Micron: HBM Leads To Profit Explosion

Micron reported Q2 revenue of $23.9B, up 196% Y/Y, with non-GAAP gross margin of 75% and free cash flow rising 705% Y/Y. Management guided Q3 to $33.5B in revenue and 81% gross margins, citing continued high-bandwidth memory (HBM) growth and robust DRAM demand in the Data Center market.

Analysis

The immediate market reaction understates where value is being created across the stack: advanced packaging, substrate, and test vendors will see multi-quarter revenue visibility and higher realized pricing before commodity DRAM recovers. That flow-through will push billings for OSATs and certain WFE names on a different cadence than traditional memory suppliers, creating outperformance pockets that aren’t correlated 1:1 with headline memory ASPs.

A key second-order dynamic is concentration risk on a small group of hyperscalers and accelerator vendors. When demand is highly concentrated, negotiating leverage and lead-times swing quickly — customers can accelerate orders in one quarter and pause in the next, producing large quarter-to-quarter revenue volatility for suppliers and straining working capital across the ecosystem.

On the supply side expect a two-speed cycle: capital allocation will favour advanced-stack HBM capacity and packaging for several years, while commodity DRAM capacity rationalization and price mean-reversion occur over 6–18 months. That implies durable orderbooks for certain equipment and OSAT players, but also sets up a margin reversion risk for winners once competitors expand capacity or governments subsidize new fabs.

Watchables for next 90–180 days: (1) OSAT/substrate book-to-bill, (2) WFE vendor backlog trends, and (3) hyperscaler procurement cadence commentary. These indicators will tell whether current margin levels are structural (multi-year) or a short-term overshoot vulnerable to a demand pause or faster competitor capacity additions.

Key Decisions for Investors

• Buy MU via a Jan-2027 call spread (long LEAP call / sell higher strike) sized 2–3% of portfolio. Rationale: asymmetric upside to capture multi-year HBM/AI memory secular demand while limiting premium outlay. Entry: within 2 weeks or on a >15% pullback; exit/roll if implied vol compresses >40% or position down 50% from peak (max loss = premium). Target: 2.5–4x premium in 12–18 months if share gains + margin durability persist.
• Long Amkor (AMKR) or ASE-equivalent exposure (stock or 9–12 month calls) sized 1–2% to play OSAT/substrate pricing power. Rationale: advanced packaging benefits before broad DRAM normalization. Stop: 20% below entry; target: 40–80% upside in 6–12 months as backlog converts to billings.
• Pair trade to de-risk macro memory cyclicality: Long MU (stock or LEAP) vs short SMH (semiconductor ETF) notional-neutral. Rationale: isolates company-specific HBM upside while hedging broad semiconductor downturn risk. Timeframe: 6–12 months; unwind if SMH outperforms MU by >15% or if OSAT book-to-bill deteriorates materially.