Why Andy Constan Says The AI Bubble is in Earnings, Not Price

Summary

Andy Constan argues that the AI bubble is in unsustainable earnings, not in price, and that earnings expectations cannot be supported by GDP growth. He discusses historical bubbles, the concentrated bet among hyperscalers, frontier models, and semis, and advises against shorting bubbles. He also shares his failed S&P short and highlights that timing the pop is extremely difficult.

• Andy Constan distinguishes the current AI bubble as an earnings bubble rather than a price bubble.
• He explains that S&P 500 earnings expectations of $400 billion exceed plausible GDP-derived corporate profits.
• Over 62% of earnings growth is expected from AI stocks, which he views as unsustainable.
• He compares the current situation to past bubbles (1987, 2000, 2008, government bonds).
• Constan advises against shorting a bubble and recommends trailing stops for those who want to play it.
• He mentions that policymakers have not acted to cool the economy, keeping inflation elevated.
• He references a trade idea from Dean Kernut involving selling OTM calls and buying OTM puts on SanDisk/Intel.
• Constan acknowledges his own failed short of the S&P 500 near 7000 and warns of the difficulty of timing bubble pops.