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

Micron is benefiting from AI-driven memory demand and expects meaningful supply expansion in 2027-2028, but the article argues the cycle may peak in 2028. The author estimates Micron could trade around $517.50 at that point using $115 trailing EPS and a 4.5x P/E, versus the current $899 share price, implying downside risk. The piece is cautious on the stock despite strong current earnings momentum.

Analysis

The market is treating Micron as a linear AI beneficiary, but the real setup is a classic capacity-hedge trap: hyperscalers and accelerator vendors are pulling forward orders today to secure supply, which can keep spot pricing elevated longer than fundamentals justify. That creates a second-order winner set in the near term—equipment, materials, and HBM-adjacent suppliers—but it also front-loads demand into 2026-2028, increasing the odds that the next downcycle is sharper than consensus expects once all three major producers’ capacity lands together.

What the market may be missing is that AI does not eliminate memory cyclicality; it changes the shape of the trough. Even if server refresh and replacement demand create a higher floor, the marginal buyer in a normalized market still prices off supply, not narrative. If pricing starts to weaken in 2027, the equity will likely de-rate before earnings peak, because the stock’s multiple can compress well ahead of EPS inflection—historically the dangerous part is not the earnings rollover itself, but the first quarter where management guidance implies utilization discipline rather than scarcity.

From a relative-value perspective, this is more compelling as a short-vs-long expression than as an outright short. NVDA and other AI compute names may be less exposed to memory oversupply than MU, because their earnings power is tied to compute bottlenecks with stronger software/architecture moats; meanwhile MU carries the capital intensity and balance-sheet risk of a late-cycle supply expansion. The asymmetry improves if the market keeps paying peak-cycle multiples for MU into 2026-2027 while peers with less cyclical earnings maintain premium valuations.

The contrarian case is that the article may be too early rather than wrong: a 2028 supply catch-up could still allow another 12-18 months of exceptional pricing power, and the stock could remain expensive if buybacks and AI memory scarcity extend the runway. But at current levels, the risk/reward appears skewed toward disappointment because the equity is already discounting a very long, very clean supercycle, while the industry structure is almost designed to overshoot into oversupply.

Key Decisions for Investors

• Initiate a staggered short MU position on strength over the next 1-3 months; use a 12-18 month horizon and size for a 25-35% downside case if the market starts discounting 2027-2028 oversupply earlier than expected.
• Prefer a pair trade: long NVDA / short MU as a relative AI-capex expression for the next 6-12 months; thesis is that compute monetization remains supported while memory pricing is more vulnerable to capacity normalization.
• For options, buy MU put spreads expiring 9-15 months out to capture a possible pre-earnings de-rating before the actual earnings peak; target a 2-3x payoff if the multiple compresses ahead of fundamentals.
• Take profits in any leveraged long exposure to the memory chain if lead times or management commentary begin signaling easing scarcity; the first sign of pricing discipline should be treated as an exit trigger, not a buy-the-dip event.
• Monitor equipment and materials names as a tactical long only until 2027 build-out visibility peaks; once capex spend shifts from growth to completion, rotate out because the same capacity wave that helps them eventually becomes the overhang on MU.