Christmas market meets AI concerns: Wall Street faces a reality check on artificial intelligence trading

As 2025 approaches its end, Wall Street stands at a crossroads. On one hand, the traditional “Christmas rally” in December has historically had a very high success rate; on the other hand, AI trading, which has driven the US stock market higher over the past few years, is facing unprecedented skepticism. This contradiction causes investors to hesitate between chasing gains and guarding against a pullback.

Historical data shows that the Christmas rally typically covers the last five trading days of December and the first two trading days of January. Since 1929, there is about a 79% probability that the stock market will rise during this period, with an average return of approximately 1.6%. In the past eight years, there has only been one decline. However, the market’s strong consensus on this pattern has instead sparked concerns. Some investors believe that overly “crowded” trades tend to yield diminishing returns and may even trigger a market reversal.

The fatigue of risk assets has already begun to show. Bitcoin is currently priced at around $89,460, down nearly 7% over the past month, failing to hold above the important $95,000 level. This weakness is seen as a signal of a cooling overall risk appetite.

Deeper focus remains on the artificial intelligence sector. Over the past three years, AI narratives have driven the market capitalization of the S&P 500 to grow by about $30 trillion. Recently, however, from Nvidia’s correction and Oracle’s stock decline after disclosing high AI investments, to market sentiment cooling towards OpenAI-related companies, investors are beginning to reassess the return cycle. Alphabet, Microsoft, Amazon, and Meta are expected to invest over $400 billion in data center construction in the next year, with depreciation costs rising rapidly, yet returns have not fully materialized.

Surveys show that less than half of AI projects have achieved returns above costs, but most companies still plan to increase their investments. The disconnect between these investments and expected returns has become the core question in the AI trading debate.

Nevertheless, directly comparing the current situation to the internet bubble of 2000 still seems exaggerated. The Nasdaq 100’s P/E ratio is about 26 times, significantly lower than the extreme levels of that time. In the short term, seasonal factors and FOMO sentiment may still support the market; but entering 2026, whether AI investments can deliver profits will be a key variable in determining the market’s direction.