You Missed Micron's 811% Run — but There's Still 40% More Upside, According to Wall Street

Micron has rallied 811% over the past year, and TD Cowen still sees about 40% upside with a price target raised to $1,500 from $660. The bullish case is based on stronger AI-driven DRAM/HBM demand, tighter memory supply, and pricing strength now expected to last through 2H 2027. Analysts also highlighted a potential $150 per-share earnings run rate in 2027 if current trends persist.

Analysis

The market is still treating memory as a mean-reverting commodity, but AI changes the earnings power equation by making demand both deeper and stickier. That matters because the real margin expansion is not just from higher spot pricing; it comes from mix shift toward HBM and longer customer commitments, which reduce the historical boom-bust whiplash and justify a structurally higher terminal multiple. In other words, MU is transitioning from a cyclical supply beta to a semi-anchored capacity bottleneck behind the AI stack.

The second-order winner is not only MU’s own equity but also the broader memory ecosystem: equipment, packaging, and advanced substrate constraints should keep capex returns unusually high longer than consensus expects. The likely loser set is downstream OEMs and cloud buyers that cannot fully pass through memory inflation, especially if GPU vendors keep prioritizing compute shipments while memory remains tight. BAC’s presence here is a reminder that the financing side of the cycle may still be underappreciated; a prolonged memory upcycle supports semiconductor capex and fee/credit activity around the supply chain, but only if the cycle stays orderly.

The main risk is timing, not direction. If AI server deployment pauses for even one quarter, memory names can de-rate fast because the stock already discounts a very strong 2027-2028 earnings path. The contrarian miss is that the bullish case may be crowded on the obvious upside, while the more powerful trade may be the duration trade: the market may still be underestimating how long scarcity can persist once packaging, power, and foundry constraints are layered on top of nominal memory capacity additions.

Key Decisions for Investors

• Maintain/establish a long MU position on any post-earnings or analyst-driven pullback; thesis horizon 6-18 months. Risk/reward favors owning duration on a supply-constrained AI bottleneck, with upside if the market starts capitalizing 2027 earnings at a software-like multiple rather than a cyclical memory multiple.
• Pair trade: long MU / short a weaker legacy memory proxy or a broader semiconductor ETF slice over 3-6 months. Objective is to isolate the AI-memory scarcity premium versus the rest of the semi complex, which is more exposed to normalization in non-AI demand.
• Buy MU call spreads 9-15 months out rather than outright calls to express the view that pricing stays firm into 2027 while reducing premium burn if the stock pauses after a large run. Best risk/reward if implied vol remains elevated but not extreme.
• Consider a tactical long NVDA vs short an AI supply-chain laggard only if memory pricing stays tight for another quarter; MU strength is a read-through that AI capex is still broadening, but NVDA remains the cleaner core beneficiary. Use as a relative-value expression, not a directional one.
• Avoid chasing here on a cash-equity basis above analyst targets without a catalyst calendar. If management commentary turns cautious on customer digestion or capex sequencing, the stock can correct sharply despite strong long-term fundamentals.