AI Stocks vs. Dot-com Bubble: CAPE at 38, Concentration Above 2000 Levels, but Companies Are Actually Profitable

April 25, 2026

TL;DR

Shiller CAPE ratio is at 38-40, second-highest in 155 years, only surpassed by the dot-com peak.
S&P 500 top-10 company concentration is nearly 50% higher than at the dot-com peak.
Unlike dot-com predecessors, current AI companies like Nvidia are highly profitable, with Nvidia alone earning $120 billion in net income.
Hyperscalers are investing $660-690 billion annually in AI infrastructure, representing the largest corporate investment program outside of wartime mobilization.
The core question is whether this massive capital expenditure will generate returns that justify the investment.
Factors associated with bubble conditions like retail investor euphoria and speculative capital concentration are present.
The key difference from 2000 is the profitability of leading tech companies, which generated significant free cash flow.
The resolution depends on the long-term return on AI infrastructure investments; if successful, current valuations may be justified, otherwise, a crash similar to 2000 could occur.

Continue reading the original article