Following Kiwoom, Mirae Asset Joins Race for MLCC ETFs, the ‘Rice of the Electronics Industry’

The introduction of MLCC-focused ETFs by asset managers suggests a broadening of investment themes related to AI beyond primary memory semiconductors. MLCCs are critical for stable power delivery in electronic devices, and their demand is reportedly escalating due to the higher power consumption of AI chips. For example, Kiwoom Securities estimates the number of MLCCs used per GPU could rise from approximately 200 for Nvidia’s H100 to 5,000 for the next-generation Rubin.

This surge in demand is reportedly outpacing supply. While MLCCs for AI servers represent a small fraction of total shipments (2-3%), they are high-specification products that occupy production lines four to seven times more than general-purpose MLCCs. Manufacturers shifting capacity to AI applications also reduces availability for standard MLCCs. Industry production capacity is expected to grow 10-15% annually, yet AI-related MLCC demand is projected to increase by over 80% per year, contributing to a supply-demand imbalance.

The imbalance has led to price increases and longer lead times. Samsung Electro-Mechanics reportedly raised MLCC prices by 30% in August, and Taiwan’s Yageo by about 50% in July. The average lead time for high-capacitance MLCCs at US distributor TTI reached 22 weeks in September, six weeks longer than at the start of the year, according to iM Securities. Despite these supply-driven price hikes, the 12-month forward price-to-earnings ratios for MLCC manufacturers like Murata Manufacturing (37.5x) and Samsung Electro-Mechanics (38.8x) remain significantly higher than those of some memory chipmakers.