1. Investment Frame
SK Hynix is increasingly treated as a strategic AI infrastructure supplier rather than a
pure commodity-memory proxy. The valuation question is whether HBM-driven profitability can
stay elevated long enough to change the market's through-cycle earnings baseline.
2. Company Profile
The company runs a broad memory portfolio with meaningful scale in DRAM, expanding
leadership in HBM, and participation in NAND. Its investment profile combines high
cyclicality with periods of sharp margin expansion when supply discipline and product mix
align.
- Core geography and customer mix are highly tied to global datacenter demand.
- Advanced memory roadmaps are increasingly linked to accelerator release cycles.
- Capital allocation discipline is central in downcycle preservation.
3. DRAM Base Business
DRAM remains the earnings foundation and the primary source of cycle beta. Pricing behavior
can shift quickly with utilization decisions across a concentrated supplier group.
This wiki treats DRAM assumptions as the anchor for downside testing because most valuation
drawdowns historically begin with faster-than-expected DRAM normalization.
4. HBM Leadership
HBM leadership is currently the key differentiator. However, market share headlines must be
matched with yield stability, qualification continuity, and packaging ecosystem throughput
to convert into sustainable earnings quality.
- Track shipment conversion, not only design-win announcements.
- Track customer concentration risk in high-end AI demand.
- Track premium durability as industry capacity expands.
5. NAND Role
NAND is an important but lower-conviction contributor in the current thesis. It can help
stabilize portfolio breadth but usually has lower margin visibility than HBM/DRAM in this
cycle phase.
6. Profitability Structure
Consolidated margin depends on the interaction of DRAM cycle behavior and HBM mix expansion.
When both move favorably, earnings can outpace revenue growth meaningfully; when either
weakens, downside is amplified by fixed-cost intensity.
7. Balance Sheet and Capex
Capex timing is critical in memory. The core objective is to defend technology leadership
without overbuilding into demand uncertainty. Balance-sheet resilience determines strategic
flexibility when cycles turn.
8. Scenario Grid
| Scenario |
Assumption Set |
Earnings Profile |
Valuation Read-Through |
| Bull |
HBM premium persists and DRAM supply stays disciplined. |
Elevated margins across multiple years. |
Higher justified multiple and stronger downside buffer. |
| Base |
HBM remains strong but DRAM normalizes gradually. |
Healthy but declining peak profitability. |
Moderate multiple support around mid-cycle earnings. |
| Bear |
Rapid supply response and weaker demand conversion. |
Faster margin reset. |
Sharp EPS and multiple compression. |
9. Monitoring Dashboard
- HBM shipment quality, yield commentary, and customer pipeline concentration.
- DRAM ASP trend relative to inventory conditions across cloud/mobile/PC.
- Capex guidance changes and implied supply additions.
- KRW earnings translation sensitivity to currency and export policy shifts.
10. Risks and Method
Main risks include quicker DRAM rebalancing, HBM execution bottlenecks, and cyclical
overexpansion of capacity. This wiki is built as a long-form context layer for the model
appendix and should be refreshed with each major company update.
Document intent: independent research context for model users. Not
investment advice.