Micron capacity buildout: These equipment suppliers stand to benefit

Micron reported strong fiscal Q4 results and forecast a stronger first quarter, citing AI-driven memory demand and tight supply potentially lasting through 2028. More than 75% of Micron's FY2027 shipments are already covered by agreements, supporting a multi-year capacity-spending opportunity for Japanese and Korean semiconductor-equipment suppliers. Japan offers relatively cheaper large-cap names following recent pullbacks, while Korean suppliers carry higher momentum but elevated forward P/E multiples of 29x-80x; key risks include memory-price normalization in 2028, new Chinese DRAM capacity and factory startup costs.

Analysis

The investable read-through is not simply "long memory equipment": HBM-led output requires more process steps and tighter yield control, but incremental spend will arrive with a two-to-four-quarter order lag and will be concentrated at the handful of tools that solve memory-specific bottlenecks. Tokyo Electron (8035) and Screen (7735) offer cleaner upside to a formal DRAM/HBM capacity-expansion cycle, while Advantest (6857) is the more defensive expression because test demand tracks unit output and increasingly complex device qualification rather than greenfield fab announcements. SUMCO (3436) remains a poor proxy: higher-value HBM mix can increase bit output without a commensurate wafer-volume recovery, leaving its operating leverage dependent on broader commodity DRAM utilization.

The key near-term risk is that contracted supply is interpreted as evidence of new fabs when it may instead reflect allocation of existing leading-edge capacity and pricing power. Memory producers can sustain elevated margins by limiting supply, which is bullish MU and SK Hynix (000660 KS) but delays the equipment order cycle; startup-cost pressure also makes management teams more likely to phase capex. Over 6-18 months, Chinese DRAM additions are more threatening to commodity DRAM pricing than to leading-edge HBM, but a commodity price decline would still reduce cash-flow confidence and be the first trigger for tool-order deferrals.

Consensus appears too willing to pay for Korean equipment momentum while discounting the memory suppliers that retain the actual scarcity rents. The more attractive asymmetry is owning MU or 000660 KS on post-results volatility, funded against premium-multiple equipment names such as Hanmi Semiconductor (042700 KS) or Jusung Engineering (036930 KS), where even modest capex timing slippage can compress both estimates and multiples. This thesis is falsified by a material reduction in memory suppliers' 2027-28 capex plans, declining HBM pricing, or evidence that Chinese supply is qualifying into premium AI-memory channels.

Key Decisions for Investors

• Initiate a 3-6 month pair: long MU / short Hanmi Semiconductor (042700 KS), sized dollar-neutral. Target 15-20% relative performance with a 7-8% relative stop; MU captures scarcity-driven gross-margin upside, while Hanmi requires sustained incremental equipment orders to defend its premium valuation.
• For Japan exposure, accumulate Screen Holdings (7735) over the next 1-3 months rather than chase Korean momentum. Use a 10% downside stop and reassess if its next order/backlog commentary does not show memory-led improvement; it provides capex-cycle upside at a materially less demanding valuation than several Korean peers.
• Maintain Advantest (6857) as the lower-beta equipment watchlist leader, but do not add aggressively until customer utilization or test-order data confirm that AI-memory shipments are translating into production volume. Its relative resilience makes it vulnerable if the market rotates from execution certainty toward discounted capex beta.
• Set an alert for MU and 000660 KS guidance: any indication of capex acceleration without a corresponding deterioration in HBM pricing supports adding Tokyo Electron (8035) and 7735; a capex pause or commodity-DRAM price reversal favors increasing the MU/042700 pair rather than outright tool exposure.