The better the results get, the harder it becomes to surprise investors.
Samsung's second-quarter profit was up more than 1,800% year-over-year. The stock fell anyway.
The business isn't struggling. The market just started asking a different question:
For three years, the question was: who's building the most? Now it's: what's actually coming out the other side?
That's why Samsung's number didn't matter. Scarcity was the engine behind this trade. Everyone wanted more computing power than the industry could supply, so spending kept looking justified. You didn't need a payoff yet. You just needed to be building.
Then Meta revealed something that challenged that assumption. It has more AI computing capacity than it needs and is looking for buyers for the excess.
One line in an earnings call can do real damage because if Meta has spare capacity, the shortage everyone priced in is no longer guaranteed. Take away the guarantee and "we spent a lot" stops meaning much on its own.
That's the wall Samsung hit. The market wasn't grading last quarter. It was grading whether the growth still justifies the spending behind it. And this time, "good" wasn't enough.
The same thing is playing out across chipmakers, hyperscalers, and the broader supply chain. Nobody thinks AI stopped working. People just got tired of waiting to see it pay off.
Growth used to be its own reward. Now it's the opening bid.
The question was always there. It just took a spare GPU to bring it into focus.