South Korea Slipped With Chip Stocks While Looking Record Cheap
South Korean stocks fell after a chip-stock selloff spread to Seoul and revived concern the AI rally had gone too far, even as that rally left the market at record-low valuations.
The pullback centered on semiconductors. Selling that started in chip names elsewhere broadened to Korea and pushed stocks lower, refocusing attention on crowding and duration in AI-linked shares, according to Bloomberg. Crowding means many investors in the same trade; duration means waiting a long time for payoff. The move was framed not as a Korea-specific downgrade but as beta to a global AI trade losing steam. Beta here means moving with the pack.
That link is direct for Seoul. The KOSPI, Korea's main stock index, has heavy weight in memory and logic-chip exposure, so selling in AI winners feeds quickly into index performance and foreign flows.
Separately, and later, the same rally left Korean equities screening cheaper than ever. South Korea's rally left equities trading at record-low valuations, as reported July 11, according to Bloomberg. Prices rose while multiples stayed depressed. A multiple is price divided by earnings, so the market climbed and looked cheaper at once.
The broader context here is that contradiction of falling prices with cheap multiples. It points to the denominator, not just sentiment. The split to watch is multiple compression versus earnings revision. If forward profit estimates for memory and the export chain rose faster than prices during the rally, a strong run can still leave multiples at lows. A chip-led drop then lowers prices without settling whether those profit forecasts reflected short-term restocking, structural AI spending, or a mix.
Looking at what this means for positioning, the two facts pull opposite ways. The selloff signal is about concentration and correlation. When one theme leads, that shared exposure can overwhelm differences between stocks, and a country bet becomes a theme bet. The valuation signal is about hurdle rates and base effects. A low multiple alone does not set a catalyst, a holding period, or the range of outcomes for memory pricing, factory use, and discipline on new plants. It sets starting compensation for holding that cycle.
In my view, the question is sequencing. A market can screen cheap for structural reasons, a governance discount, a sustained high cost of equity, or persistent low profits against book value, then bounce on flows before profits confirm. If AI spending expectations reset, the earnings forecasts that made Korea look cheap can fall faster than prices settle. That is the trap where cheap gets cheaper until earnings visibility returns. The other path is that earnings hold, volatility clears crowded positions, and the low multiple becomes compensation rather than a value-trap marker. Price action alone will not settle it. Orders, utilization, and supply-chain guidance will.
