Prediction: Here's How Much Micron Stock Will Be Worth in 2 Years

Micron has surged on AI-driven memory demand, but the article argues the stock could fall below its current $899 price within two years if the memory cycle peaks in 2028. The author estimates Micron could trade around 4.5x trailing earnings, implying about $517.50 per share if EPS reaches $115, with a possible range of roughly $275 to $930. The core concern is that massive capacity additions by Micron and peers may eventually trigger oversupply and normalize pricing.

Analysis

The market is implicitly pricing a continuation of scarcity economics well past the point where the supply response becomes self-defeating. In memory, capacity additions do not arrive gradually; they arrive in waves, and once utilization normalizes the incremental margin on each shipped bit collapses faster than investors expect because fixed-cost absorption works in reverse. The key second-order effect is that the current AI buildout is likely pulling demand forward, not eliminating cyclicality, so the next downcycle may start from a much higher nominal earnings base but still produce a severe multiple reset.

The more interesting issue is competitive positioning inside the AI stack. HBM scarcity benefits not just the memory suppliers but also GPU vendors and system integrators that can secure allocation first, while smaller cloud and OEM customers become price takers. That can subtly widen the gap between hyperscalers with procurement scale and everyone else, because memory access becomes a strategic input rather than a commodity purchase; the end result is better share for the best-capitalized AI platforms and more pressure on laggards to delay or downsize deployments.

The bearish setup on MU is time-dependent: the stock can stay expensive for months if capacity remains tight and customers keep pre-buying, but the risk/reward worsens materially into 2027 as capex turns from narrative into deliverables. The real catalyst for multiple compression is not simply a miss, but the first sign that lead times, spot pricing, or customer order behavior are inflecting ahead of the expected 2028 peak. If that happens, the market will likely de-rate MU before earnings actually roll over, just as it has in prior cycles.

The contrarian view is that this cycle may deserve a structurally higher trough multiple than history because AI data centers create a more persistent replacement demand floor than prior smartphone/PC cycles. Even so, the burden of proof is on the bulls: a higher floor can support earnings, but it does not eliminate supply-driven overshoot. At current valuation, the stock is being priced as if the new floor is already validated; that looks premature.

Key Decisions for Investors

• Initiate a tactical short in MU over 3-6 months, preferably into strength, with a stop above the recent momentum regime; thesis is multiple compression before earnings compression, offering asymmetric downside if supply commentary turns less tight.
• Pair trade: long NVDA / short MU for 6-12 months. NVDA retains scarcity pricing at the system level while MU is the most direct recipient of future overbuild risk; this isolates AI demand while hedging against memory-cycle mean reversion.
• Buy MU 12-18 month put spreads funded by selling far OTM puts. This captures the left-tail risk of a 2027-2028 de-rating while recognizing that the stock can remain elevated until capacity additions become visible.
• For long-only accounts, reduce MU sizing and rotate part of the exposure into suppliers with less direct commodity cyclicality. The risk/reward on MU deteriorates as capex converts from bullish headline to eventual supply overhang.
• Set a trigger to reassess if HBM lead times shorten or customer ordering normalizes before year-end; that would be the earliest high-signal indicator that the cycle is peaking earlier than consensus.