2026년 2월 24일 · Unknown · financial · 출처 Yahoo Finance
This article first appeared on GuruFocus.
Chairman Chey Tae-won of SK Hynix's (HXSCL) parent SK Group has signaled that the company intends to deepen its exposure to the artificial intelligence buildout, pledging to expand production of high-bandwidth memory chips as global data center demand accelerates. Speaking in Washington on Feb. 20, Chey described HBM as a monster chip that is generating enormous profits for SK Hynix, underscoring how central the product has become to the AI hardware stack. The backdrop is a sharp re-rating in SK Hynix's shares, which have more than quadrupled over the past year on record earnings, reflecting investor conviction that advanced memory remains a critical bottleneck in the AI supply chain.
Is HXSCL fairly valued? Test your thesis with our free DCF calculator.
The capital cycle supporting that view appears substantial. US technology leaders from Microsoft (NASDAQ:MSFT) to Meta Platforms (NASDAQ:META) are allocating about $650 billion this year toward infrastructure designed to strengthen their AI capabilities, a record outlay that is contributing to a global shortage of memory chips. That supply tightness has largely benefited SK Hynix, Samsung Electronics (SSNLF), and Micron Technology (NASDAQ:MU), the dominant players in the segment. SK Hynix has already sold out its entire slate of memory chips for 2026, while Micron has similarly exhausted its HBM offerings. In January, SK Hynix indicated that its 2026 capital expenditure will rise significantly from last year's level to meet demand for HBM used in accelerators designed by companies such as Nvidia (NASDAQ:NVDA) to train and run AI services.
Even so, Chey tempered the optimism with a reminder that semiconductor cycles can turn quickly. The average projection for SK Hynix's 2026 annual operating profit has climbed to $70 billion in January from roughly $50 billion late last year, and some estimates have been revised upward to more than $100 billion. Chey acknowledged that such figures sound like really good news, but cautioned they could just as easily turn into a $100 billion loss, pointing to the risk of rapid technological shifts reshaping the competitive landscape. He also flagged mounting infrastructure constraints, noting that SK Group is exploring building power plants alongside AI data centers, arguing that failing to meet energy demand could be disastrous.
View Comments