The Game Is About to Change: Samsung Electronics and SK hynix Poised for Revaluation as "GPUs Are the Heart, Memory Is the Blood Vessel" [Weekend Money]

KB Securities: "Memory Semiconductor Bottlenecks Expected to Continue"

On September 24, KB Securities reported that the core focus of artificial intelligence (AI) infrastructure investment is rapidly expanding from graphics processing units (GPUs) to central processing units (CPUs) and memory semiconductors. The firm analyzed that Samsung Electronics and SK hynix are expected to be strategically reassessed as key assets of AI infrastructure, beyond being mere memory suppliers.

Dongwon Kim, Head of Research at KB Securities, stated, "The recent surge in the stock prices of Intel, AMD, and ARM on the New York Stock Exchange was driven by growing expectations that AI infrastructure investment is expanding beyond GPUs to CPUs and memory. As monetization prospects for AI increase, especially with the proliferation of Meta's AI agent services, hyperscalers are likely to sustain their infrastructure investments, and the visibility of demand for memory is also expected to remain strong over the long term."

Kim further explained, "In particular, AI agents require much more frequent and continuous inference operations compared to conventional generative AI. This structurally increases overall memory usage in AI data centers, not only for CPUs and GPUs but also for high-bandwidth memory (HBM), DDR5, and eSSD."

He added, "AMD became the fourth U.S. semiconductor company after Nvidia, Broadcom, and Micron to surpass a market capitalization of 1 trillion dollars, signaling that the bottleneck for AI is shifting toward memory. The increase in server CPUs is driving greater demand for high-capacity DDR5 server DRAM, while the growth in GPUs and AI accelerators is simultaneously boosting HBM demand."

He also noted, "In the era of agentic AI, where high processing performance at the level of 1,000 tokens per second is required, the core competitiveness will depend on how much and how quickly data can be supplied. Samsung Electronics is responding to this by developing next-generation 'zHBM', which involves stacking memory on top of GPUs, with sample supply expected by the end of 2027."

Kim emphasized, "If GPUs are the heart, CPUs the brain, and electricity the oxygen of AI data centers, then memory is the bloodstream that keeps data circulating endlessly. As the locus of performance and efficiency shifts to memory, we expect a strategic reappraisal of Samsung Electronics and SK hynix, Korea's leading semiconductor firms."

This content was produced with the assistance of AI translation services.