South Korean stocks suffered their sharpest drop in more than three months on Tuesday as fears that an AI-driven rally pushed semiconductor valuations to unsustainable levels sparked a market-wide sell-off.

The benchmark Korea Composite Stock Price Index (KOSPI) plunged 9.99%, shedding 910.71 points to close at 8,203.84.

The Korea Exchange (KRX) activated a market-wide circuit breaker at around 2:33 PM after the KOSPI fell more than 8% from the previous session’s close. Trading in all stocks was halted for 20 minutes, marking the fourth such suspension this year and the tenth on record.

The slide came just a day after the index closed at a record 9,114.55. Samsung Electronics and SK Hynix both fell more than 12%, making South Korea’s two AI-linked memory giants the main drag on the market.

Samsung and SK Hynix together accounted for around 53% of the KOSPI’s total market value, leaving the index heavily exposed to any reversal in semiconductor sentiment.

“The sell-off appears to be driven mainly by profit-taking after the recent sharp rally, as the market had become increasingly overbought,” Ha SeokKeun, the chief investment officer at Eugene Asset Management in Seoul, said, as cited by Bloomberg.

The rout followed explicit warnings from the Korea Exchange and local regulators about overheated tech valuations. The authorities said they were considering measures to curb speculative trading linked to Samsung Electronics and SK Hynix.

The pressure intensified after SK Hynix briefly overtook Samsung as South Korea’s most valuable listed company on Monday. SK Hynix’s market capitalization reached 2.080 quadrillion won (around $1.5 trillion), narrowly ahead of Samsung’s 2.067 quadrillion won, powered by demand for AI memory chips.

The rout also followed weakness in US technology shares. A broader tech sell-off spread into Asian markets, with investors increasingly questioning whether AI-linked valuations had run too far too fast, MarketWatch said.

The market turmoil comes as investors increasingly question whether the AI boom is creating a financial bubble. OpenAI, Oracle, and SoftBank have committed hundreds of billions of dollars to new AI infrastructure projects, while chipmakers and cloud providers continue to pour money into the sector despite limited evidence that profits are keeping pace with spending.

Some analysts argue that AI investment has become concentrated within a tightly linked network of companies whose valuations depend more on expectations and capital inflows than on proven profitability.