The KOSPI Crashed 40%. Samsung Just Posted Record Profits.

In the last week of July, Samsung Electronics reported its highest quarterly operating profit on record — nearly 90 trillion won. SK Hynix posted 60.54 trillion won, a six-fold increase from the year before. Korean semiconductor exports surged 62.8% in July, with chip exports alone up 179% ^[1]. Over the same stretch, the KOSPI index lost roughly 40% of its value from its June peak, erasing $2.18 trillion in two days of trading ^[2]. If the chip sector was deteriorating, nobody told the chip sector. The question the numbers force is not whether Korea's semiconductor industry is in trouble — it is not — but what broke the market that contains it. The answer begins with what the KOSPI actually is. Samsung and SK Hynix together account for roughly half the index's market capitalization ^[3]. By June, that concentration had driven South Korea past India to become the world's sixth-largest stock market, with both chipmakers reaching trillion-dollar valuations and the KOSPI up more than 100% in a year ^[4]. The rally was almost entirely semiconductor-driven. On July 10, the index surged to 7,545.51 — SK Hynix jumped 6.65% on a US listing that was seven times oversubscribed, and Samsung rose 3.78% ^[5]. Nothing else moved the needle. The KOSPI had become, in effect, a leveraged bet on two memory-chip companies. That concentration alone would make the index volatile. What turned volatility into a rout was a layer of financial engineering sitting on top of it. Single-stock leveraged exchange-traded funds tied to Samsung and SK Hynix allowed retail investors to multiply their exposure to the same two names that already dominated the index. When global AI sentiment turned — on data-center project cancellations, China's CXMT memory-chip IPO, and internal signals that Nvidia's next-generation chip production forecasts were trending downward ^[6] — those leveraged positions began unwinding. On down days, forced retail liquidations cascaded into further selling, which triggered more liquidations. The Korea Exchange activated 37 sidecar halts in 2026 alone, temporary trading pauses designed to slow a freefall ^[7]. Finance Minister Koo Yun-cheol apologized for the role single-stock leveraged ETFs played in exacerbating the rout and said the Financial Services Commission is considering tighter leverage controls ^[2]. The foreign-investor flight compounded the pressure. Overseas investors sold roughly $110 billion in Korean equities through 2026, while the Bank of Korea raised its benchmark rate from 2.5% to 2.75% — its first hike in over three years — tightening into the sell-off ^[7]. On July 16, a single session now called Black Thursday, the KOSPI fell 6.37%: SK Hynix dropped 11.53%, Samsung 8.77%, as multiple shocks — a Morgan Stanley report of widespread data-center cancellations, China's $8.55 billion CXMT IPO, US military strikes on Iran, and the rate hike — hit a concentrated, leveraged market simultaneously ^[8]. Taiwan offers a useful comparison. The Taiex hit a record 47,018.99 on July 1, driven almost entirely by TSMC, which contributed 760 points to the index in a single session ^[9]. Taiwan's market is similarly concentrated in one semiconductor name. But the available record shows no comparable late-July collapse — no cascade of sidecar halts, no government apology over leveraged ETFs. The variable is not concentration alone. It is the financial wiring built on top of it. Morgan Stanley, in upgrading Korean stocks to overweight on August 2, attributed the rout explicitly to concentrated index weightings and single-stock leveraged ETFs ^[10]. President Lee Jae Myung acknowledged the problem directly.

Increasing concentration in the semiconductor sector has become a factor raising financial market volatility, with the impact of fluctuations in the chip sector on the whole stock market growing. — President of South Korea

Having named the risk, the government's response was to deepen it. On July 16, the same week the KOSPI suffered its worst single-day drop in years, the administration announced a $518 billion investment plan for Samsung and SK Hynix to build a new chip hub, targeting 80% of the global high-bandwidth memory market by 2028 ^[11]. A separate $150 billion shipbuilding alliance with the US, involving HD Hyundai Heavy, Samsung Heavy, and Hanwha Ocean, sits alongside it at less than a third the size ^[12]. Diversification is the smaller line item. The July rout was not the first time this architecture failed. In November 2025, the same pattern appeared: AI bubble fears, a 3.32% KOSPI drop, foreign investors pulling 1.05 trillion won, all concentrated in the same chip names ^[13]. The market's wiring produced the same outcome from the same trigger. Morgan Stanley's upgrade forecasts 36% upside to a KOSPI target of 9,000, on the thesis that the leverage washout has passed its midpoint ^[10]. The recovery thesis confirms the diagnosis rather than refuting it: it rests on the leverage cycle having run its course, not on any change to the underlying architecture. The wiring that turned a global AI correction into a 40% local rout remains in place. The government's $518 billion bet ensures the index will be even more concentrated in the two stocks that just broke it.