[TheDevs]

Growing Demand for Memory and Storage Driven by AI Implementation

The implementation of artificial intelligence (AI) is fueling a surge in demand for memory and storage solutions to support data processing and retention for AI training and inference engines. This demand is benefiting vendors of DRAM, NAND flash, HDD, and magnetic tape, with emerging non-volatile memories also poised to gain traction in end-point AI applications. As a result of this heightened demand and manufacturing cutbacks in memory and storage technology in Fall 2023, prices for solid-state memory and storage products have been on the rise.

At the 2024 Computex in Taiwan, Dinesh Bahal, GM of Micron Consumer and Components Group, highlighted the efforts of Samsung, SK hynix, and Micron in developing high bandwidth memory (HBM) products to meet the increased demand for AI applications. Western Digital (WDC) recently unveiled a six-stage AI Data Cycle framework aimed at optimizing storage infrastructures to enhance AI investments and reduce ownership costs for AI workflows.

The data workflow for AI applications involves a continuous loop of data consumption and generation, encompassing the processing of various data types such as text, images, audio, and video content. Effective and cost-efficient processing and storage of this data necessitate a range of storage and memory products, including HBM for AI training, NAND-based solid-state storage for data flows, and hard disk drives for secondary storage. WDC introduced digital storage products, including a 32TB ePMR Enterprise HDD, to support these elements of the storage supply chain for AI training and inference.