After Surging 918%, Is Micron Stock Still a Buy? Here's What History Says.

MU is functioning less like a cyclical memory name and more like a near-term capacity monopoly in a constrained AI supply chain. The second-order effect is that HBM scarcity is no longer just supporting pricing; it is forcing hyperscalers and accelerator vendors to pre-commit demand, which should keep contract terms seller-favorable for multiple quarters even if end-demand wobbles. That makes the earnings path unusually visible relative to the rest of semis, but it also concentrates risk: any sign of 2027 supply normalization or customer qualification delays would hit the multiple faster than the earnings line.

The market is likely underestimating how much MU’s strength can bleed into NVDA while pressuring AVGO’s relative positioning. NVDA benefits from having a confirmed memory ecosystem behind its next platform, reducing execution risk for Rubin and making the roadmap more financeable for cloud customers. By contrast, AVGO’s reaction suggests investors are starting to separate “AI beneficiary” from “AI priced for perfection”; that should continue to favor names with direct unit bottlenecks and visible backlog over diversified infrastructure stories.

The contrarian read is that this is not a cheap stock in the classic sense, it is a cheap stock on peak-margin assumptions. A forward P/E around 10 is only compelling if the market believes the current pricing regime persists long enough to de-lever fixed costs and fund capacity without triggering oversupply; if margin normalizes even modestly, the multiple can re-rate against a lower earnings base. The next catalyst window is the next 1-2 earnings prints: guidance quality matters more than the quarter itself, because the stock is now trading on the durability of HBM tightness, not just the existence of it.