Samsung Electro-Mechanics Secures NVIDIA AI Semiconductor Package Substrates; Sejong and Busan Plants to Be Expanded
TECHWORLD ·
✦ AI Summary
Samsung Electro-Mechanics and NVIDIA announced a strategic business partnership on the 21st to expand their next-generation AI semiconductor package substrate business.
Samsung Electro-Mechanics will newly build dedicated production lines for AI semiconductor package substrates at its Sejong and Busan plants and plans to supply NVIDIA with substrates for compute chipsets and network switch package substrates.
Samsung Electro-Mechanics forecast that package substrate shipments to NVIDIA will grow by double digits on average annually starting this year, and that by 2028 they will account for a significant portion of its total package substrate production capacity.
Samsung Electro-Mechanics and NVIDIA have launched cooperation to expand their business in next-generation AI semiconductor package substrates. The two companies announced a strategic business partnership on the 21st, saying the partnership aims to strengthen the global AI semiconductor supply chain and accelerate the expansion of next-generation AI infrastructure.
Accordingly, Samsung Electro-Mechanics will newly build dedicated production lines for AI semiconductor package substrates at its Sejong and Busan plants. Sejong and Busan are Samsung Electro-Mechanics' key domestic production bases.
Samsung Electro-Mechanics plans to supply NVIDIA with substrates for compute chipsets, as well as package substrates for network switches used in ultra-high-speed chip-to-chip connections inside data centers. Through this, the company plans to expand its supply of substrates for NVIDIA.
Samsung Electro-Mechanics forecast that the volume of package substrates supplied to NVIDIA will grow by double digits on average annually starting this year. It also expects that by 2028, NVIDIA-related shipments will account for a significant portion of its total package substrate production capacity.
The partnership carries the meaning of securing both advanced packaging technologies and supply capabilities to support the performance of AI semiconductors. Through this, NVIDIA will secure a stable supply base for cutting-edge package substrates, while Samsung Electro-Mechanics will be able to expand its business in the high-value-added market for AI semiconductor substrates.
As the computing performance of AI semiconductors becomes more advanced, package substrates are facing greater demand for advanced manufacturing technologies. As substrate area expands and the number of layers increases, it is important to implement ultra-fine circuits to reduce signal loss, and flatness maintenance technology to suppress warpage caused by thermal deformation is also important. At the same time, process difficulty is rising, and manufacturing capabilities that can secure stable yields are regarded as a key competitive advantage.
Power supply stability is also a major technical challenge. There is a need to control power noise generated during high-speed computation and improve power efficiency, and to that end, passive components such as MLCC and silicon capacitors are directly embedded inside the substrate. This method of directly embedding passive components involves a highly difficult process.
Samsung Electro-Mechanics plans to respond to the technical requirements for package substrates for next-generation AI semiconductors. To this end, it will build on its advanced FC-BGA manufacturing technology and high-reliability MLCC technology, while combining capabilities in passive component technologies, including silicon capacitors. The company plans to use this to strengthen its responsiveness to AI semiconductor package substrates and its competitiveness in high-performance substrate solutions.
Samsung Electro-Mechanics President Jang Deok-hyun emphasized the significance of the company joining NVIDIA as a key infrastructure partner. He said the company plans to combine its advanced FC-BGA manufacturing competitiveness with high-reliability MLCC technology, and expressed its commitment to supporting the highest levels of performance, quality, and on-time supply for NVIDIA's AI platform. He also said the company plans to serve as a total solution provider.
Debora Shoquist, NVIDIA's executive vice president in charge of operations, said global production capacity must be expanded to keep pace with the speed of AI infrastructure buildouts. She explained that cooperation with Samsung Electro-Mechanics will expand the network of substrate supply partners and secure additional production capacity. She added that the company will expand production scale and strengthen supply of accelerated computing platforms to match customer demand.
Source: TECHWORLD · Park Gyu-chan
Original: https://www.epnc.co.kr/news/articleView.html?idxno=408574
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.