‘70% chance’ we’re already in an AI bubble: The economist who said ‘not yet’ in January now sees all four horsemen of the bubble apocalypse
Key Points:
- Lamont, a portfolio manager and former finance professor, estimates a 70% chance that the U.S. stock market will be in an AI bubble by November 2026, citing the presence of his four bubble indicators: overvaluation, bubble beliefs, issuance, and inflows of new money.
- A key change this year is the shift from corporate stock buybacks to net stock issuance, with U.S. companies issuing a net $2.4 trillion over the past 12 months, largely driven by SpaceX’s $2.2 trillion IPO, marking a reversal from previous years.
- Major tech companies like Alphabet, Meta, and Tesla have moved from buying back shares to issuing new stock to fund AI investments, signaling potential overvaluation or a need to finance AI buildouts, though the exact motivation remains uncertain.
- Professional investors are reducing their tech exposure sharply, while Lamont notes that net issuance is still below bubble peaks seen in 1999–2000 and 2021, and he stops short of calling a market top.
- Lamont identifies three conditions that would confirm the bubble: a successful Anthropic IPO, a surge in smaller IPOs similar to 2021 levels, and valuation ratios exceeding previous records; the Anthropic IPO imminent this week is a key event to watch.