This year has begun much as 2025 ended, with a steady drumbeat of warnings about an AI bubble.
Investors see an AI crash as the market’s biggest tail risk, according to Bank of America’s latest fund manager survey, while more than half described the magnificent seven as Wall Street’s most crowded trade.
Investor angst is reflected in headlines, ranging from the Wall Street Journal (“Will stocks crash in 2026?”) to Bloomberg (“Is the AI boom a bubble waiting to pop?”) to Futurism (“Terrified investors are bracing for an AI bubble ‘reckoning’”).
Questioning headlines don’t disprove these fears, but their ubiquity complicates the idea that markets are gripped by blind euphoria. Ritholtz Wealth Management’s Ben Carlson notes only two magnificent seven stocks – Nvidia and Alphabet – actually beat the S&P 500 last year, with Microsoft, Apple, Amazon, Meta and Tesla all lagging. Additionally, valuations for most big tech stocks barely moved. Excluding Tesla, multiples were broadly flat, says Carlson, while price-earnings ratios for Nvidia, Broadcom, Amazon, Oracle and Apple actually fell.
Similarly, JPMorgan’s Michael Cembalest says in his 2026 outlook that valuations are “high, but not as high as you might think”. The S&P 500’s forward price-earnings ratio only barely rose in 2025, with almost all the index’s 18 per cent gain attributable to earnings growth. Technology valuations are “roughly half the levels seen at the dotcom peak”. They also rest on persistently high profit margins – and higher profitability is logically linked with higher valuations. Put another way, global tech PEG ratios – valuations adjusted for expected growth – are “not that different” from the overall global equity market.
That does not mean there are no excesses, or that valuations are cheap. Indeed, Cembalest says it is right to ask what could go wrong, given markets are highly concentrated and at all-time highs.
However, it does suggest 2025 was not a blow-off year, and that many bubble warnings may be overstating the case.