Enterprises are currently navigating significant challenges related to storage, as ongoing shortages, lengthy lead times, and soaring prices may compel them to postpone on-premises AI projects. Research indicates that these storage price increases, which began last year, are expected to continue through 2026, driven primarily by the growing demand associated with AI deployments and escalating data usage.
According to Falko Kuester, an engineering professor at UC San Diego, the demand for storage is continuously expanding. His organization, Open Heritage 3D, collates various forms of data—images, videos, LIDAR scans—and shares them publicly. This accessibility leads to an exponential increase in data, as scans are processed and analyzed, with estimates suggesting data volumes could reach a petabyte within the next 18 months.
The situation isn’t unique to Open Heritage 3D. A December report from Komprise highlights that 40% of companies with over 1,000 employees are currently managing more than 10 petabytes of data. Additionally, 85% of organizations expect their storage expenditures to increase in the coming year, with priorities including cost optimization and data preparation for AI.
Experts forecast a significant rise in storage prices for 2026. For instance, Brad Gastwirth indicates that DRAM and NAND prices could surge by 50% or more due to the constrained supply stemming from heightened AI infrastructure demand. Scott Tease from Lenovo anticipates users today will see prices for essential components rise dramatically, with examples including a 64-Gig DIMM expected to escalate from a low of a couple hundred dollars to nearly $800 in the near term.
The global shortage affects not just DRAM but also other memory types. As manufacturers allocate more capacity to AI server demand, smaller buyers may struggle to secure supplies. The depletion of supplier inventories is evident, and companies like Western Digital confirm that the supply chain dynamics are fundamentally shifting due to unprecedented demand from AI applications.
Amid the shortage, there is a growing interest in quad-level cell (QLC) SSDs, which are predicted to constitute 30% of the enterprise SSD market. Enterprises facing price hikes and delays are advised to maximize the longevity of their existing storage solutions. For instance, reducing unnecessary write cycles to QLC NAND Flash can help extend component life.
Overall, experts suggest that mid-sized enterprises considering new AI clusters may benefit from delaying any major purchases, allowing the market conditions to stabilize.
