Is the US Stock Party Ending? Five Bubble Warning Signs Emerge in AI-Driven Rally

According to Barron's, Capital Economics has warned that the AI-driven rally in US stocks may have entered the final stage of a bubble. Of the eight market indicators the firm tracks, five have already risen to levels close to historical extremes, including valuation, earnings expectations, index concentration, equity issuance, and foreign buying.

However, James Reilly, Senior Market Economist at Capital Economics, believes the rally in US stocks still has room to continue this year. He forecasts the S&P 500 could reach 8,250 by the end of 2026, but as the bubble unwinds, it may fall back to 6,500 by the end of 2027.

The S&P 500 has recently weakened, closing down 2.7% on Thursday (10th) from its record closing high of 7,798.99 points set on August 13. While some of the decline may stem from seasonal weakness in August and September, the index is still up over 10% year-to-date and is on track for a fourth consecutive year of double-digit gains.

Reilly notes that the recent pullback does not necessarily mean the rally will end immediately, but observing a series of bubble indicators suggests the AI stock frenzy may already be in its final phase, with medium-term prospects turning less favorable due to market overheating.

First, US stock valuations remain elevated. The S&P 500's forward price-to-earnings (P/E) ratio is currently around 21x, slightly below the peak of about 23x expected by the end of 2025. The valuation premium of US stocks over other markets has also narrowed. However, just before the peak of the dot-com bubble in 1999–2000, US stocks also experienced a period of relative underperformance, so the recent pullback alone is not enough to rule out a bubble.

Second, current valuations imply very high expectations for corporate earnings growth. The market's consensus forecast for earnings per share (EPS) over the next 12 months has risen to levels similar to those seen at the height of the dot-com bubble. This time, however, it does not reflect a recovery from depressed earnings. Since growth expectations are heavily concentrated in the tech sector and rely significantly on long-term profit realization, the risks are even more pronounced.

Third, the S&P 500 is becoming increasingly concentrated in a few tech giants. Recently, NVIDIA (NVDA-US) and Apple (AAPL-US) each accounted for over 7% of the index, while Microsoft (MSFT-US) has a weighting exceeding 5%. Reilly describes the current level of concentration among large tech stocks in the index as concerning.

Fourth, equity issuance activity is heating up rapidly—a common phenomenon in the later stages of a bubble. If OpenAI and Anthropic proceed with the highly anticipated initial public offerings (IPOs), related indicators could rise further.

Fifth, foreign investors have significantly increased their buying of US stocks, a pattern similar to capital flows observed before past market peaks.

In contrast, three other indicators—corporate fundamentals, market volatility, and leverage levels—have not yet reached extreme levels that would trigger alarms. Reilly therefore does not conclude that a market reversal is imminent, but emphasizes that with valuations, earnings expectations, market concentration, equity issuance, and foreign ownership all flashing red, the AI stock boom may be drawing closer to a turning point.