Micron: Why The Quarter Made Me Change My Mind

Micron reported record Q3 results, with revenue up 345.8% Y/Y and non-GAAP EPS of $25.11, far exceeding guidance. New Strategic Customer Agreements (SCAs) now lock in 20% of DRAM and 33% of NAND capacity, totaling $100B in performance obligations and lifting the floor for future downcycles. Management/analyst commentary flags margin mean reversion as supply arrives by 2028, though SCAs provide partial downside protection and should reduce (not remove) cyclicality.

Analysis

The market should treat this as more than a one-quarter blowout: the strategic contracts convert part of Micron’s earnings base from spot-cycle exposure into a semi-annuitized cash flow stream. That matters because equity holders usually pay for volatility reduction only after a cycle proves it can persist; if management can show contracted demand covers a meaningful chunk of incremental capacity, MU deserves a higher trough multiple and potentially lower credit spreads.

The second-order loser is any memory name still heavily dependent on discretionary buying and spot clearing prices, especially less-capitalized NAND/DRAM competitors that cannot lock in demand as efficiently. The bigger competitive effect is on hyperscale customers: they gain supply assurance, but the contract structure reduces their ability to force price concessions in future negotiations, which should support industry pricing discipline even if end-demand normalizes.

The real risk is timing mismatch: the equity can re-rate now while the supply response is still deferred, but by 2028 the industry could be back in an oversupply regime if capex and node transitions outrun AI memory demand. Near term, the catalyst is estimate revision and multiple expansion over the next 1-3 months; the falsifier is any indication that DRAM/NAND pricing momentum stalls, customer commitments get re-cut, or rivals accelerate capacity faster than expected. The contrarian miss is that this is not peak-cycle euphoria if the contracted backlog is genuinely sticky, but it is also not a permanent de-cycling of the franchise.

Key Decisions for Investors

• Add MU on post-earnings consolidation rather than chasing the gap; thesis is 1-3 months of estimate revisions and multiple expansion, with downside if pricing commentary softens or guidance stops moving up.
• Pair long MU / short SOXX for a relative-value trade over the next 1-3 months: MU has a unique contract-backed earnings floor, while the ETF will still price the broader semiconductor tape and can lag if semis de-rate.
• Long MU / short WDC as a quality and balance-sheet spread: if memory pricing stays firm, MU’s contracted mix should command a premium; if the cycle turns, WDC’s weaker mix and leverage should underperform first.
• If implied vol is still elevated, consider a 6-9 month MU call spread instead of outright stock for cleaner upside to a rerating, with risk capped if the market decides this is still just peak-cycle earnings.
• Set a hard alert if management raises capex or competitors announce faster-than-expected DRAM/HBM expansion; that is the signal the 2028 oversupply bear case is being pulled forward.