[Yonhap News Agency] 사진 확대 [Yonhap News Agency]

As the global competition over artificial intelligence (AI) performance and efficiency intensifies by the day, demand for memory semiconductors is expected to expand again. With a rebound in infrastructure-related stocks expected to continue, Samsung Electronics and SK hynix have been cited as the top picks.

On the 8th, Kim Dong-won, head of the Research Division at KB Securities, said, “Next year, Samsung Electronics and SK hynix will benefit as demand for sixth-generation high-bandwidth memory (HBM4), server double-data-rate (DDR5) memory, and enterprise solid-state drives (SSDs) is expected to rise sharply.”

He added, “OpenAI’s latest model, ‘GPT-6 Astra,’ will mark a turning point toward entering the early stages of artificial general intelligence (AGI), going beyond simply improving AI model performance. In China, highly efficient models such as Alibaba Group’s Qwen are lowering the barrier to AI adoption by offering low inference costs.”

He further predicted, “As frontier models in the United States expand the upper limits of performance and efficiency-focused models in China lower the floor for costs, AI will spread rapidly in both high-performance and mass-market segments. The surge in AI computing workloads is expected to accelerate structural growth in demand for memory products such as HBM, DRAM, and NAND flash.”

He also noted, “Although AI data-center expansion is currently focused solely on the United States, China’s AI data centers are also expected to grow. AI data centers in the United States and China are projected to expand from 95 gigawatts (GW) this year to 201 GW by 2030.”

This article has been translated by GripLabs Mingo AI.

#

Samsung Electronics and SK hynix

data centers

artificial intelligence

Samsung Electronics

SK hynix

KB Securities