Analyst sets Micron stock price

Melius Research initiated Micron with a Buy and a Street-high $700 price target, implying meaningful upside from the current $520 share price. The call is driven by expectations that the AI memory cycle will extend through the end of the decade, supported by strong HBM demand, with all 2026 output already sold and HBM4 ramping in April 2026. Shares rose 4.7% on the day, and the new target lifted the Street average to $549.23.

Analysis

The key second-order winner is still NVDA, not MU: if HBM is effectively pre-sold through 2026, the bottleneck shifts from demand discovery to execution discipline. That reduces near-term supply shock risk for AI accelerators broadly, but it also means memory vendors can sustain pricing power longer than the market expects, which should keep GPU gross margin expansion intact even if model training spend normalizes. The competitive implication is that smaller memory players without locked capacity or advanced packaging integration will struggle to participate, widening the gap between “AI leverage” winners and commodity DRAM names.

The market is likely underestimating how much of this is a multi-year capex and supply-chain story rather than a one-quarter earnings story. If HBM4 ramps successfully, the next leg is not just higher unit ASPs; it is an ecosystem repricing of substrate, advanced packaging, and test/equipment vendors that are capacity-constrained and can re-rate before Micron’s own peak earnings. The tradeable consequence is that the best risk-adjusted exposure may sit one layer down the stack where expectations are lower and supply is still tight.

The contrarian risk is that consensus is extrapolating a straight line from AI demand into the late decade while ignoring two reversal vectors: export-policy backlash and customer concentration. If China access tightens meaningfully, the upside is partially offset by market-share loss in a still-material end market, and if hyperscaler capex pauses for even one or two quarters, HBM sentiment can de-rate quickly because the tape has already priced in scarcity. The move is not obviously overdone on fundamentals, but it is increasingly crowded on narrative, so timing matters more than direction.

Key Decisions for Investors

• Stay long NVDA vs. short a basket of legacy memory names for 3-6 months; the cleaner thesis is that AI accelerator demand stays intact while commoditized memory remains vulnerable to any capex pause.
• Initiate a smaller long in semi-cap equipment / advanced packaging beneficiaries over 6-12 months; the best asymmetry is in the picks-and-shovels layer where HBM capacity expansion forces continued tool demand.
• For MU, prefer buying on a 5-8% pullback rather than chasing strength; upside remains intact, but near-term risk/reward worsens as the stock prices in a near-flawless multi-year scarcity story.
• Use MU call spreads instead of outright stock for a 3-6 month expression; this captures continued re-rating while limiting downside if export-control headlines or hyperscaler digestion hits sentiment.
• Watch for any evidence of HBM order normalization in the next 1-2 quarters; if lead times shorten, reduce exposure quickly because scarcity premiums can compress faster than fundamental estimates.