The prospect of military action by the Trump administration against Iran has triggered some of the most severe declines globally in North Asian stock markets. The underlying cause of this crash is not the direct economic impact of Middle East conflict on Asia, but rather the forced unwinding of excessively crowded AI semiconductor trades and historically high leveraged positions following an external shock.
Panic selling swept through Seoul and Tokyo on Wednesday. South Korea's benchmark KOSPI index fell more than 10% over two consecutive trading sessions, marking its largest two-day drop since 2008. This occurred after global capital had been rotating en masse from US software stocks into Asian semiconductor and hard-tech assets, with South Korea's margin debt and new trading accounts both reaching historic peaks.
Following the outbreak of conflict with Iran, the US dollar strengthened sharply, diminishing the appeal of emerging market assets. Concurrently, markets feared that sustained oil price shocks would drive inflation higher, forcing local central banks to raise interest rates and consequently increasing the cost of leveraged trades. Financial conditions in South Korea, which had been at their most accommodative level in decades, faced the risk of a sudden tightening, with highly leveraged long positions being the first to suffer. The global theme of diversification away from US assets, under the dual pressures of extreme crowding and capital flow reversal, devolved into indiscriminate selling of Asian assets.
Analysis suggests the driving force behind this sell-off is primarily capital flows rather than a deterioration in fundamentals. The long-term super-cycle narrative for memory chips from Samsung Electronics and SK Hynix, along with the AI capital expenditure trend confirmed by TSMC's robust earnings, have not seen a substantive reversal—the upward revisions to Asian corporate earnings currently remain stronger than those in the US.
Why did North Asia become the epicenter of a crash driven by capital flows? The relationship between steep market declines and capital flows is often far more significant than actual changes in fundamentals. Prior to the crash, as global investors rotated from software companies towards AI infrastructure plays, hot money was flooding into Asia seeking exposure to semiconductors and hard tech. This continued diffusion of the AI trade was the underlying foundation for this week's sharp declines in North Asian markets.
While North Asian economies are highly dependent on oil and gas imports, Europe potentially faces a more immediate energy crisis risk from the Middle East conflict in terms of direct energy supply shocks. Furthermore, barring a prolonged blockade of the Strait of Hormuz, North Asian economies possess some buffer capacity through national strategic reserves—Japan is estimated to hold roughly 254 days of oil reserves. This implies that the violent sell-off in North Asian markets is less a pricing-in of real economic impact and more a concentrated liquidation of leveraged positions.
Before the crisis erupted, the investment narrative for North Asian hard tech was overwhelmingly positive. Both Samsung Electronics and SK Hynix indicated that the supply tightness for memory chips would persist until 2027, with the market broadly expecting the companies to have entered a multi-year super-cycle; analyst earnings expectations for Samsung Electronics were consistently being revised upwards. Simultaneously, TSMC's strong earnings further confirmed that US hyperscale tech companies would continue increasing AI capital expenditure, sustaining the logic of Asian suppliers benefiting.
Hot money rapidly concentrated on a few winners. Data shows that in the week before Middle East tensions escalated, the $16 billion iShares MSCI South Korea ETF recorded over $1.2 billion in net inflows, the highest weekly figure in the fund's 25-year history. Concurrently, South Korean retail investors, breaking from a decades-long habit of avoiding blue-chip stocks, aggressively bought KOSPI constituent stocks, with the number of active accounts and margin debt both hitting record highs. By this point, the Asian AI infrastructure trade had become highly crowded—setting the stage for the subsequent stampede-like selling.
As the Iran conflict intensified, capital began to recede. The sharp strengthening of the US dollar undermined the investment case for emerging market assets. Markets also worried that persistent oil price shocks would fuel inflation, pressuring central banks to hike benchmark rates, thereby raising the cost of funding for leveraged trades. Data indicates South Korean financial conditions were previously at their most accommodative level in decades. Once funding costs began to tighten, long positions supported by margin debt faced forced liquidation pressure, accelerating the market's downward spiral.
A deeper issue is that this year's major theme of international diversification had funneled capital into North Asia with an intensity far exceeding normal levels. A geopolitical shock occurring thousands of miles away was sufficient to trigger a sharp reversal. The reallocation logic of "selling the US" thus transformed into indiscriminate selling of Asian assets.
This painful deleveraging process might be viewed as a healthy correction. It could flush out momentum-driven speculators who chase rallies and sell during downturns, allowing the market to revert to investors genuinely focused on corporate earnings and reasonable valuations. Regarding earnings fundamentals, the strength of upward revisions in Asia still surpasses that in the US; analyst expectations for Samsung Electronics' earnings continue to be revised upwards, with no substantive downgrade to its super-cycle narrative yet. Once the leverage is cleared, for capital truly focused on the semiconductor super-cycle and the long-term trend of AI infrastructure, the valuation levels after a thorough washout may provide a more solid foundation for re-entry.