Is It Too Late to Buy Micron?

Micron has surged 157% year to date and 693% over the past 12 months as AI-driven demand has tightened memory supply. The article says Micron’s PEG ratio is 0.75, below the 1.0 fair-value threshold, and notes Deutsche Bank raised its price target to $1,000 while keeping a buy rating. The core message is that the memory shortage could persist through 2030, leaving further upside despite the stock’s big run.

Analysis

The market is treating memory not as a cyclical commodity, but as a quasi-structural AI bottleneck, and that is the key second-order shift. If AI capex remains concentrated in high-bandwidth memory and advanced DRAM, the profit pool migrates upstream to the handful of suppliers with tight process control, which means pricing power can persist even if end-demand for GPUs normalizes. That dynamic also compresses the window for new supply to matter: wafer starts and node transitions lag demand by quarters, so the near-term setup remains more about allocation discipline than pure unit growth.

The bigger risk is not that Micron’s fundamentals weaken tomorrow; it is that the market is discounting a very long shortage narrative that can get disrupted by supply response, inventory restocking, or AI capex pauses. Memory is notoriously late-cycle when customers over-order, and the highest-margin phase usually ends before visible earnings peaks, so investors should watch for signs of lead times stabilizing, contract pricing flattening, or capex guidance inflecting up from competitors. If those signals emerge over the next 3-6 months, the multiple can compress faster than EPS grows.

The cross-asset read-through is that the AI infrastructure trade is becoming more selective. NVDA remains the demand anchor, but the incremental beneficiary here is the memory layer, while software and lower-intensity semis may lag if customers keep prioritizing hardware budget toward the bottleneck components. For DB, the relevance is indirect but real: if the market keeps rewarding AI-exposed industrial and tech balance sheets, European banks with limited AI leverage may continue to underperform versus the broader growth basket.

Consensus may be underestimating how much of the upside is already embedded in the stock, even if the valuation screen still looks cheap. A low PEG in a cyclical industry can be a trap if peak earnings are being extrapolated too far into the future, especially when the market is rewarding duration on the assumption that memory scarcity lasts multiple years. The better contrarian read is that the fundamental story is still bullish, but the margin of safety is narrowing as the stock moves from “mispriced cyclical” toward “crowded secular winner.”

Key Decisions for Investors

• Stay long MU, but prefer call spreads over outright equity for the next 3-6 months: upside remains meaningful if DRAM pricing stays tight, while defined premium limits damage if the market starts pricing peak-cycle risk.
• Pair long MU / short a weaker AI-beneficiary basket with less direct memory exposure over 1-2 quarters; the trade expresses the view that bottleneck suppliers outperform second-order participants as capital gets rationed toward hardware constraints.
• Sell covered calls against existing MU longs into further strength; with the stock already re-rating hard, implied upside may be less attractive than monetizing volatility while keeping core exposure.
• Add NVDA on pullbacks only if memory suppliers confirm ongoing tightness in lead times and pricing; otherwise treat NVDA as the demand proxy with lower near-term torque than MU.
• Avoid chasing INTC and other lagging semiconductor turnarounds here; if the AI capex cycle broadens later, they can work, but the current regime still rewards direct scarcity exposure over catch-up stories.