Momentum was reinforced by strong capital inflows into emerging market equities, which had already begun outperforming developed markets in 2025 and carried that strength into 2026 . The MSCI Emerging Markets Index itself gained 24% in the first half of 2026, while South Korea's KOSPI surged 101.1%
. The FTSE Korea index returned roughly 188% over the preceding 12 months, making it the top-performing equity market globally
.
The sell-off that began in early June and intensified through late June was not a single event but a series of accumulating shocks .
Stage 1 — Broadcom disappointment and Fed repricing (June 5-8)
On June 5, Broadcom reported a revenue miss, which immediately weighed on the AI semiconductor theme . The same day, the U.S. Bureau of Labor Statistics reported that the economy had added 172,000 nonfarm jobs — more than double the roughly 80,000 economists had expected
. That stronger-than-expected jobs report pushed rate expectations higher and sparked a broad tech sell-off in the U.S. that spread to Asian markets the following Monday, June 8
. South Korea's KOSPI fell over 4.5% in a single session, led by semiconductor stocks
. Foreign investors pulled an estimated $27 billion out of emerging-market portfolios in May and early June
.
Stage 2 — AI-cost anxiety, forced deleveraging, circuit breakers (June 23-26)
By June 23, the sell-off intensified dramatically. South Korea's KOSPI fell 10.5%, led by Samsung and SK Hynix, wiping out months of gains . Korean regulators had previously warned about extreme levels of margin debt and retail borrowing in leveraged ETFs
. When the sell-off hit, forced deleveraging amplified the decline, triggering a cascade of liquidations
. On June 26, an MSCI gauge of emerging market equities fell as much as 3.9% in a single day, the biggest drop since June 8
. South Korea took the heaviest blow: Samsung Electronics and SK Hynix each slid more than 10% at one point, dragging the KOSPI down as much as 9% and triggering a 20-minute trading halt — the fifth trading halt the Korea Exchange had imposed in 2026
.
AI-cost anxiety hardened into a genuine risk-off move. Apple's price-increase announcements (driven by a memory chip shortage) stoked fears about consumer demand for premium AI devices, which in turn raised doubts about chip demand .
The rally and the crash shared a common driver: extreme concentration in a handful of AI semiconductor giants.
South Korea was the epicenter. Samsung Electronics and SK Hynix became both the engine of the KOSPI's record highs and the source of its near-10% plunge . The market became dangerously narrow: when foreign investors aggressively sold chip stocks, the entire index dropped sharply
.
Taiwan followed a similar pattern. TSMC (Taiwan Semiconductor Manufacturing Co.) was a core AI beneficiary, and the Taiwan market became similarly concentrated around the AI chip cycle, amplifying both upside and downside . JPMorgan strategists noted that after steep year-to-date returns, just three stocks — Samsung Electronics, TSMC, and SK Hynix — accounted for the bulk of the gains
.
Japan, while classified as a developed market, has a tech-heavy semiconductor supply chain (Tokyo Electron, Advantest, and other chip-equipment makers) that traded in lockstep with the AI theme and was swept up in both the rally and the rout .
The structural risk is clear: when multiple markets are tied to the same AI valuation cycle, a reversal in sentiment can hit them all at once .
As the sell-off unfolded, major institutional voices began issuing explicit warnings about the risks embedded in the AI trade.
BlackRock (June 30): The world's largest asset manager downgraded its view on emerging-market equities from "overweight" to "neutral" . BlackRock explicitly flagged rising AI concentration risk in Korean and Taiwanese stocks, warning that the extreme concentration in AI-tied corporations could shake the broader market
. This was a direct institutional acknowledgment that the very structure that drove the rally made the market vulnerable.
Bank of America has been cautioning about AI concentration since at least January 2026, when BofA described the heavy concentration of capital flowing into AI as creating "systemic vulnerabilities" and suggested "transition investing" as a hedge . By February, BofA's European equity strategy team issued a note declaring that "doubts around the AI revolution are emerging," with the market narrative shifting from an "upside-only" perspective to concerns that AI might actively destroy corporate profits
.
Earlier, in late 2025, BofA's Savita Subramanian argued that this was an "air pocket" rather than a full bubble — meaning a period of turbulence driven by rising data-center debt and potential returns disappointment — but warned investors to brace for volatility in 2026 . In May 2026, BofA analyst Michael Hartnett warned that the rising excitement surrounding AI stocks was driving U.S. stock markets toward levels of concentration "reminiscent of some of the most significant speculative bubbles in history"
.
Bank for International Settlements (June 30): The BIS issued a historically framed warning, stating that the scale and pace of the current AI investment boom "bear resemblance" to past investment booms that ended in busts . The BIS warned that if returns on AI disappoint, the spending surge could become "a protracted investment bust," with knock-on effects across the financial system
. The BIS also noted that the AI boom is unfolding through a highly concentrated ecosystem of hyperscalers, suppliers, and private lenders linked by debt and increasingly opaque financing arrangements
.
EM tech enters the second half of 2026 with a record-breaking first half behind it but also with extreme concentration, elevated valuations, forced deleveraging risk, and explicit warnings from BlackRock, BofA, and the BIS . The late-June sell-off demonstrated how quickly the hot trade can unwind when multiple shocks hit at once. The lack of market breadth means any further negative AI news could trigger another cascading decline
.
Investors will be watching three things closely: whether AI-related earnings (especially from chip companies) can justify the remaining valuations; whether regulators in Korea and elsewhere tighten margin-lending rules to prevent a repeat of the forced deleveraging cascade; and whether the market begins to broaden beyond the handful of AI-champion stocks that dominated the first half of the year .