Wall Street sees warning signs of an AI bubble, but valuations tell different story

Wall Street’s growing anxiety about an artificial intelligence stock bubble is colliding with an inconvenient fact: Many of the technology companies driving the rally are considerably cheaper relative to expected earnings than their predecessors were during the dot-com boom. The debate has intensified after the Nasdaq 100 (NDX), which reached record highs earlier this month, tumbled 5.5% last week amid renewed concerns about the sustainability of AI investment, Bloomberg News reported Sunday.

For investors, the distinction matters because today’s market risks may have less to do with excessive stock valuations than with the enormous borrowing, concentrated market leadership and ambitious earnings expectations supporting the AI expansion. A slowdown in corporate spending on AI infrastructure, disappointing returns on those investments or tightening credit conditions could threaten technology shares even without the extreme valuations that preceded the 2000 market collapse.

Valuations offer conflicting signals

The five largest technology companies in the S&P 500 (SP500) trade at an average forward price-to-earnings ratio of approximately 23, compared with about 65 for their counterparts shortly before the dot-com bubble burst in 2000, according to Bloomberg. Forward earnings multiples for the S&P 500 (SP500) and Nasdaq 100 (NDX) have declined this year as analysts anticipate substantial profit growth.

However, a longer-term valuation measure offers a more troubling comparison. The cyclically adjusted price-to-earnings ratio, which compares stock prices with a decade of inflation-adjusted earnings, stands near 41, approaching its 1999 peak of 44. Nicholas Colas, co-founder of DataTrek Research, has identified the elevated measure as a particularly concerning signal.

Debt, market concentration raise concerns

Michael O’Rourke, chief market strategist at JonesTrading Institutional Services, argues that declining earnings multiples among major technology companies may reflect the financial risks associated with their aggressive AI infrastructure investments. He has raised concerns about borrowing and financing arrangements in which technology suppliers help fund purchases of their own products.

Meanwhile, the market has become increasingly dependent on a handful of companies. Nvidia (NVDA), Apple (AAPL) and Microsoft (MSFT) collectively represent approximately 21% of the S&P 500’s market capitalization, a concentration that required nearly 10 companies to match in October 1999.

Investor enthusiasm hasn't reached consensus extreme

Howard Marks, co-chairman of Oaktree Capital Management, sees evidence of speculative behavior but remains unconvinced that the market has entered a full-fledged bubble. Although investors appear reluctant to miss further technology gains, widespread discussion of bubble risks suggests they remain conscious of the potential dangers, according to Bloomberg News.

Market sentiment has also been tempered by inflation, geopolitical tensions and economic uncertainty. The Nasdaq 100 (NDX) has experienced several pullbacks this year, including a correction of at least 10%, complicating comparisons with the unrestrained enthusiasm of previous speculative episodes.

Nevertheless, some developments resemble earlier periods of market excess. Broadcom (AVGO) Chief Executive Hock Tan has criticized speculative investors who buy AI-related shares without adequately understanding the underlying businesses.

Other market observers have pointed to companies adopting AI-oriented names, unusually strong momentum in selected stocks and growing differences between market winners and losers. Despite these warnings, the Nasdaq 100 (NDX) has outperformed the S&P 500 (SP500) by roughly 250 percentage points over the past decade, making technology exposure difficult for investors to abandon.

Credit markets could hold key

The financial system may provide a more important warning signal than stock prices alone. Brian Reynolds, chief market strategist at Reynolds Strategy, has identified credit-market disruptions as a potential catalyst for ending the expansion, particularly if borrowers encounter financing difficulties or investors face margin calls. AI infrastructure provider CoreWeave (CRWV) illustrates those vulnerabilities. Its shares nearly doubled between January and early May before last week’s selloff erased those gains and pushed the stock into negative territory for the year, while yields on its debt climbed.

The central question is whether AI-related profits can ultimately justify the capital being committed to the technology. Previous technological revolutions, including the commercialization of the internet, produced lasting economic benefits even as investors suffered substantial losses from overvalued companies.

For now, the combination of strong earnings expectations, mounting financial commitments and unusually concentrated stock-market gains leaves investors confronting a difficult possibility: Artificial intelligence could transform the economy while still delivering disappointing returns to shareholders who pay too much for its promise.