JPMorgan initiates SK Hynix at Overweight with $245 price target

JPMorgan just handed SK Hynix one of the more bullish initiations in the semiconductor space this year, slapping an Overweight rating and a $245 price target on the memory giant. The target is set for June 2027, and the thesis boils down to one word: AI.

The bank’s analysts project a 34% compound annual growth rate in earnings per share over the next two years, fueled by what they describe as an AI-driven memory upcycle that could stretch beyond five years. For a company already posting gross margins around 76%, that kind of earnings trajectory makes the current valuation look, well, modest.

The numbers behind the conviction

SK Hynix currently trades at a price-to-earnings multiple of roughly 7.1x.

JPMorgan’s $245 target implies approximately a 20% premium over SK Hynix’s native share price when translated through its American Depositary Receipt listing. The target works out to about 7x average EPS for fiscal years 2026 and 2027.

Memory average selling prices are forecasted to climb from Q1 2024 through at least Q4 2028. That’s a four-plus year pricing tailwind.

Why AI changes the memory math

JPMorgan’s research highlights an expectation that memory content per AI system will increase significantly in next-generation architectures.

Over 50% of SK Hynix’s production capacity is already locked up through long-term agreements, providing revenue visibility that supports multi-year growth projections.

Shareholder returns and competitive positioning

SK Hynix has also revamped its capital return policy, committing more than 50% of free cash flow to investors.

The company’s US market expansion through its ADR listing has attracted broader institutional interest. Analyst price targets from various firms range from $240 to $320, meaning JPMorgan’s $245 sits near the conservative end of the spectrum.

The memory market is essentially a three-player game: Samsung, SK Hynix, and Micron. SK Hynix has carved out a strong position in high-bandwidth memory, the specific type of DRAM that AI accelerators require.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.