'Selling Samsung, SK hynix now would be a big mistake,' experts say

AI-related stocks, including Samsung Electronics and SK hynix, remain promising investments, analysts say, with the AI industry's expansion and productivity gains expected to continue at least through 2030.

Yu Dong-won, head of the global asset allocation division at Yuanta Securities Korea, told attendees at the Herald Money Festa 2026 conference, held recently at Dongdaemun Design Plaza in Seoul, that investors who sell Samsung Electronics and SK hynix now "could find themselves bitterly regretting it two or three years down the road."

"The AI cycle has not yet reached its peak, and investors should continue to hold related positions," Yu said, urging investors not to cut their semiconductor exposure in the domestic market too sharply.

"Even if you broaden your portfolio beyond semiconductors, the question should be how much less of them to hold — not whether to abandon them altogether for other sectors," he said. "Among Korean industries, semiconductors stand out for their low debt ratios, high growth potential, direct ties to US consumption and the largest export volumes."

Addressing concerns about an AI bubble, Yu said historical patterns from the internet cycle suggest the current AI boom is far from over. "If share prices recover to their previous highs within six to nine months of a decline, it is hard to say that cycle has ended," he said. "The internet took about 30 years to spread globally, but AI is expanding far more quickly. I believe the AI cycle will continue at least through 2030."

Yu identified South Korea, the United States and Taiwan as the most promising global investment destinations. "The next four to five years call for a focus on the US, South Korea and Taiwan," he said, pointing to the NASDAQ 100, Kospi and Taiwan's Weighted Index as the key benchmarks to watch.

Choi Chang-gyu, head of the ETF consulting division at Mirae Asset Global Investments, also spoke at the Herald Money Festa 2026, advising investors holding Samsung Electronics and SK hynix to trust the scale of capital expenditure commitments. "The more data centers are built, the more demand will grow for both HBM and NAND," he said. Capital expenditure, or capex, refers to spending on AI infrastructure such as data centers and servers.

Choi said interest rates remain a headwind but are not enough to derail the AI industry. "Interest rates are holding things back, but they will not break the AI industry," he said. "I do not see any fundamental problem arising for the AI sector."