Micron reported fiscal Q2 revenue nearly tripled YoY to $23.9B and adjusted EPS of $12.20, well above Street EPS of $9.31. Management guided a record $33.5B revenue and $19.15 EPS for the current quarter (analyst expectation $10.77), citing AI data-center demand consuming >50% of memory shipments and supply constrained (Micron able to fulfill ~50–66% of key-customer demand). Industry notes (SK Hynix, Bloomberg) point to supply shortages through 2030 and HBM CAGR ~42% to 2033; Motley Fool projects upside to ~$119 EPS by 2030 and a theoretical price of ~$2,499 (21x) implying ~490% upside from current levels.
The current HBM-led shortage is not just a pricing shock for DRAM; it reorders capital allocation across the entire compute stack. Suppliers that can rigidly prioritize HBM (Micron, select SK Hynix lines) capture outsized pricing power while legacy PC/phone BOMs see supply and margin erosion, forcing OEMs to either absorb higher costs or reduce unit volumes. Because new fabs and HBM tooling have 18–36 month lead times, the market can remain structurally tight across multiple buying cycles even if spot ASPs soften briefly. Second-order winners include wafer fab equipment and specialty materials providers whose revenue is lumpy but high-visibility to multiyear AI capex plans; second-order losers are ODMs and mid-tier PC/smartphone OEMs that lack long-term supply contracts and must compete on price. Cloud and hyperscaler customers will pay a dual tax—higher memory costs and slower refresh cycles—so expect negotiated allocation agreements and multi-year supply contracts to become a dominant commercial structure, reducing spot volatility but concentrating revenue to a smaller set of suppliers. Key catalysts to watch are (1) capacity announcements with concrete tooling timelines, (2) HBM ASP trajectory vs PC/phone DRAM ASPs, and (3) signs of inventory digestion at hyperscalers. Tail risks that could reverse the trend include accelerated fab completions via government subsidy, a material jump in memory architecture efficiency (reducing GB/TB per model), or export-control shifts that fragment supply and temporarily inflate spot prices but undermine long-term addressable markets. Consensus scenarios appear to assume linear extrapolation of today’s tightness into 2030; that understates both the timing risk of supply coming online and the optionality embedded in customer contract structures. Position sizing should therefore be convex to the timing variable—larger exposure to outcomes where tightness persists through multiple build cycles, smaller exposure to a near-term mean reversion outcome.
AI-powered research, real-time alerts, and portfolio analytics for institutional investors.
Overall Sentiment
strongly positive
Sentiment Score
0.75
Ticker Sentiment