SK hynix said at a conference call for fourth-quarter 2025 earnings on the 29th that NAND flash is undergoing a structural shift beyond its past role as a simple storage device to a storage solution that directly supports AI compute flows.
The company said, "As AI inference technology advances, it has become difficult to meet customer needs with the existing GPU-, CPU-, and memory-centric architecture alone," and noted, "To quickly input and output AI data and improve compute efficiency, there is a growing need for high-performance storage with enhanced caching capabilities." Accordingly, demand is rapidly increasing for high-performance enterprise SSDs (eSSD) that support high-speed input and output of AI data.
A particularly notable change is a fundamental shift in AI server architecture. SK hynix said, "Storage is establishing itself as a core component beyond a supporting role in GPU-centric compute pipelines," and added, "Storage is being redefined as a device that determines AI compute performance and power efficiency."
In response, the company is preparing next-generation storage solutions such as ultra-high-performance eSSDs and high-end SSDs (high SSDs), alongside development of existing NAND products. The strategy is to preemptively secure future technologies to strengthen competitiveness.
In addition, along with developing ultra-high-capacity SSDs to overcome power and space limits in data centers, the company is accelerating the advancement of high-bandwidth flash (HBF) technology. SK hynix said, "In response to the expansion of AI infrastructure, we will continue comprehensive technological responses across NAND storage."