The KOSPI's stunning AI-fueled rally is unraveling as investors question whether the semiconductor supercycle has overshot reality
South Korea’s benchmark KOSPI index plunged as much as 8.2% intraday on July 7, briefly crossing the threshold into bear market territory, defined as a 20% decline from recent highs. The culprit: a sweeping reassessment of artificial intelligence demand that hit the very stocks responsible for the index’s meteoric rise.
SK Hynix and Samsung Electronics, the twin pillars of South Korea’s semiconductor industry, led the selloff. Together, those two companies account for roughly half of the KOSPI’s entire market capitalization.
From supercycle to super selloff #
The KOSPI had surged more than 100% year-to-date earlier in 2026, riding a wave of global AI infrastructure spending that turned South Korean memory chip makers into the market’s main characters. The index peaked above 8,800 to 9,000.
Goldman Sachs was so bullish that it raised its 12-month KOSPI target to 9,000 back in May, forecasting a staggering 300% earnings growth for semiconductor companies on the back of the AI memory chip supercycle.
In June, the KOSPI dropped roughly 10% in a move sharp enough to trigger trading circuit breakers.
The AI demand question #
Concerns about potential oversupply in the memory chip market have started surfacing. There’s also the broader question of AI adoption rates. Enterprise customers have been aggressive in building out AI infrastructure, but the revenue models for many AI applications remain immature.
The KOSPI’s problem is uniquely acute because of its structure. When roughly half your index is two semiconductor stocks, any wobble in the AI narrative doesn’t just affect a sector. It moves the entire market.
What this means for investors #
The divergence between Goldman Sachs’ bullish 9,000 target and the market’s corrective behavior is worth watching closely.
Investors with exposure to AI-adjacent assets should be paying close attention to two variables going forward. First, memory chip supply dynamics: any signals of inventory buildup or slowing order growth from hyperscalers would validate the bearish case. Second, AI spending announcements from major cloud providers during the upcoming earnings season will either restore confidence or deepen the current skepticism.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our