An AI Boom: Massive Investment, Overvalued Stocks, and the Likelihood of a Stock Market Bubble
The disruptive technology of AI is rapidly transforming the economy and revolutionising our lives in the process. Tech stocks have boomed, with massive capital investment in AI infrastructure, particularly in data centres and semiconductor factories. Optimistic projections for the future come with the caveat, however, that the returns on investment may not match expectations, as supply exceeds demand for AI.
In early November 2025, the AI giants lost billions in market value, only to rally by 10 November. Spooked stock markets saw a sell-off in AI stocks, which have accounted for 75% of the S&P 500’s returns and also dominate the tech-heavy Nasdaq Composite index. Michael Burry, founder of Scion Asset management, shorted NVIDIA and Palantir by placing a $1.2 billion bet against the tech giants. Regulatory filings (13Fs) by Wall Street’s largest hedge funds showed that they had reduced their holdings of some Magnificent Seven stocks, including NVIDIA, Amazon, Alphabet, and Meta during the third quarter (Q3), which ended on 30 September 2025. Japanese technology investor Softbank dumped its entire stake in NVIDIA, as did Peter Thiel’s hedge fund, Thiel Macro LLC, even though the chipmaker had become the first company ever to be valued at $5 trillion. Not all investors thought along similar lines. “Oracle of Omaha” Warren Buffett’s Berkshire Hathaway thus purchased 17.8 million shares of Alphabet during the same period. Further capital injections and circular financing arrangements by Big Tech companies themselves have helped artificially prop up stock prices, which ultimately reflect investor expectations and future performance. The fundamental value of a stock is indeed defined as the discounted sum of “expected” future dividend, which is a predictor of future cash flows.
Comparisons have been made with the dot-com bubble of 1996 to 2000, when overinvestment on untried and untested concepts that failed to deliver led to a crash, followed by a brief recession in 2001. On this occasion, the situation is somewhat different. The technology is well-tested, and the Magnificent Seven companies are well-capitalised. Only some AI start-ups remain vulnerable. Growth in the AI sector is at present constrained by limits to growth in computational capacity. Data centres consume prodigious amounts of electricity, as well as water, and regulatory frameworks and existing infrastructure (power grids) are failing to keep up with the demand. Then there’s also the matter of increased greenhouse gas emissions.
AI is here to stay, but the explosive growth in AI may not necessarily translate into meaningful productivity gains, commensurate with its heavy consumption of data, money, and energy. Market concentration in intangible AI, to the detriment of heavy manufacturing and other sources of tangible goods, does not bode well for sustained economic growth and the wellbeing of citizens. Today’s stock mania might still give rise to an investment bubble, followed by a market correction. Although this is unlikely to happen in the short term, the impact of such a bubble would be devastating, given the current dominance of AI in the stock markets, the huge amounts invested in tech stocks, and the growing dependence of the economy on AI spending. Only the passage of time will confirm, on way or another, whether we are on the right trajectory, or not.
Ashis Banerjee