Market history has a specific failure mode for technology cycles: equity prices concentrate early on a sector's expected long-run gains, overshoot the timeline that actually materializes, and then correct. The mechanism turns on timing, not the technology's eventual impact. European Central Bank economists have applied that frame to AI, calling conditions in the current market "worrisome" in an analysis that draws on historical precedent.
The ECB analysis rests on comparison with prior technology cycles. When a technology genuinely reshapes economic output, early equity pricing tends to run ahead of when that output arrives. Valuations can be a reasonable bet on the eventual destination and still give back ground, because the path takes longer than models built at the peak of enthusiasm assume.
The economists' point separates two questions that markets tend to conflate. Whether AI delivers economic gains at scale is one question. Whether current prices reflect a timeline that actually holds is another. On the second question, the ECB economists' read is cautious: history suggests that even valuations which accurately capture a technology's long-run power will tumble before that power fully arrives.
For equity markets pricing in broad AI-driven productivity gains, that distinction carries weight. The ECB analysis does not argue that the AI thesis is wrong. It argues that being right on the thesis has not, historically, been enough to avoid a correction.