On August 25, SK hynix announced it has developed the industry's first 321-layer 2Tb QLC NAND flash* memory and begun mass production. This breakthrough marks the first time a company has surpassed 300 layers in QLC NAND, setting a new milestone in storage technology.

**NAND flash is categorized as single-level cell (SLC), multi-level cell (MLC), triple-level cell (TLC), QLC, and penta-level cell (PLC) depending on how many data bits can be stored in one cell. As the amount of information storage increases, more data can be stored in the same volume.

The new device offers the highest integration among current NAND products, doubling capacity compared to existing models while enhancing efficiency and performance. To overcome the typical speed limitations of high-capacity NAND, SK hynix expanded its plane* architecture from four to six, boosting parallel processing. As a result, the chip delivers up to 2x faster data transfer speeds, 56% better write performance, 18% higher read speeds, and more than 23% improved power efficiency—making it especially competitive for AI-driven data centers where both speed and energy savings are critical.

**A plane refers to a cell and its peripheral circuitry that can operate independently within a single chip. By increasing the number of planes from 4 to 6, the simultaneous read performance of the chip—a key factor in data processing—is significantly improved.

SK hynix plans to first apply the 321-layer NAND to PC SSDs, followed by enterprise SSDs for data centers and UFS solutions for smartphones. The company also aims to strengthen its AI storage lineup with its unique 32DP* packaging technology, enabling ultra-high-capacity eSSDs built on 32 stacked NAND dies.

**32DP (32 Die Package): A method of simultaneously packaging 32 dies in one package to increase chip capacity.

According to Jeong Woopyo, Head of NAND Development, the start of mass production significantly expands SK hynix's high-capacity storage portfolio while ensuring cost advantages. He added that the company is positioning itself as a "Full Stack AI Memory Provider" to meet the surging demand for AI workloads and data center performance.