SK Hynix rises as $28 billion US ADR sale draws strong investor demand: Bloomberg

SK Hynix’s planned $28B U.S. ADR listing has drawn demand several times larger than the shares on offer, signaling strong appetite for the AI memory-chip leader. The 177.9M ADRs (each ADR = one-tenth of a share) are set to price Thursday and begin trading on Nasdaq Friday, potentially aided by conversion restrictions that may limit arbitrage and allow a premium. Despite the fundraising optimism, SK Hynix shares were down nearly 3% in Seoul, while the broader KOSPI fell over 5%.

Analysis

The immediate winner is not the issuer so much as the bookrunner stack: a mega-foreign listing signals the ECM window is open, but the fee pool is still a one-off and likely too diluted across syndicate members to change quarterly EPS for BAC/C/GS/JPM. The real value is franchise signaling — if this clears with multiple-times cover, tech allocators are still willing to fund AI capex stories, which can spill over into semis, hardware suppliers, and future follow-ons. That tends to support relative multiples for GS and JPM more than the broader bank basket because league-table share and tech distribution matter more than balance-sheet lending here.

The contrarian miss is that headline demand does not equal durable secondary performance: the ADR conversion limits create scarcity that can inflate the U.S. print versus Seoul, but that premium can fade quickly once arb desks map the mechanics. Over 1-3 months, the key catalyst is whether the Nasdaq line holds a premium and whether this becomes a template for additional Asian tech issuance; over 6-18 months, it is a structural positive for ECM franchises, not for net interest income. Falsifiers: weak first-week aftermarket, a discount versus Korea, or a shift in AI memory pricing/order growth would turn this into a liquidity event rather than a durable sentiment catalyst.

Key Decisions for Investors

• Long GS/JPM vs short XLF for 1-3 weeks: express the view that ECM franchise share, not generic bank beta, captures the most incremental value from the listing. Target a modest relative outperformance; stop if the deal prices with only token oversubscription or the first trading sessions are flat-to-down.
• Buy SMH or SOXX on any post-pricing tech pullback and hold 1-3 months: the message to allocators is that AI memory capex remains fundable, which should support semi multiples even if the issuer itself is a Korean name. Falsify on weak U.S. trading premium or evidence of HBM order deceleration.
• Do not chase BAC/C standalone unless you want pure index exposure: their upside here is mostly beta, while incremental economics are likely smaller than the market implies. Prefer them only as lagging legs in a bank basket if the sector rips on risk-on flows.