Circuit-breakers fired on both of South Korea's main stock exchanges for two consecutive days — an unprecedented event in the country's market history. The KOSPI, South Korea's main equity benchmark, fell 5.98% on the 29th to close at 5,663.24, after having already plunged 10.84% the previous day to 6,023.66. Over 25 trading days, the cumulative decline has reached 33.9% — the second-deepest on record, narrowly behind the 34.7% collapse during the covid-19 pandemic in March 2020, and steeper than the drops seen during the 1997 Asian financial crisis, the 2008 global financial crisis, and every other major market shock since.
Three forces are driving the rout: growing scepticism about artificial-intelligence infrastructure spending and the semiconductor cycle; forced selling triggered by leveraged financial products; and the market's pre-emptive pricing of a possible surprise interest-rate rise by the US Federal Reserve. Most analysts believe the sell-off has overshot the underlying deterioration in fundamentals, and the consensus advice is to avoid panic-selling, shun new leverage, and accumulate positions gradually rather than all at once.
What happened
The KOSPI opened higher on the 29th before reversing sharply, shedding 5.98% to close at 5,663.24 and touching an intraday low of 5,262.77. This followed Tuesday's collapse of 732.09 points, or 10.84%, to 6,023.66 — the second-largest single-day fall in the index's history, with the intraday decline reaching 11.29%. The junior KOSDAQ market fell 7.72% on the same day, briefly breaching the 700-point level. South Korea's volatility index, the VKOSPI — sometimes called the "Korean fear gauge" — surged 7.58% on Tuesday to close at 83.43, after briefly clearing 90 during the session.
The back-to-back triggering of both sell-side circuit-breakers and trading halts across the KOSPI and KOSDAQ over two consecutive days has no precedent. Measured over 25 trading days, the 33.9% decline trails only the pandemic sell-off of March 2020 (34.7%) and comfortably exceeds the falls recorded during the 2011 US sovereign-credit-rating downgrade (-20.3%), this year's Middle East conflict shock (-14.7%), the 2022 global monetary-tightening episode (-14.1%), and the 2018 US-China trade war (-13.3%).
Cause one: scepticism about AI and semiconductors
The epicentre of the crash is the chip sector. Doubts about the sustainability of Big Tech's capital expenditure on AI infrastructure have been multiplying; markets have begun punishing even strong earnings. Alphabet fell roughly 7% on July 22nd despite reporting solid second-quarter results. The mood darkened further when ChangXin Memory Technologies (CXMT), a Chinese memory-chip manufacturer, completed its stock-market listing. The day after the listing, Samsung Electronics and SK Hynix fell approximately 13% and 14% respectively, in a move quickly labelled a "China semiconductor shock".
Whether that reaction was proportionate to the actual threat is, however, open to question. The prevailing expert view is that CXMT does not yet pose a decisive competitive challenge. Its high-bandwidth memory (HBM) technology lags South Korea's leaders by roughly five years; its mainstream DRAM process technology by about three. Its indigenously developed lithography equipment also falls short of the frontier. Crucially, CXMT's existence and ambitions were not news: the risks it poses have been discussed in the industry for years, so its listing does not fundamentally change the competitive landscape overnight.
That said, the risk cannot be dismissed entirely. The 12tn–14tn won (approximately $9bn–$10bn) raised through the listing could be deployed to expand commodity DRAM capacity, where Samsung has significant exposure. A resulting wave of low-cost supply could accelerate a price-decline cycle and compress margins. The sharp sell-off in Korean chipmakers, then, reflects less CXMT's current capabilities than an anxious market pricing in the possibility of a future price war, and doing so with considerable excess.
SK Hynix published record quarterly earnings on the 29th, yet fell short of consensus estimates. On its results call, management directly challenged market scepticism about the durability of AI infrastructure investment. Separately, concerns about vendor financing and circular funding arrangements around Nvidia weighed on sentiment: Nvidia's shares fell almost 5% in New York, and its credit-default-swap spreads widened.
Cause two: forced selling from leveraged products
The second driver is technical rather than fundamental. Single-stock leveraged exchange-traded products linked to Samsung Electronics and SK Hynix attracted large retail inflows, creating what market participants describe as a "tail-wagging-the-dog" effect, in which derivatives flows distort the underlying equities. Analysts estimate that a substantial portion of the KOSPI's decline from the 9,000-point range to the 6,000s can be attributed to leverage-induced amplification. As prices fell, margin calls and forced liquidations on credit-financed positions and leveraged products triggered further selling — a self-reinforcing spiral.
The authorities have taken note. South Korea's Financial Services Commission said on Tuesday that if conditions did not stabilise it would consider imposing caps on leveraged-investment exposure.
