Root Conditions. Escalation. Peak | Andy Constan on What Five Bubbles Taught Him About AI

| Stock Investing | May 13, 2026 | 9.98 Thousand views | 1:04:10

TL;DR

Andy Constan argues that markets have entered the 'peaking phase' of an AI-driven bubble, drawing parallels to five historical bubbles he has experienced, emphasizing that while bubble regimes are identifiable through root conditions and escalation events, timing the exact peak remains impossible due to human nature and market mechanics.

💥 The Three Phases of Bubble Formation 3 insights

Root conditions create fertile ground

Major productivity shifts, regulatory changes, or technological breakthroughs—like the AI/ChatGPT moment in January 2023—establish the foundation for potential bubbles, though they don't guarantee one will form.

Escalation events trigger parabolic moves

Specific catalysts transform favorable conditions into full bubble regimes, such as the Fed easing financial conditions during the March 2023 SVB banking crisis despite 62 months of above-target inflation.

Peaking phases can persist longer than expected

Even after warning signs appear, markets can rally significantly; the NASDAQ surged 60% in the six months following the October 1998 LTCM bailout before reaching its ultimate 2000 peak.

📉 Historical Parallels and Mechanical Risks 3 insights

The 1987 crash was a bubble despite popular perception

Stocks gained 31% before the October crash and ended the year flat, driven by LBOs, financial deregulation, and portfolio insurance mechanics that accelerated both the ramp and the collapse.

Modern options activity echoes 1987 dangers

Zero days-to-expiration (0DTE) options today rhyme with portfolio insurance strategies that exacerbated the 1987 crash through mechanical selling dynamics.

Bond bubble followed COVID acceleration

Post-GFC zero interest rate policies created a bond bubble that went parabolic during the pandemic shutdown before resolving rapidly in the following year.

🧠 Human Nature and Market Reality 3 insights

FOMO is the engine of bubble regimes

Bubbles persist because human nature makes investors susceptible to envy when peers achieve sudden wealth through semiconductor or AI stocks, creating unsustainable behavior.

Markets are always in equilibrium

Constan rejects 'overbought' or 'oversold' concepts, arguing that at any moment, investors hold exactly what they want to hold, making price levels inherently neither expensive nor cheap.

The holy grail doesn't exist

Successfully timing bubble tops is impossible, and investors should default to passive, well-diversified portfolios rather than attempting to predict when regimes will shift.

Bottom Line

Investors should recognize the current AI bubble's peaking phase but avoid trying to time the top, instead maintaining passive diversification while acknowledging that human nature and mechanical market risks make bubbles unpredictable and dangerous to short prematurely.

More from Excess Returns

View all
We Asked a Renowned Financial Historian What Investors Get Wrong About Bubbles
1:16:30
Excess Returns Excess Returns

We Asked a Renowned Financial Historian What Investors Get Wrong About Bubbles

Financial historian Edward Chancellor explains that historical technology booms from railroads to the dot-com era follow a predictable capital cycle pattern: massive overinvestment driven by competitive 'fear of missing out' leads to fragmented markets and destroyed profits, even when the underlying technology proves transformative for society.

3 days ago · 9 points