Should You Buy Micron Before Sept. 30? 1 Number Could Send the Stock Soaring

Micron reports fiscal Q4 earnings on Sept. 30, with investor focus centered on whether AI-driven HBM memory demand and pricing strength extend into 2028. HBM3E and HBM4 are already fully booked through 2027, while five-year take-or-pay customer agreements cover roughly 20% of DRAM and 33% of NAND volume, reducing cyclical downside. Micron is also investing $100B in a New York fabrication facility expected to begin production in 2029 or 2030; favorable 2028 supply-demand commentary could materially support the shares after earnings.

Analysis

MU’s event risk is principally a duration trade, not a quarterly-estimate trade: a credible extension of tight HBM economics into 2028 would move the market from valuing a 2026-27 earnings spike toward assigning a higher mid-cycle multiple. The key distinction is whether incremental supply is contractually absorbed at returns above the cost of new wafer capacity, rather than merely whether demand indications remain strong. Commentary separating HBM from conventional DRAM/NAND is essential; broad-memory pricing can soften even while HBM remains constrained, limiting the consolidated-margin upside implied by bullish narratives.

Strategic contracts reduce downside volatility only to the extent that counterparties remain economically incentivized to take volume and that pricing bands preserve returns during a downturn. They also create a less obvious risk: suppliers that lock capacity too early can forgo upside if HBM pricing remains structurally tight, while customers may accelerate qualification of Samsung Electronics (005930 KS) and SK Hynix (000660 KS) to reduce dependence. The second-order beneficiary of sustained HBM tightness is NVDA, whose system demand is less likely to be constrained by memory availability; the loser is AI infrastructure buyers whose deployment schedules are delayed by component allocation rather than compute demand.

Near term, implied volatility and an elevated pre-event positioning premium make outright MU equity less attractive than defined-risk structures. Over 1-3 months, evidence of 2028 contracted volume, customer prepayments, or disciplined industry capex would support estimates and multiple expansion; vague demand language without allocation or pricing detail is a likely sell-the-news outcome. Over 6-18 months, the central risk is synchronized capacity expansion across the memory oligopoly, which historically converts supply discipline into price erosion faster than consensus models anticipate.

Key Decisions for Investors

• Use a post-earnings entry for a long MU / short SOXX pair rather than chase pre-report beta: initiate only if MU sells off more than 8-10% despite confirmation of 2028 contracted HBM demand and no material gross-margin-guide reset. Target 15-20% relative outperformance over 3-6 months; exit if management signals meaningful uncommitted HBM capacity for 2028 or industry supply growth above demand.
• For event exposure, buy an MU call spread 3-6 months out, financed with an out-of-the-money put spread, sized to a maximum defined loss. The thesis requires disclosed 2028 allocation/pricing visibility; absent that disclosure, avoid paying elevated front-week implied volatility.
• Maintain NVDA as the cleaner downstream confirmation trade if HBM availability improves without pricing collapse: long NVDA versus short an AI-capex customer basket only after management identifies supply relief that can accelerate system shipments. Falsifier: evidence that memory cost inflation is compressing accelerator-system demand rather than relieving unit constraints.
• Watch Samsung Electronics (005930 KS) and SK Hynix (000660 KS) relative performance following the call. If MU frames supply as broadly tight but the Korean peers underperform materially, treat that divergence as a warning that MU-specific qualification or execution issues—not industry duration—are driving the narrative.