[비즈한국] One of the triggers for last week’s semiconductor stock adjustment was news related to Meta. As it was announced that Meta plans to sell its excess computing resources from data centers and AI semiconductors—which it poured into developing its own AI models—to external parties, the market interpreted this as a signal that "AI overinvestment may have already taken place." In the aftermath, the S&P 500 semiconductor sector fell 6%, and the KOSPI semiconductor sector dropped 9% last week.
So, has the investment peak truly passed? According to Hana Securities, the capital expenditure (CAPEX) growth rate for the five major U.S. hyperscalers—Microsoft, Amazon, Alphabet, Meta, and Oracle—is expected to rise from 81% in the first quarter of this year to 90% in the third quarter. The point at which the investment growth rate begins to fall below the revenue growth rate—the moment when a genuine debate over passing the peak will emerge—is expected to be the third quarter of next year.

Free cash flow (FCF) already began to decline after peaking at $51.8 billion in the third quarter of last year, shrinking to $19.1 billion in the first quarter of this year, with a transition to a deficit expected in the third and fourth quarters. This means the inflection points the market is worried about are still six months to a year away. Lee Jae-man, a researcher at Hana Securities, diagnosed that it is "too early to reflect" concerns about an investment peak and capital recovery at this point. Furthermore, he noted that the KOSPI's adjustment was excessive, citing that the 12-month projected operating profits for Samsung Electronics and SK Hynix are rising both year-on-year and month-on-month.
In terms of numbers, the scale of the adjustment is already nearing the worst levels in history. Since 2023, the deepest decline the KOSPI experienced from its peak was 20%. Assuming a 20% drop from the recent peak of 9115 points, the figure would be 7290 points; the intraday low on the 3rd was 7378 points. Even using the most severe adjustment experienced in the last three years as a benchmark, the calculation shows less than 100 points left until the bottom. Indicators measuring how far the stock price has deviated from its one-month average price tell the same story.
Last week, the KOSPI fell to a level nearly 9% below its 20-day moving average (disparity ratio of 91.3%), which is the most oversold state since last year. Stock prices tend to return to their average the further they stray from it; the 9240-point target suggested by Hana Securities is a value calculated based on the point of recovery, assuming the oversold level returns to the post-2025 average (103.3%). It is less of a target index promised by a brokerage and more of a technical calculation of how far a compressed spring might bounce back.
Reports from two brokerage firms released on the same day interpreted interest rates differently, despite looking at the same U.S. employment data. Hana Securities focused on the bond market's reaction. As June employment figures came in weaker than expected, U.S. 2-year Treasury yields, which are short-term rates, fell faster than 10-year yields. As a result, the gap between short- and long-term interest rates widened from 29bp in June to 35bp. Hana Securities viewed this as a signal that concerns over interest rate hikes have slightly diminished. Historically, growth stocks have been relatively strong in such trends. Since 2015, in similar phases, the monthly average return of the S&P 500 growth index was highest at 1.7%, with sectors like semiconductors, hardware, and pharmaceuticals/biotech in the U.S., and semiconductors, pharmaceuticals/biotech, and defense in Korea performing well.
Conversely, BNK Securities saw different aspects within the same employment data. While the number of non-farm payrolls in June increased by only 57,000, the unemployment rate actually fell to 4.2%. The interpretation is that this was largely due to a decrease in the number of people seeking work, rather than a significant increase in jobs. Researcher Kim Sung-no pointed out that considering inflation pressure, the Taylor rule interest rate would remain at the 6.82% level seen in May. He pointed out that the market also reflects the possibility of a 25bp interest rate hike in September or October. One side read the possibility of a rebound in growth stocks by looking at falling bond yields, while the other side saw that interest rate caution has not ended by looking at the falling unemployment rate and inflationary pressures.
There are also warnings that the concentration in semiconductors has become too extreme. BNK Securities assessed that as money flocks only to semiconductors, more stocks in the market are being traded at excessively cheap prices. It notes that companies whose stock market capitalization is lower than the cash they hold are emerging one after another. At the same time, the KOSPI Volatility Index (V-KOSPI) has risen to historical highs, indicating growing market anxiety. Consequently, they are paying attention to bio/healthcare.
In June, bio/healthcare exports reached a record high of $1.92 billion. The global healthcare index also hit an all-time high. However, domestic healthcare stock prices remain suppressed. The 12-month forward price-to-earnings ratio (PER) and price-to-book ratio (PBR) have dropped to their lowest levels since 2015. This means performance and export indicators are improving, but stock prices are being neglected. Researcher Kim Sung-no stated, "A perspective that seeks new investment opportunities outside of semiconductors is necessary."
So, how should individual investors respond? Brokerages point out that while the semiconductor profit cycle is still alive, the formula of the last few months—where one could load their entire account into semiconductors and succeed—may be reaching its expiration date. With this as a premise, it is necessary to revisit a few principles.
First, now is a period for partial approaches, neither dumping stocks nor panic buying. While the semiconductor profit outlook itself is still being upgraded, Samsung Electronics is scheduled to announce its earnings this week, and SpaceX will officially be included in the Nasdaq 100 before the market opens on the 7th. It is a high-speed inclusion occurring just 15 trading days after listing, following the fast-track rule introduced by Nasdaq last month, with JPMorgan estimating an inflow of approximately $4.3 billion in passive funds. As volatility factors are lined up, there is no reason to adjust the allocation all at once.
Second, it is time to check the opposite side of the concentration. Hana Securities named secondary batteries, shipbuilding, power equipment, and defense, alongside semiconductors within the KOSPI, as sectors where profit outlooks and operating profit margins are rising together, while BNK Securities pointed to bio/healthcare, where valuations are at historical lows despite record-high exports. Redistributing the weight of the portfolio across multiple sectors supported by earnings, exports, and valuations is the practical task given to individual investors during this adjustment phase.
Third, it is good to set criteria for judgment in advance. One can determine whether this adjustment is a brief breather or the beginning of a downward trend by watching whether the hyperscalers' FCF turns to a surplus in the first quarter of 2027 as expected, and whether the rebound in the short-long term interest rate gap continues. In a phase where the volatility index is at an all-time high, betting on a rebound with leveraged products is the choice furthest from all these principles. While the final decision on stock selection and investment rests with the investors themselves, it is worth remembering that when the market looks only at one sector, the numbers often point elsewhere.