In the stock market the other day, some hot sectors—such as memory storage, optical modules, servers, and chips—once they even brushed up against AI, their stock prices dared to front-load profits for years.

It looked completely insane and had lost its mind.

Now, Samsung’s profits have surged, but the stock price still falls; even if SK hynix’s performance is strong, it can’t stop capital from retreating. Many related domestic companies have also reported increased-revenue results, but they still couldn’t save today’s slump.

Because what the market trades is never how much you made yesterday—it’s whether you can keep exceeding expectations tomorrow. When overseas begins to worry that AI capital expenditure will slow down, that storage expansion is becoming excessive, and that returns on compute investment are insufficient, the most crowded AI hardware chain in A-shares naturally gets hit first.

So this drop isn’t in the AI industry itself—it’s the illusion of “AI always deserves a high price.” The real turning point has already appeared:
From now on, the market won’t just listen to stories—it will look at orders; it won’t only focus on revenue—it will look at profits; it won’t only ask whether there’s AI—it will ask whether AI is actually making money. After the tide goes out, what remains won’t be the companies that can tell the best stories, but those that truly have customers, real cash flow, and can truly turn AI into a business.

#AI技术 # industry #bubble