The current concentration of the US stock market has reached an alarming level, and many early warning signals have appeared. The top eight AI-related companies account for nearly 30% of the S&P 500 index.
The top 30 companies account for 43% of the S&P 500, and in the past two years, excluding these 30 companies' gains, the U.S. stock market gains were zero.
This data shows that the upward momentum of US stocks is highly dependent on a few AI concept stocks. This concentration is reminiscent of the bubbles of the past and exacerbates systemic risks.
Nvidia, as a leading company in the AI wave, has shown weakness in its stock price recently. While US stocks are constantly rising highs, Nvidia fell against the trend. This is the global leader in artificial intelligence. If Oracle had not given the questionable US$500 billion order to give the entire chip sector a lot of money, then the performance would have been even worse.
Goldman Sachs clearly pointed out that the core risk of the current AI trading market lies in its high dependence on capital expenditures of technology giants. Analysts expect capital expenditure growth to slow significantly in the fourth quarter of 2025 and 2026. In extreme stress testing scenarios, if tech giants' capital expenditures fall to 2022 levels, the expected sales of the S&P 500 in 2026 will fall by about 30%.
Compared with the historical bubble period, the current valuation of AI stocks is already at an extremely high level. The top five technology giants in the market value include Nvidia, Microsoft, Apple, Google and Amazon. The current average expected price-to-earnings ratio is 30 times. Of course, Tesla, among the seven sisters, has remained above 100 for a long time, and it is entirely supported by the concepts of artificial intelligence and robots. If Tesla is added, it will be the highest in the stock market in history.
September may become a historic turning point in the market, and multiple factors are gathering: the growth rate of capital expenditure of technology giants slowed down, the Fed's policy shift, and the heating of geopolitical risks.
Soft employment data consolidates traders' bets on the Fed's interest rate cut, but also raises deep concerns about the overall economic situation. Weak economy means slowing down in corporate profit growth, which is not good for the stock market, which is already expensive.
Take a step back, even if artificial intelligence is really a technological revolution, then it should be the downstream application side. The construction of upstream chip and computing power centers has come to an end. The slowdown in capital expenditure is 100% certain, and the production capacity on the production side is still increasing significantly, and the oversupply is almost 100%. This is the biggest risk.
