This tech company is set to overtake Nvidia as the world’s most profitable

KB Securities projects Samsung Electronics will post KRW488 trillion of operating profit in 2027, narrowly surpassing Nvidia’s KRW485 trillion, while SK hynix is forecast to rank third globally at KRW358 trillion. Combined corporate taxes from the two chipmakers are estimated to rise to KRW203 trillion in 2027, and KOSPI operating profit is seen reaching KRW1,044 trillion, supporting a 7,500 target for the index. The outlook implies strong upside for Korean equities and a potential stabilizing effect on KRW/USD via semiconductor-driven dollar inflows, though foreign investors recently sold KRW66 trillion of Korean assets in February and March.

Analysis

This is less a simple Korea bull case than a repricing of the global memory-DRAM/AI capex stack. If the earnings power materializes, the bigger trade is not just Korea beta; it is the levered beneficiaries of a prolonged AI infrastructure cycle with high exposure to HBM/advanced packaging, while the losers are businesses dependent on imported components and a stable won. The FX implication matters: persistent dollar inflows from semiconductor capex and profits can dampen KRW volatility, which in turn reduces the funding-risk premium embedded in Korean equities and local credit.

The market is still underestimating second-order beneficiaries outside semis. A stronger KOSPI and easing FX pressure typically flow into domestic financials, brokers, and capital-goods suppliers, but the more interesting read-through is to cyclicals tied to capex execution and grid/power buildout. If the index-level earnings expansion broadens beyond memory, the valuation gap versus peers is less about cheapness and more about governance/flow constraints closing at different speeds; that argues for staged re-rating rather than a one-shot rerating.

The main contrarian risk is that consensus may be overconfident on the duration and purity of the memory upcycle. History says supply response eventually compresses margins faster than top-line growth, and any stumble in AI capex or cloud inventory digestion would hit the high-margin Korean names hardest because expectations are so elevated. In that context, the near-term catalyst is mostly flow-driven—foreign underweight re-accumulation and KRW stabilization—while the medium-term risk is a negative surprise in HBM pricing or capex discipline over the next 2-4 quarters.

Key Decisions for Investors

• Go long EWY vs short EEM for a 3-6 month horizon: express a Korea-specific earnings rerating versus broader EM, with upside if foreign flows normalize; stop if KRW weakens sharply or semicap guidance rolls over.
• Buy KOSPI exposure on pullbacks via KOR or a Korea ETF basket, targeting 12-18 month upside from valuation expansion; best entry is after any foreign-flow-driven selloff, not at fresh highs.
• Long NVDA / short a basket of Korean memory-levered names only if you believe the market is overpricing the second derivative of memory profits; otherwise avoid outright short NVDA because it remains the cleaner AI capex hedge and the article is structurally supportive of the ecosystem.
• Long KRW hedged financials and brokers in Korea for 6-12 months: stronger local equity turnover and calmer FX should support domestic liquidity-sensitive names with lower earnings volatility than semis.
• Use call spreads on Korea exposure rather than outright equity longs if you want convexity: the upside case is a multi-quarter rerating, but the downside is sharp if memory pricing peaks early or foreign investors fade again.