Cause three: a surprise Fed rate rise, pre-emptively priced in
The third factor is a risk that markets had largely ignored until recently. The US Federal Open Market Committee met on July 28th–29th (local time) to decide on interest rates. A hold had been the base-case assumption, but renewed inflation fears — stoked by another leg up in oil prices following fresh Middle East tensions — shifted that calculation sharply. According to CME's FedWatch tool, the implied probability of a rate rise at this meeting jumped from 13% to 38% in the space of a week. A hold remains the consensus outcome, but Bank of America described the meeting as a "live event", and some economists put the odds close to even.
Domestic monetary conditions are also tightening. The Bank of Korea raised its policy rate to 2.75% two weeks ago, reaffirming a hawkish stance, and some brokerages expect a further increase to 3.00% before year-end. The mere possibility that both domestic and foreign monetary policy could tighten simultaneously appears to have been priced into equities before any decision was announced. The FOMC result was due to be released in the early hours of July 30th, Korean time.
What kind of moment is this?
First, the consensus among brokerages is that there is no clear evidence of fundamental economic damage — at least not yet. KB Securities described the correction as excessive relative to underlying fundamentals. Heungkuk Securities argued that the market was simultaneously pricing in the worst-case scenarios on both AI investment sustainability and Chinese semiconductor competition. As noted above, CXMT's technological capabilities do not yet justify the alarm. The more accurate assessment is probably not that "fundamentals are fine" but rather that "damage exists but is not catastrophic — though it should not be ignored either". Kiwoom Securities observed that the cumulative sell-off has eroded market resilience to the point where investors are psychologically primed to seek out bad news with each successive decline.
Second, the scale of the drop is genuinely severe by any historical measure. At 33.9% over 25 trading days, this is the second-deepest drawdown on record, approaching the pandemic lows. Some market commentators are labelling it a systemic-risk event on a par with the dotcom bust or the post-2010 crises. However, historical analogies must be handled carefully. The Asian financial crisis, the 2008 meltdown, and the pandemic all involved direct damage to the real economy or the financial system itself. The current episode appears to be primarily a liquidity shock amplified by leveraged products. During the pandemic, for instance, margin calls on equity-linked securities (ELS) ran to 1tn won per day, and the Bank of Korea launched unlimited repurchase-agreement operations — a tool it had not deployed even during the 1997 crisis — as the entire financial system wobbled. There is no comparable evidence of systemic stress today.
Third, fear has clearly reached extreme levels. The VKOSPI had already hit 97.99 on June 29th — its highest reading since the global financial crisis — and surged back above 90 intraday during this week's falls. Simultaneous large-scale net selling by both retail investors and foreign institutions has taken on the character of a classic panic.
How should investors respond?
The advice converging from across the market can be distilled into a few principles.
Selling at current levels offers limited benefit. After a decline of this magnitude, further liquidation amounts to a bet on additional downside. Yet aggressive buying is equally hard to justify right now: the Fed's decision remains unknown, and there is no clear signal that leverage-driven selling pressure has been fully exhausted.
The textbook approach is to wait for confirmation of a turn rather than trying to pick the exact bottom. Identifying a change of direction and acting on it, rather than guessing the lowest point, is the more defensible risk-management strategy.
At the same time, history repeatedly shows that extreme oversold conditions and peak fear have provided investment opportunities. The fact that many analysts are characterising this as a liquidity shock rather than a fundamental collapse, and that the VKOSPI is near levels associated with previous market troughs, is — paradoxically — an argument for selective buying. The practical implication is to accumulate gradually across several tranches rather than deploying capital in a single move. Spreading entry points reduces the cost of being early if prices fall further, while limiting the opportunity cost if a recovery comes sooner than expected.
The clearest lesson from this episode concerns leverage. Credit-financed positions, margin lending, and leveraged structured products amplify gains in bull markets, but, as this week has demonstrated, they also amplify declines by triggering forced selling that feeds on itself. In a market this volatile, avoiding fresh leverage is the single most consistently repeated piece of advice.
What to watch: the FOMC decision, due in the early hours of July 30th Korean time, and the details of any government measures to restrict leveraged investment. Both will be near-term pivots determining whether this sell-off stabilises here or extends further.
Comparison of major market shocks
**Current crash (July 2026)** | **Covid-19 pandemic (March 2020)** | **Asian financial crisis (1997)**
25-trading-day decline | -33.9% (2nd worst on record) | -34.7% (worst on record) | No comparable data (monthly: -27%)
Nature | Leverage-driven liquidity shock + semiconductor competition fears (partly exaggerated) | Real-economy shutdown + financial margin-call crisis | National foreign-exchange and financial-system crisis
Fundamental damage | No evidence of critical damage; some risk factors warrant monitoring | Real economic activity effectively paralysed | Actual national near-default
Policy response | Considering caps on leveraged-investment exposure | Unlimited repo operations (unprecedented even in 1997); 10.8tn won "Together Korea Fund" established | IMF bail-out