[Edaily Reporter Kim Kyung-eun ] SamsungElectronics(005930)and SK hynix(000660): Forecasts from domestic and international securities firms regarding target stock prices have diverged by more than twofold. While global investment bank Morgan Stanley has significantly lowered its target price for SK hynix to 1.6 million won, domestic securities firms are suggesting prices as high as 4 million won. Amid growing concerns over a slowdown in artificial intelligence (AI) investment, differences in perspective regarding the sustainability of the memory semiconductor boom appear to be widening.

According to the financial investment industry on the 11th, Morgan Stanley recently lowered its target price for SK hynix by 38.5%, from 2.6 million won to 1.6 million won. It also lowered its target price for SamsungElectronics by 18.9%, from 381,000 won to 309,000 won. This move reflects concerns about semiconductor profitability amid growing regulatory uncertainty surrounding AI data center investments following the U.S. midterm elections.

In contrast, domestic securities firms continue to maintain an optimistic outlook. KYOBOSECURITIESCO.,LTD. and SKSecurities set the highest target price for SK hynix at 4 million won—2.5 times higher than Morgan Stanley’s 1.6 million won.

Yuanta Securities Korea also set the highest target price for SamsungElectronics at 690,000 won, a 2.2-fold difference from Morgan Stanley’s figure. On the 29th of last month, Yuanta Securities Korea raised its target price for SamsungElectronics from 630,000 won to 690,000 won. Additionally, SKSecurities and KB Securities are maintaining high target prices of 610,000 won and 600,000 won, respectively.

There is also a significant gap in target prices among domestic securities firms. BNK Investment & Securities set SamsungElectronics’ target price at 270,000 won, IBK Investment & Securities at 350,000 won, and iM Securities at 370,000 won. The gap between Yuanta Securities Korea, which set the highest target price, and BNK Investment & Securities, which set the lowest, is 2.6 times.

The divergent outlooks among securities firms stem from differing views on long-term growth potential. While positive outlooks on the semiconductor market are generally maintained, opinions are split regarding the impact that an expansion in semiconductor supply after 2028 will have on profitability.

Morgan Stanley projected that while memory supply will increase starting in late 2027 as new production facilities come online, the pace of AI data center construction will slow. The firm factored in the possibility that a data center project with a capacity of approximately 5 gigawatts (GW)—originally scheduled to begin operations in 2028—could be delayed until 2029 due to regulatory uncertainty following the U.S. midterm elections.

Analyst Sean Kim, who authored the report, stated, “The most critical debate in the memory semiconductor market right now is whether the ‘2028 earnings cliff’ can be avoided,” adding, “Concerns are growing not so much about next year’s earnings themselves, but rather about whether growth can be achieved and sustained after 2028.”

In contrast, domestic securities firms believe that the imbalance between memory supply and demand will not be easily resolved despite increased supply. Shinhan Investment Securities projected that the supply growth rates for DRAM and NAND flash next year will be 21% and 23%, respectively, assessing that these levels will be insufficient to meet demand. Demand for High-Bandwidth Memory (HBM) is also projected to rise by 69%, from 35 billion gigabits (Gb) this year to 59 billion Gb next year. In particular, analysts believe that the expansion of long-term supply agreements (LTAs) will prevent a sharp deterioration in market conditions similar to what was seen in the past.

Some analysts also suggest that tighter U.S. AI regulations are unlikely to lead to an actual reduction in investment. KB Securities concluded that, given the data center projects already underway, the likelihood of a sharp contraction in AI infrastructure investment by 2028 is low.

The key factors that will determine the future direction of stock prices are expected to be the pace of AI investment expansion and changes in actual semiconductor orders. SKSecurities analyzed that during past downturns in the memory semiconductor market, a recurring pattern emerged: after stock prices peaked, customer orders contracted, followed by a decline in memory prices. However, the firm explained that this time, no signs of order contraction have been observed even after stock prices peaked last June.

Han Dong-hee, an analyst at SKSecurities, said, “While concerns about delays in AI data center construction persist, a slowdown in the pace of construction will enhance the value of AI infrastructure and strengthen the case for investment,” adding, “Memory stocks will continue to prove their worth through earnings, and the current stock price level presents an opportunity.”