SK Inc. and SK Square offer a way to own underlying businesses at a holding-company discount, assuming management acts in minority shareholders' interests. SK Square's 20% stake in SK hynix has benefited from portfolio cleanup, buybacks and executive-pay alignment; its shares outperformed SK hynix over the past three years. SK Inc. trades at a massive discount to its underlying SK hynix stake. x.com  x.com 

SK Square is a discounted way to own SK hynix, one of the world's most important memory companies. Its NAV is about ₩1.8 million per share, almost all of it the SK hynix stake, while the stock trades near ₩1.0 million, a 40-45% discount. Korean holding-company discounts never fully vanish, but management now targets a discount below 30% by 2028 and is returning capital via buybacks, cancellations or dividends. The case only needs SK hynix to keep compounding and the gap to narrow slightly. x.com 

Owning SK Hynix directly beats buying SK Square for its discount. Holding companies trade below NAV for a reason: there is no good way to force an arbitrage on their assets, just as Grayscale vehicles traded below the value of their crypto. SK Square's roughly 20% Hynix stake may look worth more than the company, but that discount can persist. x.com 

SK hynix should reprice strongly because its blowout earnings make the case that memory is structurally different now. At the current earnings pace, it would earn back its entire market cap in two years; buybacks and dividends also return cash to shareholders and add to value. x.com 

SK hynix stands to benefit from higher HBM prices as supply remains constrained. Needham expects HBM price increases next year and tight memory supply through 2030; SK hynix has finalized 2027 HBM prices and volumes with major customers. SK hynix also plans roughly KRW100 trillion in FY2026 shareholder returns, including at least KRW30 trillion in dividends and KRW60 trillion in share acquisitions and cancellations. x.com  x.com 

SK Square's NAV discount is overstated because the company would owe tax on its holdings. For US investors, the tax reporting implications also make it not worth owning. x.com  x.com 

SK hynix's earnings and buybacks could drive a tripling by 2027 before a re-rating, on Jonah Lupton's assumptions. Its Q3 report expects memory demand to outpace supply through 2030 as advanced processes and fab lead times constrain growth. It plans about KRW 100 trillion in FY2026 shareholder returns, including KRW 60 trillion or more in share acquisitions or cancellations; the buybacks would first reduce shares outstanding. x.com  x.com  x.com  x.com  x.com  x.com 

Sold-out HBM and cash returns make SK hynix look too cheap for a memory cycle that has not ended. Its HBM capacity is sold through 2027, 12-layer HBM4 is shipping, and next year's HBM prices are forecast to rise 121%. A completed ₩40T buyback adds shareholder returns; a special dividend funded by a Kioxia stake sale is possible. At roughly 4x earnings by one estimate, the shares could go higher if the AI memory cycle persists. x.com 

SK hynix's expected widening supply-demand gap makes the 2028 downcycle implied by its valuation seem impossible. SK hynix management says its shares are undervalued relative to earnings and cash flow, and that assessment rings true. x.com  x.com 

Samsung's HBM4 ramp makes SK hynix's ability to defend its HBM lead the key question in the memory race. Chosun Biz reports SK hynix's HBM4 shipments may have missed expectations for a second straight quarter amid Nvidia Rubin delays and chip tuning. Counterpoint puts Samsung's Q2 HBM share at 33%, up from 21%; TrendForce expects strong SK hynix Q3 profits. Samsung is the preferred memory stock right now, though there may be no single clear winner. SK hynix's October 28 report is the next check. … Show more