooth the impact. Compared with the period around 2000, the buffers of interest rate cuts and fiscal policy are "significantly smaller" today. This moves the discussion from valuation multiples to macro-stability: fiscal deficits and the neutral rate of interest have already risen, narrowing the room for central banks to offset asset price declines with large-scale easing.
The blog runs in parallel with the ECB President's remarks, without cancelling them out. The same week, President Christine Lagarde stated in Geneva that Europe cannot afford to miss the AI transformation, citing a survey indicating euro area firms plan to allocate around 9% of total investment to this field this year. The President emphasized real investment and competitiveness, while the blog focused on asset price trajectories and cross-border exposure. Both can hold simultaneously: Europe increasing capital expenditure does not eliminate the concentration risk of households holding U.S. equities through ETFs, and a higher probability of correction does not mean Europe should stop investing.
What must be distinguished is the policy level: the blog is not a Governing Council interest rate statement, nor is Lagarde's speech an official pricing of CAPE. For markets, the informational value of this blog lies in recasting "correction" from an emotional judgment into two standard macro mechanisms, while quantifying the euro area's fund exposure to the Magnificent Seven. It does not provide a timing or magnitude for the adjustment, nor does it advocate for immediately tightening financial conditions. Its policy implications point toward monitoring fund redemptions, the concentration in insurers and pension funds, and the synchronization between U.S. equity and euro area financing conditions.
Comments