Above the Noise: Perspective on bubble fears

AI investment is not yet evidence of a bursting bubble because earnings, demand and market valuations do not currently show the excesses seen in 2000.

• AI investment has surged alongside earnings, constrained chip and memory supply, and rising revenues for model providers, while the S&P 500’s price-to-forward earnings ratio is roughly one third of its technology-bubble peak.
• CoreWeave’s higher borrowing costs reflect the risks of a highly leveraged neocloud business; Gartner forecasts that more than 80% of neocloud providers could cease operating independently by 2030.
• The stress around CoreWeave has not spread broadly: investment-grade and high-yield credit spreads have not widened significantly, hyperscalers appear better placed to finance AI investment, and the S&P 500 ended the week of October 16 only 2.5% below its record high.
• The rise in 10-year Treasury yields was likely not primarily a debt-sustainability shock because real yields rose without a comparable increase in inflation expectations; Treasuries later rallied as scrutiny of the AI trade increased.
• Midterm elections have historically mattered less than growth, earnings, inflation and Federal Reserve policy: since 1960, the president’s party lost an average of 22 House seats, while stocks gained an average of 16% in the following year.

Takeaway: Investors should distinguish financially stronger AI beneficiaries from highly leveraged weaker players, monitor leverage, inflation expectations, credit spreads and lending standards, and avoid assuming that the AI excess has already reached its breaking point.