Something happened in the Nasdaq yesterday that should give every AI stock investor pause. Fifty-six out of 100 stocks in the Nasdaq 100 finished higher — and yet the index closed down 2.2%. AI stock earnings expectations, it turns out, can’t be carried by the majority when a handful of heavyweight semiconductors are in freefall.
Every growth stock goes through at least two distinct phases. In the first, the story carries the numbers. Investors buy the vision: total addressable market, disruption potential, the slope of the adoption curve. Valuation multiples expand because the narrative is compelling and the future feels wide open. This is where AI stocks spent most of 2023 and 2024.
The second phase is harder. The story no longer does the lifting — the numbers have to. Revenue must grow fast enough to justify the multiple. Margins need to show up. Guidance has to be credible. It isn’t that the story stops mattering; it’s that the story now needs proof of life. Investors stop buying the pitch and start grading the receipts.
That shift appears to be underway. Micron is down 13% not because the AI memory thesis is wrong, but because a stock up 760% in a year enters every earnings call with the burden of proof on its side. SK Hynix’s reported consideration of HBM4 production cuts lit the fuse in Seoul — and the selloff spread to Philadelphia before New York opened. When sentiment is fragile, a whisper of supply caution becomes a shout.
Tonight’s Micron earnings are the first real checkpoint. Strong memory demand figures give the AI infrastructure story something concrete to point to. A miss — or cautious guidance — risks turning a two-day correction into something more sustained.
For investors, the question isn’t just whether you believe in AI. It’s which phase you think each AI-related stock is in. Buying a narrative is a different bet from buying a business that’s beginning to prove one. Knowing the difference is at least as important as picking the right sector.
My take: The story used to carry the numbers; now the numbers have to carry the story.
Not advice. Just how I see it.