If U.S. long-term interest rates remain high, funds in the stock market could concentrate in a handful of stocks, according to an outlook. In particular, Samsung Electronics and SK hynix, which have strong shareholder returns without relying on external capital, could emerge as havens.

On the 20th, a display board in the dealing room at Hana Bank's headquarters in Jung District, Seoul shows the KOSPI and the share prices of SK hynix and Samsung Electronics. On the 19th, SK hynix holds a board meeting to approve a share repurchase and cancellation plan and discloses its shareholder return policy./Courtesy of Yonhap News.

Hwang Su-uk of Meritz Securities said, "The supply-demand burden for long-dated Treasurys stemming from expanded artificial intelligence (AI) capital expenditures (CAPEX) has not been resolved despite the U.S. Treasury's increased buybacks of Government Bonds," and noted, "Until U.S. long-term rates stabilize, it will be difficult for global risk asset valuations to expand."

Hwang cited falling oil prices, a slowdown in AI capital spending, and intervention by the Federal Reserve (Fed) as conditions for stabilizing long-term rates. However, with a low likelihood of oil prices falling and little chance that an AI investment slowdown will materialize, Hwang said attention should be paid to the possibility of Fed intervention.

Hwang pointed to the need to watch whether the Fed is coordinating measures such as easing risk-weighted asset (RW) rules on commercial banks' corporate bonds, adjusting capital rules for market risk (FRTB), and relaxing additional capital requirements for global systemically important banks (GSIBs).

In the end, Hwang projected growing demand for corporations that can deliver growth exceeding AI expense. "The higher rates go, the scarcer become companies that can fund investments with their own cash flow rather than relying on external capital, and return increased profits to shareholders," Hwang said. "Flows may concentrate in the semiconductor 'top two,' where profit forecasts are being raised on AI memory demand alongside large-scale shareholder returns."

※ This article has been translated by AI. Share your feedback here.