Korean Stocks Sink as Memory Chipmakers Plunge on Deepening AI Spending Fatigue

The Quick Take

• South Korean equities fell sharply, led by memory chip giants as investor sentiment on the AI spending cycle deteriorated
• The sell-off deepens a pattern of AI-fatigue-driven volatility that has weighed on SK Hynix and Samsung Electronics
• Korean chipmakers' vulnerability to AI capex cycle sentiment creates spillover risk for Taiwan, Japan semiconductor peers

South Korean equities registered a broad decline on Monday, with the memory chipmaking sector leading the fall as investors increasingly question the durability of AI-driven demand for high-bandwidth memory (HBM) and DRAM. According to Bloomberg, sentiment continues to worsen on the artificial intelligence boom — a reversal from the euphoric period of 2023-2025 when SK Hynix and Samsung Electronics were priced as primary beneficiaries of hyperscaler data center buildouts. The sell-off reflects growing concern that AI infrastructure spending is front-loaded, with the demand cycle potentially peaking before a sustainable revenue inflection materializes for chip suppliers.

The Korean memory sector's AI fatigue dynamic has direct implications for the global semiconductor supply chain. TSMC's logic chip orders, Micron Technology's HBM ramp, and Japan's Kioxia NAND flash shipments all face the same underlying sentiment headwind. For equity markets, the key concern is whether the current AI capex cycle — dominated by NVIDIA GPU purchases and custom silicon from hyperscalers — is creating overcapacity downstream in DRAM/NAND even as logic chip demand remains buoyant. Korean chipmakers uniquely straddle both cycles, making them a leading indicator for the broader semiconductor sector's risk appetite.

Forward signals to monitor include SK Hynix's Q2 earnings report and guidance on HBM order visibility, which will either validate or refute the AI-fatigue thesis for memory. Samsung's foundry utilization rates will indicate whether the logic chip side can compensate for any memory demand softness. The macro variable: hyperscaler capex announcements — particularly from Microsoft, Google, and Amazon — in the coming earnings season will determine whether AI infrastructure spending is decelerating or merely pausing ahead of the next wave of AI model deployments.

Synthesized from 1 source.