Abstract

The power demand of AI servers and GPU modules is rising in line with their performance. In response to this, there is a growing movement to replace Aluminum Polymer Capacitors with MLCCs. With the expansion of the High CV mid-to-large MLCC lineup, MLCCs can now replace areas which were previously covered by aluminum polymer capacitors. Capable of implementing the required capacitance within lower heights and smaller areas, they hold an advantage in space saving. Particularly, considering their transient performance due to their low ESR features, high-frequency ripple behavior, design flexibility, and component supply stability, the trend toward MLCC-centric architectures is becoming increasingly clear across the overall AI server market.

1) Space Saving Effect

MLCCs are of a smaller size and lower height in comparison to Aluminum Polymer Capacitors. Using multiple MLCCs within the same area can lead to higher capacitance, improving space utilization efficiency greatly.

※ The figure above is an example and the layout can be modified according to the customer's design requirements.

2) Capable to respond to High-Speed Transients

The capability to correct voltage immediately is crucial, since AI GPUs go through very large, instantaneous power fluctuations. MLCCs have a very low internal ESR (equivalent series resistance) and ESL (equivalent series inductance), minimizing voltage drop and enabling a rapid recovery, thus providing a faster and more stable voltage response than Aluminum Polymer Capacitors. They are also suited for AI server power architectures, since they maintain a low impedance in the high-frequency band (1–2 MHz).

3) Supply-Chain Stability

Aluminum Polymer Capacitors have limited suppliers and a high price volatility. MLCCs, on the other hand, have a lower long term risk because they have multiple global suppliers.

MLCCs meet the high-speed and high-frequency power demands of AI servers. They also provide superior performance and design flexibility in comparison to Aluminum Polymer Capacitors. With space saving advantages and stable supply chains, MLCC-centric architectures will continue to establish themselves as the core trend in AI servers.

[ Product Lineup ]