The Nasdaq fell 5.5% in a day, the Philadelphia Semiconductor Index plunged 12%, Nvidia fell 9%, and AMD dropped 17%—a selloff like this cannot be explained by “short-term deleveraging” alone, but calling it a “full-scale collapse of the AI trade” would also be going too far.

A more accurate characterization is that deleveraging amplified the selloff, while earnings validation is what is truly shifting the pricing logic. Analysts interviewed by Jiemian News broadly agree: the conditions for a full-scale collapse are not yet in place, but “faith-based pricing” is giving way to “show me the earnings.” Hedge funds’ exposure to AI has fallen to “underweight” for the first time since early 2023, and Goldman Sachs’ prime brokerage data also confirms that selling had already intensified in October. This means positioning was already becoming less crowded before the plunge; Thursday was simply a concentrated release.

In the short term, margin liquidations and systematic selling can magnify losses beyond what fundamentals warrant. But over the medium term, the gap between AI infrastructure’s “cash burn” and its ability to generate cash won’t automatically close just because of one bout of deleveraging. My view: this was a “stress test” triggered by deleveraging, and its result is that the AI trade has officially entered its late-stage washout phase. Chip stocks have high short-term sensitivity, but the lowest earnings visibility; if you want to participate, waiting for volatility to subside and for the next earnings report to show “who is actually making money” is more practical than trying to call the bottom now.

#每周来晒 #美股史诗级抛售