Samsung Electro-Mechanics, LG Innotek projected to top 1 trillion won in operating profit

Samsung Electro-Mechanics and LG Innotek are emerging as unexpected winners in the AI era as they rapidly strengthen their presence in the supply chain for advanced semiconductor substrates used in GPUs and AI servers.

Both companies are expected to post record earnings this year, according to the securities industry on Wednesday.

Daishin Securities projected Samsung Electro-Mechanics’ operating profit at around 1.5 trillion won ($1 billion), up 26.1 percent from last year.

Hana Securities estimated LG Innotek’s operating profit at around 1.1 trillion won. If the projections hold true, it would mark LG Innotek’s return to the 1 trillion won level of operating profit for the first time in four years since 2022.

The two companies’ stock prices have also surged sharply. Samsung Electro-Mechanics and LG Innotek closed at 1.63 million won and 1.044 million won, respectively, on Wednesday, nearly doubling in less than a month.

The rally has been driven by soaring demand for flip-chip ball grid array (FC-BGA) substrates used in AI servers. FC-BGA is a key component that connects semiconductor chips to mainboards, serving as a stable bridge between increasingly AI chips, requiring thousands of fine circuits to be densely connected, and mainboards that have relatively wider wiring and simpler structures.

High-performance AI chips, such as Nvidia GPUs, in particular, require multilayer and high-density substrate technologies to process massive amounts of data at ultrahigh speeds. However, only a limited number of companies are capable of producing such substrates reliably and at scale without defects.

Analysts say worsening supply shortages of FC-BGA are rapidly increasing the corporate value of Samsung Electro-Mechanics and LG Innotek, which are among the few suppliers capable of meeting demand.

Samsung Electro-Mechanics is also drawing attention for its competitiveness in multilayer ceramic capacitors (MLCC).

MLCCs act like mini electricity dams by storing power and supplying it instantly when needed, reducing voltage fluctuations. They prevent malfunctions by helping maintain stable power when semiconductors draw large amounts of electricity simultaneously. The components are used in nearly all electronic devices, including smartphones, computers, automobiles and servers.

AI servers and high-performance GPUs require even more high-specification MLCCs to handle more intense power consumption and heat generation. While ordinary smartphones contain around 1,000 MLCCs, AI servers and electric vehicles use tens of thousands.

Both Samsung Electro-Mechanics and LG Innotek are also accelerating the development of glass substrates, often referred to as the dream substrate of next-generation semiconductors.

Compared to existing FC-BGA substrates, glass substrates are less prone to warping and can reduce signal loss by more than 20 to 30 percent, making them a core technology for next-generation AI semiconductors.

Samsung Electro-Mechanics has established a pilot line at its Sejong plant and begun producing prototypes. The company also plans to begin mass production as early as 2027 through a joint venture with Japan’s Sumitomo Chemical Group.

LG Innotek is developing and testing production systems centered at its Magok research and development center in Gangseo District, western Seoul, and its Gumi plant in North Gyeongsang. It aims for mass production as early as 2028.

The company is also seeking global clients by participating for the first time in the Electronic Components and Technology Conference 2026, the world’s largest semiconductor packaging conference, held in Orlando, Florida, from Tuesday to Friday. There, it is showcasing its next-generation substrate technologies.

“Both companies are highly likely to be re-evaluated not simply as parts suppliers but as key infrastructure partners as the AI and humanoid robot industries expand,” said Kim Dong-won, a researcher at KB Securities.

This article was originally written in Korean and translated by a bilingual reporter with the help of generative AI tools. It was then edited by a native English-speaking editor. All AI-assisted translations are reviewed and refined by our newsroom.

BY LEE YOUNG-KEUN [[email protected]]