According to the latest research by TrendForce, the massive amount of data generated by AI is putting pressure on global data center storage facilities. The Nearline HDD (nearline hard drive), which has traditionally been the cornerstone for storing massive amounts of data, is now facing a supply shortage. This has led to the SSD (solid-state drive) becoming the focus of the market, especially the large-capacity QLC SSD, whose shipments may experience an explosive growth in 2026. 

In the traditional data center storage hierarchical architecture, HDDs, due to their extremely low cost per unit of storage capacity (GB), have consistently held the dominant position as the storage solution for "cold data" (data that is rarely accessed but needs to be archived for long-term preservation). Cold data includes backup files, historical records, and other data that is not frequently accessed but needs to be archived for long-term storage. As the application of Inference AI (AI inference) expands, the demand for cold data storage has also rapidly increased. 

SSD, with its high-speed reading and writing performance, mainly handles the "hot data" and "warm data" that need frequent access. When comparing QLC SSD and Nearline HDD, the former not only has better performance but also can save approximately 30% of power consumption. 

TrendForce, a market research firm, stated that due to the fact that major HDD manufacturers worldwide have not planned to expand their production lines in recent years, they were unable to promptly meet the sudden and massive storage demands stimulated by AI. Currently, the delivery time for NL HDDs has significantly extended from the original several weeks to over 52 weeks, further widening the storage gap for CSPs. 

In North America, CSPs have long planned to expand the adoption of SSDs for warm data applications. However, due to the severe shortage of HDDs in this wave, CSPs have even begun to consider using SSDs for cold data. Nevertheless, to achieve large-scale deployment, they must first overcome the dual challenges of cost and supply chain.

TrendForce (a market research firm) states that if CSPs want to introduce QLC SSDs for cold data storage, they need to consider the modification of data management algorithms, the adaptation of the software stack, and the calculation of the total cost of ownership (TCO). It is necessary to adhere to the price bottom line to achieve cost balance. For SSD suppliers, although this wave of order transfer is an excellent opportunity to improve the profit structure, due to the limited production capacity of high-capacity products, suppliers will not be willing to significantly reduce prices. Therefore, it is expected that there will be a price game between buyers and sellers, driving the overall Enterprise SSD contract price to increase by 5-10% in the fourth quarter of 2025.