South Korean chip stocks, notably SK Hynix and Samsung Electronics, are experiencing historic gains and are poised for further significant rallies, according to Nomura. The investment bank predicts over 110% upside for both companies, driven by a “structural growth phase” in memory demand, particularly for high-bandwidth memory (HBM) chips crucial for AI. With demand projected to far outstrip supply, the companies are set for unprecedented revenue and profit expansion in the coming years.
South Korea's leading semiconductor stocks are not just breaking records; they're setting new benchmarks, propelling the Kospi index to unprecedented heights. And the rally, according to Nomura, is far from over. The Kospi index, which surged an impressive 75% in 2025, has already chalked up a 78% gain this year, largely driven by the spectacular performance of two domestic chip powerhouses: SK Hynix and Samsung Electronics.
These two giants have seen their shares skyrocket, with SK Hynix climbing 183% year-to-date (following a 274% gain in 2025) and Samsung Electronics surging 134% this year (after a 125% rise in 2025). The future looks even brighter as demand for high-performance memory chips, essential for artificial intelligence applications, continues its exponential growth.
Nomura, holding a "buy" rating on both companies, projects Samsung – the world's largest memory manufacturer – could see its shares rally by over 110% this year, while SK Hynix might climb as much as 117% from Monday's closing prices. The brokerage firm has set ambitious 12-month price targets: 4 million won for SK Hynix and 590,000 won for Samsung Electronics.
This optimistic outlook stems from the memory industry entering a "structural growth phase" catalyzed by the launch of ChatGPT in December 2022. This event ignited a massive increase in demand for high-bandwidth memory (HBM), with Samsung and SK Hynix standing as key global producers. SK Hynix, a frontrunner in the HBM sector, is anticipated to lead the charge in advanced HBM4 chip production, with Samsung also positioned as a formidable competitor.
Analyst reports from Nomura indicate that customer demand for HBM and other high-performance memory solutions already surpasses the industry's medium-to-long-term supply capabilities. They foresee memory demand potentially rising by "several thousand-fold" over the next five years, significantly outstripping the projected five to six-fold increase in industry supply.
Further fueling this demand, the adoption of Retrieval-Augmented Generation (RAG) and emerging agentic AI applications is accelerating the need for conventional servers and solid-state drives (SSDs). RAG, a process where large language models consult external knowledge bases for accurate responses, reduces AI errors and hallucinations. Nomura identifies a "triple memory super-cycle spanning DRAM, HBM and SSD memory" that began in the third quarter of 2025.
Dynamic random access memory (DRAM) plays a crucial role in the AI ecosystem, supporting model training and real-time inference. Moreover, the evolution of agentic AI is expected to spur new categories of semiconductor demand, including CPUs and commodity memory, which will further boost the usage of AI servers and drive substantial memory-related growth.
Against this robust structural backdrop, Nomura believes memory vendors have entered an unparalleled period of rapid revenue growth and margin expansion. The brokerage anticipates annual revenue and earnings growth of approximately 30% for memory suppliers over the next three to five years, building on an estimated seven to eight-fold profit increase in 2026. This growth is expected to be underpinned by a 30% annual rise in volume, stable or slightly increasing commodity memory prices via long-term agreements, and enhanced HBM profitability.
Early financial results from SK Hynix and Samsung already underscore this trend. SK Hynix reported a fivefold year-on-year increase in operating profit in the first quarter of 2026, while Samsung's operating profit soared by more than 750% during the same period.
