Jeon In-gu

Jeon In-gu views Samsung Electronics and SK Hynix favorably but expects extreme volatility because memory demand and volumes are known while memory unit price direction is unpredictable. After KOSPI peaked near 9,300 and SK Hynix roughly halved, popular national stocks historically do not collapse all at once and tend to produce at least a second and third rally, so he sees the current rebound as a second upleg. However, near SK Hynix 2 million won strong selling may emerge unless earnings, guidance upgrades, or memory price increases appear, and he expects a stressful second half with money gradually rotating to other sectors.

Jeon In-gu views Samsung Electronics and SK Hynix favorably but expects extreme volatility because memory demand and volumes are known while memory unit price direction is unpredictable. After KOSPI peaked near 9,300 and SK Hynix roughly halved, popular national stocks historically do not collapse all at once and tend to produce at least a second and third rally, so he sees the current rebound as a second upleg. However, near SK Hynix 2 million won strong selling may emerge unless earnings, guidance upgrades, or memory price increases appear, and he expects a stressful second half with money gradually rotating to other sectors.

Samsung Biologics and Celltrion have steadily rising revenue and operating profit while their share prices have fallen, pushing PER toward value-stock levels. They may increasingly emphasize shareholder returns and become dividend-like stocks, are expanding overseas, and benefit structurally from aging demographics and older cohorts with greater spending power.

Samsung Biologics and Celltrion have steadily rising revenue and operating profit while their share prices have fallen, pushing PER toward value-stock levels. They may increasingly emphasize shareholder returns and become dividend-like stocks, are expanding overseas, and benefit structurally from aging demographics and older cohorts with greater spending power.

As part of a personal Plan B, about 40% of the portfolio can go to companies capable of growing annual revenue and profit by more than 10% that have been neglected by AI and have become cheap. He names Netflix and Airbnb as examples of companies that were previously too expensive but are now reasonably priced, offering long-term stability even if share prices fall.

As part of a personal Plan B, about 40% of the portfolio can go to companies capable of growing annual revenue and profit by more than 10% that have been neglected by AI and have become cheap. He names Netflix and Airbnb as examples of companies that were previously too expensive but are now reasonably priced, offering long-term stability even if share prices fall.

Nvidia's GPU rental prices are still rising even for older models, showing explosive cloud demand. Nvidia is now backstopping GPU value to bring in Wall Street funding for data centers, which should accelerate data center construction. Since about 40% of AI data center cost is GPUs, Nvidia is the biggest beneficiary, while Samsung Electronics, SK hynix, and Micron also benefit near term.

Korea's 18.5GW data center mega project is led by SK Telecom for about 15GW, and SK Telecom's CEO met Jensen Huang with SK Chairman Chey Tae-won. He argues that canceled US data center projects threatened GPU/HBM demand, so Korean builds offset that demand. Sovereign AI data centers are becoming national strategic assets, and if non-US builds accelerate, GPU/HBM prices, memory stocks, and KOSPI can rise again. SK Telecom is the clear frontrunner with an Anthropic stake and has already rallied, while other Korean telecoms have not yet detailed funding and operation plans but may catch up once they announce them.

Korean beauty can become the new leading striker after Samsung Electronics and SK Hynix. At a June beauty seminar he found a consensus that European sales were surging, Korean brands were increasingly entering Amazon's top 100, and the mix was shifting from cheap color cosmetics toward higher-value skincare. Europe opening as a continent means the trend likely persists for years. Many beauty companies double net profit every three years yet trade at two-year-forward PERs of 5-10x, offering cheap high growth with no AI linkage. APR is his preferred leader because of its brand portfolio, larger US top-100 presence, and better US response, balancing growth and stability for long-term holding of at least three years.

Korea's 18.5GW data center mega project is led by SK Telecom for about 15GW, and SK Telecom's CEO met Jensen Huang with SK Chairman Chey Tae-won. He argues that canceled US data center projects threatened GPU/HBM demand, so Korean builds offset that demand. Sovereign AI data centers are becoming national strategic assets, and if non-US builds accelerate, GPU/HBM prices, memory stocks, and KOSPI can rise again. SK Telecom is the clear frontrunner with an Anthropic stake and has already rallied, while other Korean telecoms have not yet detailed funding and operation plans but may catch up once they announce them.

The AI theme is not broken but rotating. Memory earnings look best in 2027 with 2028 guidance not clearly better, so investors are shifting money toward areas with visible high growth. He says US cloud, data center, optical, and software names are currently favored because big tech earnings and cloud growth remain strong, while software stocks bottomed and rebounded. Korea lacks pure large-cap cloud, data center, and global software names, so this rotation is expressed through US sectors.

Jeon In-gu expects high oil prices to persist around or above $80 for a while because the Iran-US and Hormuz Strait situation is difficult to resolve before the November US midterms, and Iran may benefit from keeping pressure on Trump. He cautions that oil forecasting is unreliable, but the current geopolitical setup points to sustained high oil prices rather than a quick resolution.