A view of the Samsung Electro-Mechanics Suwon plant./Courtesy of Samsung Electro-Mechanics

Samsung Electro-Mechanics is accelerating the expansion of its high-value semiconductor substrate business as artificial intelligence (AI) spreads. The strategy centers on FC-BGA (flip-chip ball grid array), a key component for servers and data centers, pursuing both capacity expansion and technology upgrades.

According to the industry on the 3rd, Samsung Electro-Mechanics executed more than 1 trillion won in capital expenditures (CAPEX) last year and is expected to make a similar-sized investment this year. The plan is to shift the center of gravity to semiconductor substrates. Demand for high-performance semiconductor substrates is structurally increasing with the expansion of Generative AI and high-performance computing (HPC).

Samsung Electro-Mechanics' FC-BGA production line for servers and data centers is virtually running at full capacity. Some supplementary investments are underway, but supply is failing to keep up with demand. President Chang Duck-hyun of Samsung Electro-Mechanics said at the regular shareholders' meeting in Mar. 2026 last month, "Demand for the product exceeds production capacity by more than 50%," adding, "We are pursuing both line enhancements and plant expansion." Samsung Electro-Mechanics is pushing to expand centered on its Busan and Sejong business sites.

FC-BGA is a key component that connects high-performance semiconductors and substrates, requiring high-speed signal processing and heat dissipation control. As AI server semiconductors see sharp increases in data throughput and power consumption, they require higher layer counts and precision than before, making it a field with high technological barriers to entry.

In step with this, Samsung Electro-Mechanics' investment direction is also changing. Moving away from general-purpose components such as MLCCs (multilayer ceramic capacitors), the scale of investment in package solutions has expanded significantly since 2022, making an FC-BGA-centered investment stance clear. This is seen as a strategy to shift the business axis to a core infrastructure area that determines the performance of AI Semiconductor.

The supply-demand imbalance for advanced semiconductor substrates is being reflected in prices. Samsung Electro-Mechanics recently raised the selling prices of some FC-BGA product lines. Rising materials and supplies costs and supply shortages have combined to strengthen its pricing power.

Mirae Asset Securities said, "The (Samsung Electro-Mechanics) substrate is already sold out, and the next step is price increases," raising its FC-BGA price estimate by about 10%. The industry interprets the list price hike as a signal of entering a full-fledged phase of profitability improvement.

MLCCs are showing a similar trend. As demand for high-spec products for AI servers increases, supply is getting tight and prices continue to rise. High-capacity, high-reliability MLCCs are key components that maintain the power stability of AI servers, raising expectations for profitability improvement on rising demand.

Changes in the business portfolio are also accelerating. Samsung Electro-Mechanics is expanding its business areas from a traditional smartphone and PC-centered structure to AI servers, automotive electronics, and aerospace. In automotive electronics, it is expanding camera module production capacity in response to the spread of Autonomous Driving, and it is also pushing to expand its Mexico plant to respond to the North American market.

Securities houses expect Samsung Electro-Mechanics to continue its earnings improvement on the back of expanding AI demand. There are also forecasts that operating profit for the year will exceed 1 trillion won.

However, external variables such as geopolitical risks in the Middle East are cited as burdens. Given its business structure with a high export ratio, it could be affected by logistics costs and oil price fluctuations. Even so, the prevailing view is that increased investment in AI data centers and growth in the automotive electronics industry will offset a large part of this.

※ This article has been translated by AI. Share your feedback here.