AI Boom Creates Two-Tier Market for Ceramic Capacitors, Squeezing High-End Supply

A sharp divergence is splitting the market for multilayer ceramic capacitors (MLCCs), essential components in virtually all electronics. Spurred by the artificial intelligence boom, prices for high-end, high-capacitance MLCCs used in AI servers are reportedly skyrocketing, with some distributors telling Chinese media outlet Times Weekly that prices are changing “hour by hour.” In contrast, prices for common MLCCs used in consumer electronics like smartphones and PCs are beginning to fall.

AI Strong, Consumer Weak

This growing divide reflects an “AI strong, consumer weak” demand pattern, according to TrendForce analyst Chen Weishang. The insatiable appetite for AI hardware has created a supply crunch for the most advanced components. Lead times for some high-end X6S MLCC products have reportedly stretched from eight weeks to as long as 20 weeks, with some high-end AI-related orders in China now booked into the second quarter of 2027.

In response to the tight supply, major global manufacturers have begun raising prices. In July, industry leaders like Murata, Yageo, and Samsung Electro-Mechanics reportedly increased prices on some products by 30% or more. This trend, however, is not universal, as the oversupply in the consumer electronics segment has led to price declines for lower-specification components.

Servers Drive Unprecedented Demand

AI servers are the primary catalyst for the high-end MLCC shortage. These systems require significantly more, and higher-value, capacitors than traditional servers to ensure power stability and signal integrity for high-performance GPUs.

• According to Frost & Sullivan, an AI server uses 8 to 12 times more MLCCs than a conventional server.
• An 8-GPU AI training server requires approximately 48,000 MLCCs.
• An Nvidia GB300 NVL72 rack uses about 440,000 MLCCs, with the next-generation Rubin VR200 NVL72 platform expected to use around 600,000.

Ad

This volume increase translates to a massive jump in value. The MLCCs in a single GB300 NVL72 rack are valued at approximately $1,530, a figure projected to rise to $4,320 for the VR200. Murata, the market leader, anticipates that the per-unit MLCC demand for AI servers will grow at an average annual rate of over 30% for the next five years.

The Technology Gap

While the demand surge presents an opportunity for Chinese manufacturers like Fenghua Advanced Technology and Chaozhou Three-Circle (CCTC), a significant technology gap remains. Lu Jing, a partner at Frost & Sullivan, told Times Weekly that unlike standardized products like DRAM, the MLCC market is highly fragmented, making it harder for new entrants to quickly gain market share. Progress is expected to be gradual, achieved one component at a time.

Japanese and South Korean firms—primarily Murata, Samsung Electro-Mechanics, and Taiyo Yuden—dominate the most advanced segments. Murata's technical leadership is particularly strong; a Northeast Securities report notes that Nvidia's GB200/GB300 GPUs embed tiny 01005-size MLCCs directly into the substrate, a process only Murata can currently achieve at scale. According to Times Weekly, Murata executives have previously estimated they hold a 10-year technology lead over Chinese competitors in the high-end space.

While Chinese firms reported strong revenue and profit growth for the first half of 2026, their high-end offerings are still in the process of customer certification for top-tier AI applications. Their immediate impact is more likely to be increased price competition in the mid-range market rather than alleviating the supply bottleneck for the most cutting-edge components.