A Game of Thumb investor named the top two risks to the AI sector in 2026

US investor Steve Iceman, who predicted the 2008 global financial crisis and capitalized on it, has warned of two critical problems that could bring down stocks in the artificial intelligence sector next year. The market participant, who became a prototype for one of the main characters in the movie "The Downside Game," said in his YouTube podcast that the AI industry faces not only structural limitations, but also the risk of its basic hypothesis being wrong.

Iceman considers the structural risk to be a shortage of electricity: consumption growth in the U.S. has accelerated from 1% to 3% per year, which requires the commissioning of facilities equivalent to the power consumption of several major cities. At the same time, the process of designing and building power plants takes at least two to three years, and there are already precedents for new data centers in Silicon Valley going without power. If the lack of electricity slows down the introduction of new computing capacity, cloud giants will begin to reduce orders for chips, and "the entire growth cycle will reverse," the investor believes.

The intellectual risk, Iceman argues, relates to the possible dead end of the strategy of infinitely scaling neural networks. In his view, just as the entire mortgage bond industry rested on the belief that U.S. home prices would never fall, so the current frenzy of demand for AI hardware holds on the assumption that the larger language models become, the better they perform. "The first chink in that armor" came when OpenAI released ChatGPT-5 and that one didn't turn out to be much better than its predecessor, the investor noted. If the view that further scaling of neural networks is futile gains traction, the market flywheel could spin in the opposite direction, he warned.

Professional investors see an AI bubble as the main market risk of 2026, a Deutsche Bank survey has found. This opinion was expressed by more than half of 440 asset managers. Jim Reid, Head of Global Economics and Case Studies at Deutsche Bank, noted that this scenario dominates all other threats by a record margin. At the same time, the economist emphasized that the current level of market overheating is still lower than what was observed during the post-pandemic liquidity glut in 2021.

Amid doubts about the justification of multi-billion-dollar AI spending, Wall Street's major indexes - the S&P 500 and Nasdaq - plunged to three-week lows in the middle of last week. Baird Private Wealth Management strategist Ross Mayfield noted at the time "growing anxiety about trading in the AI sector." "The main driver is the level of capex and the looped nature of some of the spending, with OpenAI at the center of it. As we head into the New Year, the more global issue is around the sustainability and return on all this investment," he added.

This article was AI-translated and verified by a human editor