Overview
- Capital Economics published a note on Sept. 19 that outlined a downside scenario in which an AI-driven collapse would produce at least a 30% peak‑to‑trough fall in the S&P 500 and leave the index near 6,500 by the end of 2027.
- The firm emphasized this is a scenario, not a prediction, and said the U.S. market would be the epicenter with smaller losses abroad because other markets are less concentrated in large AI‑exposed tech firms.
- Analysts point to recent Fed tightening as a plausible trigger that could force selling by raising borrowing costs for highly leveraged companies and by making speculative valuations harder to justify.
- Large technology 'hyperscalers' have issued roughly $220 billion of debt over the past year to finance AI buildouts, which raises those firms’ sensitivity to higher rates and could amplify market stress if sentiment turns.
- Valuations are cited as extreme—CAPE near 41—prompting comparisons to the dot‑com era and warnings from investors like Ray Dalio; if a severe bust occurred, the Nasdaq/QQQ could take many years to recover while bond yields and the dollar may move only modestly.