BNK Investment & Securities analyst Lee Min-hee issued a 'hold' rating with a ₩1.85 million target price for SK Hynix stocks last week, when the stock was trading above ₩2.2 million — effectively a sell recommendation. The analyst cited anticipated slowdown in hyperscaler capital expenditure starting in the second half as the primary reason, stating semiconductor momentum would peak in the second quarter. This followed Morgan Stanley's recommendation to reduce exposure to memory semiconductor stocks, intensifying domestic concerns about demand reaching its peak.
Lee stated in an interview on the 13th that "semiconductors will peak in the second quarter, with momentum slowing from the second half." The analyst explained that "the biggest backdrop is the increased possibility of hyperscalers adjusting their capex pace," adding that "hyperscalers' competitive infrastructure investment is no longer effective." Memory semiconductor prices had risen as hyperscalers rushed AI server infrastructure investment, but demand is expected to decline from the second half according to the analyst's forecast.
The analyst pointed to deteriorating leading economic indicators in financial markets, including the US yield curve spread. Lee explained that "US Treasury yields continue to rise, and this indicator leads the growth rate of hyperscalers' AI capex," noting that "most capex costs are raised in the corporate bond market, but Meta and Microsoft's corporate bonds are losing popularity and cash is drying up."
Meta's recent entry into cloud services is interpreted as negative news in the semiconductor market. Meta announced plans to provide surplus computing resources built as AI infrastructure to external customers. This has led to interpretations that hyperscalers will ease infrastructure competition while considering profitability. Lee analyzed that "(Meta's move) shows recognition that there has been 'overinvestment' so far."
The analyst also predicted decreased demand from IT companies producing finished products like smartphones. Apple raised prices of major products by up to 20% last month, stating that "price increases are unavoidable due to memory prices." The analysis suggests that while high-end smartphones can pass costs to consumers through price increases, mid-to-low tier smartphones face difficulty raising prices, which will fuel decreased semiconductor demand.
Lee assessed that the premium from SK Hynix's American Depositary Receipt (ADR) listing would not be significant. The analyst stated, "If ADR listing automatically brings a premium like US-based Micron, wouldn't Samsung Electronics also list via ADR?" adding that "governance structure differs from US companies and capital markets differ, so market valuation itself is different." Lee explained that "while overseas investor accessibility to stocks can improve, valuation will not change."
SK Hynix stock price declined from early trading that day, falling as much as 15% during the session. The stock broke below even the ₩1.85 million target price set by the analyst. Lee had also issued a 'hold' opinion report on SK Hynix in May, receiving criticism from retail investors. The analyst said complaints poured into the company following this report as well, stating "how can (an analyst) predict all stock prices accurately?" and adding "my role is to explain what risks exist and what the current situation is — investors make the judgment."
What target price did BNK Investment & Securities set for SK Hynix stocks?
BNK Investment & Securities analyst Lee Min-hee set a target price of ₩1.85 million for SK Hynix stocks last week with a 'hold' rating, when the stock was trading above ₩2.2 million.
Why does the analyst expect semiconductor demand to slow in the second half?
The analyst cited increased possibility of hyperscalers adjusting capital expenditure pace as the primary reason, noting that rising US Treasury yields are constraining financing as Meta and Microsoft corporate bonds lose popularity, and that Meta's cloud business entry signals recognition of prior overinvestment in AI infrastructure.
Related News