Here's Why I Will Never Pay Anywhere Near $1,100 for Micron Stock

Micron reported fiscal 2026 Q3 revenue of $41.4B (+346% YoY) and EPS of $24.67 (+1,368% YoY), driven largely by AI-related HBM; management guided Q4 revenue of $50B (+342% YoY) and EPS of $30.73 (+985% YoY). Despite the growth, the article flags a valuation disconnect—MU trades at 25.6x trailing earnings (vs. Nasdaq-100 at 34.1x) and a low forward P/E of 7.6x on bullish estimates—arguing AI memory demand may weaken as some companies curb AI spend and route workloads to cheaper models. Net: strong fundamentals but rising risk that data-center/AI spend slows in coming quarters.

Analysis

MU is in the classic late-cycle memory setup where the market is discounting near-term scarcity but not durable scarcity. The key mechanism is operating leverage: when pricing is tight, every incremental bit of HBM margin flows through aggressively, but the same cost structure cuts the other way once supply normalizes. That makes the equity vulnerable to any hint that hyperscaler budgets are shifting from “build more” to “optimize what we already bought.”

The bigger second-order effect is on the broader AI hardware stack. If memory availability improves, NVDA and the system vendors get relief on bill of materials and can ship more units, but only if end demand holds; if end demand weakens, better memory supply just accelerates the pricing reset. On the software/cloud side, GOOGL and MSFT are better positioned to absorb customers trading down to cheaper models, but that behavior is still negative for memory intensity and capex growth, so it is a leading indicator to watch for MU rather than a direct long signal for the cloud names.

Contrarian risk: the market may be underestimating how long HBM tightness can persist if all suppliers are capacity-constrained at the same node. That creates squeeze risk for any short built purely on “memory always mean-reverts.” The falsifiers are straightforward: another quarter of firm HBM pricing and no inventory build would argue the cycle is still early; a capex downtick from MSFT/GOOGL or any guide reduction from AI infrastructure buyers would validate the bearish thesis within 1-2 quarters.

Key Decisions for Investors

• No outright long MU here; wait for either a spot-memory rollover or a hyperscaler capex guide cut before adding risk. Time horizon: next 1-3 months.
• If expressing a bearish view, use a defined-risk MU put spread 30-60 days out rather than a naked short. The thesis only needs modest multiple compression on any evidence of AI spending digestion.
• Relative-value watch: long NVDA / short MU only if HBM supply remains constrained but final demand stays intact. If hyperscaler capex weakens, avoid this pair because both legs can fall together.
• Monitor MSFT and GOOGL next capex commentary as the key falsifier for MU. A lower capex trajectory is the cleanest signal that the current earnings run-rate is peak-cycle, not structural.