Prediction: Micron's Sept. 30 Earnings Will Confirm the Memory Shortage Isn't Over

DRAM and NAND prices reached record levels in August, with PC DRAM rising more than 4% month over month to $25 and NAND increasing 1.4% to $30.10, reinforcing a persistent AI-driven memory shortage ahead of Micron's Sept. 30 earnings. Micron's fiscal Q3 DRAM revenue rose to $31.3B and NAND revenue to $9.9B, supported primarily by sequential ASP gains of roughly 60% for DRAM and 80% for NAND. TrendForce expects Q3 PC DRAM prices to rise another 18%-23% sequentially, while SK Hynix forecasts the DRAM market may not rebalance until 2030, supporting bullish expectations for Micron's forward guidance.

Analysis

MU is increasingly a pure pricing-duration trade rather than an AI-volume trade. Its comparatively lower HBM mix creates greater near-term operating leverage to broad DRAM and NAND spot/contract inflation, but also leaves the multiple vulnerable if investors rotate toward structurally differentiated HBM suppliers such as SK Hynix (000660 KS) and Samsung (005930 KS). The key earnings variable is not the reported quarter; it is whether management’s December-quarter and FY27 gross-margin framework implies that contract pricing remains above cash-cost expansion.

The second-order risk is that record pricing pulls forward customer procurement and stimulates capacity additions just as AI server demand normalizes. NAND is historically less disciplined than DRAM, so a modest ASP increase can sustain earnings, but a renewed supply response would compress MU’s earnings power disproportionately within 6-18 months; consensus low-P/E framing is therefore potentially a peak-cycle valuation trap. Conversely, persistent HBM capacity absorption can keep commodity DRAM wafer availability tight longer than the market expects, benefiting MU even without HBM leadership.

Near term, positioning appears favorable into the September 30 print only if estimates have not already absorbed another guide-up. A weak spot-price reaction after earnings, customer inventory commentary indicating normalization, or gross-margin guidance below consensus would be more informative than a headline beat and would invalidate a long. The more durable expression of the scarcity thesis is selective exposure to memory-capex beneficiaries—AMAT, LRCX and KLAC—once suppliers translate pricing power into fab and advanced-packaging spending budgets.

Key Decisions for Investors

• Tactically long MU into September 30 only via a defined-risk call spread, such as Oct/Nov 5-10% out-of-the-money calls financed by higher strikes; target 2:1 upside/downside. Exit on guidance that fails to raise the next-quarter revenue or gross-margin outlook, rather than waiting for spot-price data.
• Pair trade for 1-3 months: long MU / short SOXX in equal beta-adjusted dollars. This isolates memory ASP and margin-duration upside from broad AI-semiconductor multiple risk; stop if MU underperforms SOXX by 8% following earnings or if management cites inventory digestion.
• Do not chase SNDK solely on NAND pricing. Establish a watch alert for evidence of sequential margin expansion and sustained enterprise-SSD mix gains; absent that disclosure, NAND price appreciation may chiefly offset prior-cycle losses rather than generate durable free cash flow.
• For a 6-18 month horizon, accumulate AMAT, LRCX and KLAC on post-earnings semiconductor volatility rather than MU at cycle-peak enthusiasm. The thesis requires disclosed memory-fab/packaging capex increases; cancel if suppliers prioritize shareholder returns and existing-fab utilization over new capacity.