[Edaily Reporter Park Sun-Yeop ] Analysts have suggested that the recent sharp decline in the KOSPI is not so much due to the semiconductor industry having peaked and entered a downturn, but rather a price correction resulting from the simultaneous withdrawal of funds that had been excessively concentrated in semiconductor stocks, along with leveraged investments. The consensus is that for the KOSPI to rebound in earnest, earnings reports must confirm that major U.S. tech companies are continuing to invest in artificial intelligence (AI).
Kim Doo-eon, an analyst at Hana Securities, stated in a report on the 21st, “The profits of semiconductor companies have already been confirmed,” adding, “What the market now needs to confirm is orders from U.S. hyperscalers, the end users.” Hyperscalers refer to big tech companies such as Alphabet, Microsoft, and Amazon that operate large-scale data centers and purchase AI semiconductors and memory.

Kim explained that the recent decline stemmed from a combination of factors. After the KOSPI rose sharply, foreign investors moved to take profits and rebalance their portfolios; as stock prices fell, this was followed by the liquidation of funds tied to margin loans and leveraged products.
He assessed that single-stock leveraged exchange-traded funds (ETFs) acted less as the initial trigger of the decline and more as an amplifier that exacerbated the drop. Since leveraged ETFs must adjust their holdings daily to track twice the daily return of their underlying assets, this process placed additional selling pressure on the market during the downturn, as seen at SamsungElectronics(005930)and SK hynix(000660).
Corporate earnings themselves were not bad. SamsungElectronics’ second-quarter operating profit increased approximately 19-fold compared to the same period last year, and TSMC’s second-quarter profit also rose by 77%. However, the stock prices of both companies actually fell following the earnings announcements. This was because market attention had shifted from the earnings that had already been reported to how much AI investment would continue in the future.
The rise in memory prices was also met with mixed reactions in the market. While it represents a factor improving earnings for SamsungElectronics and SK hynix due to higher selling prices, it imposes a cost burden on Big Tech and AI infrastructure companies that need to purchase memory.
However, analysts noted that there is no need to interpret the fact that some AI companies are reviewing hedging strategies to reduce the risk associated with long-term memory supply contracts as a sign of slowing demand. They explained that this is more akin to financial risk management aimed at minimizing potential losses that could occur if future spot memory prices fall below the minimum prices stipulated in the contracts.
The recent weakness in semiconductor stocks on the U.S. stock market was also interpreted not as a sign that the AI investment boom has ended, but rather as a transition toward new market leaders. Investment capital is shifting from “AI suppliers”—such as semiconductor companies, which have already seen significant gains—to “AI demanders,” which are executing AI investments and generating revenue through services.
Accordingly, the earnings reports from U.S. Big Tech companies are expected to be the first test of the KOSPI’s rebound. The market is expected to scrutinize not only revenue and earnings per share but also whether AI service revenue is growing, whether data center capital expenditure plans are being maintained, and whether companies can sustain profitability and cash flow even amid large-scale investments.
If Big Tech companies maintain their AI investments, rising memory prices could once again be interpreted not as a cost burden but as evidence of strong demand. Concerns that semiconductor companies’ profits have peaked are also likely to subside.
However, the report emphasized that a market rebound should be distinguished from a full recovery to previous highs. Hana Securities presented three scenarios for the time it will take for the KOSPI to recover to its previous high. They estimated it would take 1,148 days if the current bull market ends and a trend similar to past crises emerges, and 150 days under the base scenario, which reflects the strong resilience of this bull market. They noted that if the situation amounts to nothing more than a temporary supply-demand adjustment within the bull market, the timeframe could be as short as 32 days.
Researcher Kim stated, “Once the supply-demand rebalancing is complete, a technical rebound could begin,” but added, “The driving force behind recovering the previous high will not be simple oversold conditions, but rather orders from hyperscalers.” He further emphasized, “Since the profits of semiconductor companies—the suppliers—have already been confirmed, it is now time to confirm the investment willingness of Big Tech—the demand side.”