“The Era of AI Inference is Coming” Semiconductors are Not Over Yet

Park Seung-young, Team Head, stated, “I just think the 5,500 to 7,400 range is quite strong.”

This analysis suggests that despite improvements in corporate earnings, high interest rates will dampen investor sentiment and limit the upward movement of the index.

Park Seung-young, Team Head, explained, “Samsung Electronics and Hynix are almost at their July peak levels. This means that expectations for Samsung Electronics and Hynix are still quite high.”

While margin loans in other sectors, such as KOSDAQ, are declining, high expectations for these two large semiconductor stocks continue to act as a heavy factor hindering the overall market's spread.

The analysis indicates that while in the past, expectations for the memory cycle drove the market through these two stocks, there are currently limitations due to a lack of market breadth, no upward revisions in earnings forecasts, and the exhaustion of share buyback effects.

The current rise in raw material prices indicates entry into the late stage.

A warning has been raised that the market is underestimating geopolitical risks. Park Seung-young, Team Head, cautioned against excessive optimism, stating, “There's an expectation that even if a war breaks out, Trump will back down. Waiting, thinking 'it'll be fine,' is even more dangerous.”

The market reflects expectations that geopolitical issues would be resolved if former President Trump were elected, but this optimistic view itself could be a risk factor.

An analysis suggests that there is institutional bias in interpreting crude oil supply and demand data, and Iran's production variable could significantly impact oil prices. Park Seung-young, Team Head, pointed out, “There's supply and demand data provided by the U.S. Energy Information Administration and data provided by OPEC. Both issue monthly forecasts, but when the U.S. forecasts, it always shows more supply. And when OPEC presents its forecast, it always shows a supply shortage.”

The U.S. Energy Information Administration (EIA) forecasts an oversupply, while OPEC forecasts a supply shortage, showing conflicting perspectives. In particular, Iran's production capacity of 3-4 million barrels acts as a key variable in oil price fluctuations.

Korean companies are mostly producers of intermediate goods, making them vulnerable to raw material price fluctuations, necessitating a hedging strategy. Park Seung-young, Team Head, explained, “It's difficult for those who invest specifically in Korean stocks to hedge. Because most of our companies produce intermediate goods, and for intermediate goods, raw material prices and selling prices always move together. But if you look at the earnings, due to FIFO, the cost of goods sold reflects raw material prices from about six months ago.”

Due to the first-in, first-out (FIFO) accounting method, changes in raw material prices are reflected in actual corporate earnings approximately six months later. To counter this, he suggested buying U.S. energy stocks or utilizing commodity ETFs.

The opinion was presented that with WTI oil prices reaching $94, a strategic response is necessary, anticipating the possibility of breaking $100 in the future. Park Seung-young, Team Head, advised, “It's probably a better position to think that it might briefly go up to $100 and then come down. Maintain your stock allocation similarly, and sell off anything that has risen too much in advance.”

Rather than reducing the entire stock allocation, it is important to proactively sell off stocks that have already risen significantly and to respond flexibly, considering that the upside of the Korean stock market is limited.

An analysis suggests that the current market situation is markedly different from this past spring. Park Seung-young, Team Head, stated, “Now, stock prices have risen a lot, and third-quarter earnings forecasts aren't going up much, because most of them were raised in the first quarter. So expectations are already largely priced in. From now on, we need to look at next year's picture, but since we don't yet know what next year's earnings will be, it's definitely different from the spring.”

While stock prices rose in the spring due to earnings expectations, now prices have already risen significantly, and third-quarter earnings forecasts have already been reflected, extinguishing expectations. Moreover, increasing uncertainty about next year's earnings is acting as a factor weighing down the market's upside.

Based on the not-so-hawkish tone of remarks from the New York Fed President and the Fed Chair, the likelihood of a September rate hike was assessed as low. However, the possibility of one additional hike by year-end remains open.

This reflects the Fed's cautious stance, considering both persistent inflationary pressures and concerns about an economic slowdown.

Park Seung-young, Team Head, emphasized, “I don't think it's right to allocate more than 10% of your total portfolio.”

Alternative assets tend to move primarily during periods of dollar weakness, and current concerns about U.S. fiscal instability are somewhat exaggerated, which should also be considered.

Park Seung-young, Team Head, said, “I don't think the U.S. fiscal situation and the state of U.S. Treasury yields are as bad as the market worries.”

The ratio of interest payments to U.S. fiscal revenue is about 17%, which is lower than the risk threshold of 20%. Furthermore, the proportion of foreign-held Treasury bonds and short-term bonds also remains around 20%.

Considering these data, the analysis concludes that it is still premature to judge that confidence in U.S. Treasury bonds will decline.

Park Seung-young, Team Head, commented, “I think if U.S. Treasury yields rise significantly from here, many people will buy Treasury bonds.”

As factors suppressing foreign demand, such as from China, are resolved, bond prices are expected to fall, naturally attracting buying interest. The environment for issuing long-term U.S. Treasury bonds is still assessed as stable.

The trend of AI model competition in the AI market is changing, and OpenAI's role is once again drawing attention. Park Seung-young, Team Head, explained, “In the past, good models were released every year. But recently, cost-effective models have been coming out.”

Last year, performance competition was dominant, but this year, cost-effective models became important, leading to a shift of developers from Claude to OpenAI (Codex).

However, the release of new models by OpenAI could once again trigger performance competition in AI models.

Park Seung-young, Team Head, explained, “Don't say, 'Oh, the semiconductor market is range-bound now' when I just said it would be range-bound, but it won't fall much. So buy when it falls a lot, and when it rises too much, don't think, 'Oh, it'll keep hitting the top of the range and going up.'”

Even in a range-bound market, semiconductor stock prices are likely to experience limited corrections rather than sharp declines, so it's advisable to hold or buy at appropriate times rather than selling out of concern.

This is based on market capitalization ratios, and for Samsung Electronics, prioritizing DRAM growth and dividend yield over foundry expectations is a more rational investment perspective, he added.

The first is stock-centric asset allocation, where the equity portion should be the highest. Park Seung-young, Team Head, explained, “You should allocate assets centered on stocks. This is the first principle. You shouldn't hold the most bonds. Stocks should be the highest, centered on this.”

The second is sufficient diversification. Risks should be reduced by diversifying across 6-8 asset classes. Finally, the principle is to always plan investment strategies considering taxes. He emphasized, “Lastly, you should always consider taxes.”

Park Seung-young, Team Head, said, “I think you should hold half Korean stocks and half U.S. stocks. Even if you invest in the index with half Korean stocks and half U.S. stocks, Samsung Electronics and Hynix will always make up 50% of the Korean portion, and U.S. big tech will always be 35-40% of the U.S. portion.”

Even with index investing, semiconductor and big tech weights are automatically secured, so when these leading sectors are deemed to have risen too much, it is advisable to consider rotating into other sectors. This is a strategy to flexibly respond to market changes and capture profit opportunities.