INDEPENDENT JOURNAL

A clearer view of what matters

An independent news site covering AI, technology, economy, society, science, health, sports, culture and world affairs.

AI · 2026-10-22

Samsung Electro-Mechanics and NVIDIA Forge Strategic Partnership… Expanding AI Semiconductor Package Substrate Lines in Sejong and Busan

Samsung Electro-Mechanics and NVIDIA have formed a strategic partnership to supply package substrates for AI semiconductors and ultra-high-speed networks. The two companies will expand dedicated production lines at the Sejong and Busan plants, and Samsung Electro-Mechanics plans to supply related substrates for multiple years starting this year.

Samsung Electro-Mechanics and NVIDIA Forge Strategic Partnership… Expanding AI Semiconductor Package Substrate Lines in Sejong and Busan
AI-generated editorial illustration
As AI data centers grow, not only computing chips but also the components that connect those chips and help them operate reliably become increasingly important. That is the backdrop to the strategic partnership between Samsung Electro-Mechanics and NVIDIA, which will jointly expand domestic production capacity for package substrates.

According to Chunjilbo, the two companies will make a joint investment to expand dedicated production lines for AI semiconductor package substrates at Samsung Electro-Mechanics’ Sejong and Busan plants. Starting this year, Samsung Electro-Mechanics plans to supply NVIDIA with next-generation AI semiconductor and ultra-high-speed network package substrates over multiple years. The supply items include substrates used in AI compute chipsets as well as substrates for network switches that handle high-speed data transfer between chips in data centers.

A package substrate is a component that electrically connects semiconductor chips, the substrate, and external circuits while also supporting the chip. As AI semiconductors become more powerful and chip sizes and stacking levels increase, substrates are required to deliver denser circuitry and more stable signal transmission. It is also important to suppress warping caused by heat during manufacturing and maintain flatness.

Power management is another major challenge. To reduce power noise generated during high-speed computation and improve efficiency, components such as multilayer ceramic capacitors (MLCCs) and silicon capacitors are sometimes embedded within the substrate. Samsung Electro-Mechanics plans to combine its package substrate manufacturing technology with passive component technology to meet NVIDIA’s next-generation substrate requirements.

Samsung Electro-Mechanics expects shipments to NVIDIA to increase by more than double digits on an average annual basis over the coming years, starting this year. It also projects that in 2028, NVIDIA-related shipments will account for a significant share of Samsung Electro-Mechanics’ overall package substrate production capacity. This is the company’s outlook, and the actual supply volume and degree to which production capacity is reflected remain to be confirmed in the future.

This cooperation is significant because it allows NVIDIA to secure additional advanced substrate supply chains and production capacity, while Samsung Electro-Mechanics expands its high-value-added package substrate business for AI semiconductors. If the joint investment-led expansion of production lines proceeds as planned and leads to long-term supply, it could help broaden the component supply base needed for the expansion of AI infrastructure. However, the specific investment amount and the scale of the line expansion were not provided in the report.

Related search terms

#Samsung Electro-Mechanics NVIDIA partnership#AI semiconductor package substrates#Sejong Busan production lines#NVIDIA supply deal#high-speed network substrates#MLCC silicon capacitor substrate#AI data center components
← Back to articles