SK Hynix Is Buying Back 3.3% of Its Shares and Canceling Every One

SK Hynix approved a $29B (40T won) buyback to repurchase ~24.07M shares (~3.3% of outstanding) and cancel them, a move that sent Seoul shares up ~12% on Thursday. The company also plans to raise its 2025–2027 shareholder-return payout to over 50% of cumulative free cash flow (from within 50%), alongside cash dividends. This comes as Q2 strength delivered record revenue of 79.3T won (+257% YoY) and 76% operating margins, supported by AI memory demand and net cash of ~69T won.

Analysis

The market is reacting to the repurchase, but the real signal is that management is choosing to return capital aggressively while simultaneously funding another wave of capacity. That combination usually only shows up when a company believes its current margin structure is not just a peak quarter, but a multi-year cash-generation regime; if true, the equity should start trading less like a low-multiple cyclical and more like a cash-yield compounder with upside from operating leverage.

Second-order, this is bad news for peers with weaker balance sheets because it raises the bar for competing on both supply growth and shareholder returns at the same time. If HBM/DRAM tightness persists, the winners are the suppliers that can keep spending without diluting owners; if supply eventually loosens, the first-order casualty is pricing, but the lagged effect is multiple compression for the whole memory complex, especially names where earnings still look peak-ish.

Near term, the trade hinges on execution: the buyback cadence over the next quarter, the formal capital-return framework in the next earnings print, and any sign that memory ASPs are flattening. Over 6-18 months, the key falsifier is a rollover in AI memory demand or a capex bill that starts to outrun free cash flow, because then the “over 50%” pledge shrinks quickly and the stock reverts to a standard cyclical valuation. The contrarian view is that the move may still be under-owned because investors are treating this like a one-off buyback rather than a structural shift toward permanent share cancellation and higher per-share claims on a very cash-rich business.

Key Decisions for Investors

• Long SKHY / SKHYV on weakness over the next 1-2 weeks; use post-rally consolidation rather than chase the open. Target 10-15% upside into the next earnings/board update if the payout framework is formalized as expected; invalidate if HBM/DRAM pricing softens or the stock gives back the breakout level.
• Pair trade: long SKHY vs short MU over 1-3 months to isolate the strongest balance-sheet/capital-return story in memory. This expresses the view that SK Hynix can support both capex and per-share returns better than the cleaner, less cash-rich alternative if memory pricing normalizes.
• For lower-risk expression, use a 2-3 month call spread on SKHY/SKHYV rather than outright stock to capture the catalyst from buyback execution and Q3 framework disclosure while limiting downside if the market starts discounting the program as purely financial engineering.
• Watch NVDA as a secondary beneficiary, but only tactically: if HBM supply remains tight into the next quarter, NVDA retains pricing power and shipment visibility. If memory tightness begins to ease, reduce that exposure first, since the spillover works through supply availability before it shows up in earnings.
• Set an alert for the next earnings release: if management does not reaffirm the 'over 50%' return floor or if free cash flow guidance falls while capex stays elevated, treat the thesis as broken and fade the stock on any relief rally.