Precedent · Market shock
The Dot-Com Bust
2000–2002
The collapse of the internet/telecom bubble. The Nasdaq fell ~78% from its March 2000 peak, roughly $5T in market value evaporated, a mild recession arrived in 2001, and tech endured a multi-year capex winter. A valuation-and-sentiment unwind, not a credit event.
The signature
Each variable's peak deviation from the pre-event baseline, with the curve shape, the lag before it moved, and how long the recovery ran.
| Variable | Peak deviation | Shape | Lag / Recovery | Confidence |
|---|---|---|---|---|
| S&P 500 S&P -49%, Nasdaq -78% peak-to-trough 2000-2002 | −49% | U | 30d lag · 720d | high |
| 10Y-2Y Curve Fed cut hard; curve re-steepened | +120% | Step | 30d lag · 540d | medium |
| Real GDP Mild 2001 recession | +0% | V | 180d lag · 365d | high |
| Unemployment Rate ~4% to ~6% | +50% | Ramp | 120d lag · 540d | high |
Methodology
Encoded against the pre-peak baseline. The signature is a valuation collapse concentrated in growth/tech with a relatively modest macro spillover: equities grind down in a deep U (multiples compress faster than earnings), GDP grazes recession, unemployment ramps from ~4% to ~6%, and the Fed cuts hard so the curve re-steepens. Shapes: U (equities), ramp (unemployment), step (curve).
What's different now
The AI-capex parallel is the reason to read this. The transferable lesson is the valuation-unwind dynamic — multiples can compress for two years even when the technology is real — not the magnitude. Today's AI leaders carry cash flows the 2000-era names never had, so the question is concentration and price, not whether the technology matters.