DRAM is a volatile memory with a 1T1C cell structure, playing a core role in computing by enabling high-speed and large-capacity data input/output in close proximity to CPUs and GPUs. Although its speed is about one order slower than SRAM and it is several times more expensive than NAND, it possesses capacity levels of several GBs and speed characteristics of several Gbps, while maintaining appropriate power and cost competitiveness. Consequently, its application areas are expanding not only in product groups such as computing, mobile, graphics, and consumer electronics but also in the AI field.

The DRAM business model has been sustained for decades through the scaling of cell size and chip size. From a design perspective, improvements in interface performance and innovations in architecture technology have driven product diversification. Furthermore, HBM products have emerged as the core memory of the AI era as a result of applying advanced process technologies and post-process technologies (e.g., TSV, wafer bonding).

Recently, to overcome the limitations of current tech nodes, various tech platform options such as the vertical gate cell scheme, chiplets concept, and 3D cell stacking structure are being proposed.

NAND has high-capacity storage as its key characteristic; although it is slower than DRAM, it plays a core role in the storage device market due to its non-volatility, low cost, and high density. In particular, it has established itself as an essential storage device for mobile devices, PCs, and data centers, and in the era of AI and big data, it continues to expand its application areas through high-speed interfaces and low power technologies.

The NAND business model has maintained sustainability for decades through cell density and 3D stacking technologies. Enhancements in interface performance and innovations in architecture (e.g., NVMe, PCIe 5.0) have led to product diversification, while the securing of high speed and high density through advanced process technologies (e.g., Charge Trap, PUC) has positioned NAND as a core storage memory for the AI era.

For 3D NAND technology, research on new structures and materials is underway to overcome physical stacking limits, and next-generation cell and process technologies such as hybrid bonding and FeNAND are also being developed in parallel for high-performance product development.