Tomorrow, after the US market closes on April 29 (after 4:00 AM Beijing time on April 30), four companies will announce their results together: $Amazon(AMZN.US) , $Microsoft(MSFT.US) , $Meta Platforms(META.US) , $Alphabet(GOOGL.US) .

This kind of scene only happens two or three times a year, and the Nasdaq is basically decided by this one night. I don't really want to discuss who can beat expectations—at this level, it would be a surprise if they *didn't* beat. What really determines the stock price is another question: how much more money will be poured into AI this year, and when will that money start turning into revenue. This time last year, Meta got hit with a big red candle overnight because its capex increase was too aggressive. A year has passed, and the market's patience with capital expenditure has actually become shorter, because we've already entered the stage where "returns should be visible."

Microsoft's and Google's cloud businesses are the two hardest pillars in this narrative. As long as growth doesn't drop, the market can accept even aggressive AI investment. Conversely, whoever dares to revise capex upward again in their earnings report without showing corresponding revenue acceleration will be tomorrow's cautionary tale. Amazon is the one I'm least confident about. Consumer data for the retail line has been sluggish for the past two months. Even if AWS holds steady, a single miss on retail profit margins could drag down the entire company. Meta is betting on size, not performance. In short, reduce positions before earnings, don't preset a direction, and keep cash to deal with the two spikes at the open.

On such a night of concentrated earnings announcements, betting the right direction won't earn you much more than usual, but betting the wrong way can make you work a week for nothing.