Last year, memory chips were the most sought-after item at Huaqiangbei, China's largest electronics market located in Shenzhen. However, this year, particularly since April, attention has shifted to a smaller component: multilayer ceramic capacitors (MLCCs).
According to SCMP, suppliers and dealers are "enjoying a surge in MLCC prices" due to the global AI boom. This tiny component is made up of multiple layers of dielectric materials, including ceramics and alternating metal electrodes, which are then sintered into a compact block. It serves to store energy, filter noise, and stabilize voltage on the circuit board.
According to Arrow DXP, MLCCs are among the most critical passive components on AI chips, GPUs, and AI server boards. Their primary role is to stabilize the power supply to the processor. During AI training or inference, the GPU can experience extremely rapid changes in power consumption, creating current spikes. MLCCs, positioned close to the GPU, CPU, and high-performance memory (HBM), act as noise suppressors, providing instantaneous current and reducing voltage drops, ensuring stable chip operation even under fluctuating computational load.
In addition to power stabilization, MLCCs also have the task of filtering high-frequency noise in the system. Modern AI servers use many high-speed interfaces such as PCIe Gen5, NVLink, CXL, and connect to HBM memory, generating a large amount of electromagnetic noise and signal fluctuations. Thanks to their fast response characteristics in the high-frequency range, capacitors help eliminate noise on the power and signal lines.
According to ALF Achemic, a next-generation AI server typically uses up to 28,000 MLCC capacitors, 10-15 times more than a conventional server, making this component increasingly sought after.
"I haven't seen any signs that MLCC prices are about to cool down," a distributor surnamed Wu, who supplies Murata Manufacturing ( Japan ) brand capacitors to traders and secondary suppliers in China, told the SCMP. He added that in the spot market, the average price of capacitors has risen from 10 yuan ($1.48) to 40 yuan ($5.9) per 1,000 units, a fourfold increase.
Some capacitor models have even seen price increases. According to a trader surnamed Zheng, who sells MLCCs from Samsung Electro-Mechanics (Semco) and domestic companies like Chaozhou Three-Circle Group, the price of a 0805 22μF capacitor has increased more than 20 times since the beginning of the year. Another trader added that frequent shortages, with production failing to keep up with demand since April, have made his business increasingly difficult.
The global MLCC market is currently dominated by five major companies: Murata Manufacturing, Samsung Electro-Mechanics, Taiyo Yuden, TDK, and Kyocera AVX. Of these, Murata of Japan is the world's largest manufacturer, supplying MLCCs for consumer electronics, electric vehicles, and AI servers. Samsung Electro-Mechanics ranks second, offering high-capacity MLCCs for AI GPUs, HBM memory, and data centers. The other two Japanese companies, Taiyo Yuden and TDK, also play a significant role thanks to their strengths in high-performance, ultra-small MLCCs. Kyocera AVX is a major supplier of capacitors for industrial, automotive, and specialized electronic infrastructure applications.
What's happening at the electronics market in Huaqiangbei reflects what investment bank Goldman Sachs calls "the biggest and longest-lasting AI growth cycle in history." Specifically for capacitors, leading manufacturers from Japan and South Korea are receiving a surge in orders, raising concerns about a serious and prolonged supply crisis.
Last week, market research firm TrendForce reported that the capacitor industry is facing a potential supply shortage in the second half of this year due to ongoing design adjustments by next-generation AI platforms during final validation stages, significantly increasing the number of high-end MLCCs required per motherboard. For example, the number of MLCCs in one type of MLCC in the AMD MI450 AI chip, which specializes in training and inferring large language models, increased by 632%, from 1,440 units per motherboard to 10,544 units, after engineers replaced all electrolytic aluminum and tantalum capacitors.
According to TrendForce, MLCC demand will reach a "new peak" at the end of the year, as Google, Amazon Web Services, and Meta ramp up production of their own custom AI chips. Typically, capacitor manufacturers prioritize contracts with these companies, making the spot market susceptible to price increases and serious delivery delays.
( According to vnexpress.net )
Source: https://baodongthap.vn/con-sot-tu-dien-mlcc-tai-trung-quoc-a242512.html
