Samsung Electro-Mechanics Aims to Become Key Supplier of AI Components with Silicon Capacitors

by JINYOUNG PARK Posted : June 14, 2026, 09:03Updated : June 14, 2026, 09:03
Kim Won-ki, head of Samsung Electro-Mechanics' silicon capacitor development group, presents at the company's product seminar on silicon capacitors on June 11.
Kim Won-ki, head of Samsung Electro-Mechanics' silicon capacitor development group, presents at the company's product seminar on silicon capacitors on June 11. [Photo: Samsung Electro-Mechanics]

Samsung Electro-Mechanics is accelerating its entry into the artificial intelligence (AI) infrastructure market with its silicon capacitors, which are essential components for AI data centers. Following the establishment of mass production systems for multilayer ceramic capacitors (MLCC) and semiconductor package substrates, the company has completed its portfolio of key components needed for AI semiconductors.

On June 14, Samsung Electro-Mechanics announced that it has signed a supply contract worth approximately 1.5 trillion won with a major global client.

Silicon capacitors are critical devices that enhance power stability and maximize component integration in high-speed and high-density electronic devices, including AI, high-performance computing (HPC), and optical communications.

The competitiveness of AI semiconductors has increasingly relied on power efficiency and stability. As the power consumption of GPUs and high-bandwidth memory (HBM) used in AI data centers has surged, reliable power supply technology has become a key competitive advantage. AI server semiconductors experience significant instantaneous power fluctuations, making components that supply power and eliminate noise close to the semiconductor chips increasingly important.

Silicon capacitors have emerged as next-generation power stabilization components to meet this demand. They not only store electricity for immediate supply but also eliminate minute electrical signal noise, ensuring stable operation of semiconductors. While traditional MLCCs performed power stabilization externally, silicon capacitors can be directly integrated within semiconductor packages, allowing for faster and more efficient power control.

The value of silicon capacitors is rising in line with the trend toward high integration in AI semiconductors. Made from silicon wafers, they are thin, space-efficient, and maintain excellent characteristics even in high-frequency environments.

Compared to MLCCs, silicon capacitors have equivalent series inductance (ESL) that is over 100 times lower, minimizing power loss and noise. This is why they are considered optimized components for high-performance semiconductors such as GPUs and AI accelerators.

Samsung Electro-Mechanics has enhanced its competitiveness by incorporating memory semiconductor process technology. The company applied the ISC technology used in DRAM production to silicon capacitors, maximizing surface area through deep etching of silicon wafers and forming dielectrics and electrodes internally to achieve high capacitance in a small area. This differentiated technology is crucial as AI semiconductor packages become increasingly complex and space constraints grow.

Quality competitiveness is also highlighted as a major strength. Samsung Electro-Mechanics has established its own testing facilities capable of assessing performance at the individual product level, based on quality management know-how accumulated through its MLCC business. Given that semiconductors for AI data centers must operate reliably over long periods, Samsung Electro-Mechanics emphasizes this reliability as a differentiating factor.

With this, Samsung Electro-Mechanics has established its capability to supply all solutions for key components in AI data centers. The company is recognized as the only supplier capable of providing both MLCCs and package substrates. For semiconductor clients, this means they can procure essential components in an integrated manner without going through multiple suppliers, offering significant advantages in design optimization and supply chain stability.

As investments in AI infrastructure expand, the importance of power stabilization solutions is also increasing. The silicon capacitor market is expected to grow at an annual average rate (CAGR) of 18% from 2026 to 2031. With the advancement of AI semiconductor packages, the application scope is anticipated to broaden into various fields, including AI servers, autonomous driving, and high-performance computing.

Samsung Electro-Mechanics plans to leverage its recent large supply contract to expand its global customer base. The company aims to enhance its lineup of high-capacity, multifunctional silicon capacitors and increase market share through diversification of its global clientele.




* This article has been translated by AI.