While the The booming AI market has propelled Nvidia to become the second largest semiconductor supplier behind Intel., Samsung has fallen to third among suppliers, partly due to falling memory demand over the past year. Although memory is expected to rebound in 2024, Samsung also hopes that a stronger local presence will help the South Korea-based semiconductor supplier regain some of its lost momentum.
To that end, Samsung recently received $6.4 billion in CHIPS Act funding, and the company is directing those funds toward strengthening its North American manufacturing presence. Far from being a stranger on American soil, the company is expanding its manufacturing facility in Taylor, Texas, which has been in operation since 1996. The additional investment via the CHIPS and Science Act means Samsung will invest more than $40 billion of dollars in the region in the coming years as one of the largest foreign direct investments in a new project in United States history. Samsung expects the larger factory to produce chips for the automotive, consumer technology, IoT, aerospace and other industries.
Samsung’s reshoring efforts are keeping pace with those of US-based manufacturers such as Intel and Micronas well as TSMC foundry based in Taiwan.
Samsung hopes strengthening its U.S. supply chain can restore some of the momentum the company lost during the latest electronics industry downturn, which was particularly brutal for the company’s memory business. ‘business. Like other memory suppliers, Samsung has been hit by slowing demand and falling prices over the past year.
Improve memory
In response, Samsung also channeled more R&D into advanced memory for AI applications.
Samsung is preparing new 3D structures for DRAMs smaller than 10 nanometers (nm), enabling larger single-chip capacities that could exceed 100 gigabits (GB). After mass production of 12nm-class DRAM in May 2023, Samsung is working on its next-generation 11nm-class DRAM, which is expected to offer the highest density in the industry.
Additionally, Samsung is working to reduce cell sizes and refine channel hole etching techniques in its quest to develop 1000-layer vertical NAND (V-NAND). Development is on track for Samsung’s ninth generation of V-NAND to deliver the industry’s highest layer count based on a dual-stack structure.
Samsung also introduced its next-generation HBM3E DRAM, named Shinebolt. The HBM3E offers an impressive speed of 9.8 gigabits per second (Gbps) per pin, meaning it can achieve transfer rates of up to over 1.2 terabytes per second (TBps).
AI and mobile
Given Samsung’s history in mobile products, it’s no surprise that the company is looking to the mobile sector for future AI products. Samsung recently announced its collaboration with Arm on a next-generation Cortex™-X processor developed on Samsung Foundry’s latest Gate-All-Around (GAA) process technology. The two companies plan to reinvent 2 nanometer (nm) GAA for custom silicon in next-generation data centers and infrastructure and develop an AI chipset solution for future artificial intelligence (AI) mobile computing applications ) generative.