A growing chorus of City analysts is warning that the AI stock rally may be running on borrowed time. With valuations stretched to historic extremes and revenue growth failing to keep pace, the question is no longer if a correction will come, but how severe it will be.
The numbers are stark. The Nasdaq has surged nearly 40% over the past year, driven largely by a handful of AI-focused giants like Nvidia, Microsoft, and Alphabet. Nvidia alone has seen its market capitalisation balloon past $3 trillion, making it more valuable than the entire FTSE 100. Yet its price-to-earnings ratio hovers above 70, more than triple the historical average for tech stocks.
“We are witnessing a classic bubble narrative,” said Helena Chen, equity strategist at Barclays. “Hype around generative AI has created unrealistic expectations. Many companies are pricing in future dominance that may never materialise.”
Indeed, the disconnect between stock prices and underlying fundamentals is widening. A recent analysis by Morgan Stanley found that while AI-related capital expenditure is soaring, actual revenue from AI products remains modest. Cloud providers are spending billions on GPU clusters, but only a fraction of that spend is translating into customer sales. For every dollar invested in AI infrastructure, only about 20 cents is being recouped in revenue, according to the bank’s estimates.
The regulatory landscape adds another layer of uncertainty. The EU’s AI Act, which came into force this month, imposes strict compliance costs on high-risk systems. In the US, the Federal Trade Commission has signalled tougher scrutiny of AI monopolies. “Regulation is the wild card,” noted James Patel, tech policy analyst at Renaissance Research. “If lawmakers clamp down on data scraping or model transparency, it could slash profit margins overnight.”
Yet bullish investors argue that this is different from the dot-com bust. AI, they say, is a genuine productivity revolution, not a fad. “The internet bubble burst because companies had no path to profitability,” said Sarah Kim, portfolio manager at Fidelity. “Today’s AI leaders generate massive cash flows and have real revenue diversification.” Nvidia, for instance, reported $26 billion in quarterly revenue last month, up 262% year-on-year. But critics point out that such growth rates are unsustainable; eventually, hyperscalers will pause their GPU buying spree.
Central bank policy could be the catalyst. The Bank of England and Federal Reserve have signalled they will keep interest rates higher for longer to tame sticky inflation. Higher rates compress valuations for growth stocks, and AI shares are the most vulnerable. The recent rotation out of tech into value sectors suggests the market is already hedging its bets.
What would a correction look like? Historical precedents are sobering. The Nasdaq lost 78% of its value from 2000 to 2002. Even a mild re-rating could see AI stocks fall 30-40%, wiping out trillions in market cap. But Chen believes the adjustment will be more orderly this time. “Institutional investors have better risk management tools now. We may see a series of sharp but contained drawdowns rather than a crash."
For the average retail investor, the advice is caution. Diversify, avoid leverage, and question the narrative. As one hedge fund manager put it: “When your cab driver tells you to buy Nvidia, it’s time to sell.” The AI bubble may not burst tomorrow, but the air is definitely hissing out.
The coming months will test whether this is the dawn of a new industrial age or just another chapter in the old story of financial euphoria. Either way, the reckoning is coming.
