Login
Sign Up
Data compiled by Woofun AI shows that consumer-grade LPDDR5X memory prices surged 89% in the second quarter, marking a significant escalation in the AI memory squeeze effect across smartphone and PC markets. Market research firm SigmaIntell disclosed that supply-demand imbalances drove broad-based price increases for consumer storage products, with LPDDR4X 4GB rising 75% month-on-month and LPDDR5X 12GB climbing 89%. This volatility is primarily driven by the adoption of LPDDR memory in next-generation server GPU platforms, such as NVIDIA's Vera Rubin, which has intensified competition for limited supply.
Manufacturers have expanded these price hikes to include consumer-grade memory, with some customers already accepting the new pricing structures. Supply-side pressure stems from capacity reallocation, where producers prioritized high-value-added products like HBM, server DRAM, and enterprise SSDs, thereby tightening the availability of consumer-grade memory. Consequently, smartphone and PC brands are adjusting their memory orders.
This shift suggests that DRAM price increases may slow in the second half of the year, as demand for low-end terminals cools more significantly than mid-to-high-end products.
Meanwhile, semiconductor wafer resources remain concentrated on HBM, compressing consumer-grade DRAM capacity, while general server DRAM prices continue to rise due to cloud providers replenishing inventory, keeping channel inventories at two to three weeks. In the NAND market, SSD prices rose approximately 50% month-on-month in Q2, driven by strong demand for enterprise NAND and AI cluster expansion. UFS prices increased up to 100% due to enterprise SSD demand squeezing supply, while reduced supply of low-capacity eMMC further tightened consumer-grade NAND availability. Overall, AI infrastructure expansion is altering storage industry price transmission paths, with shortages previously focused on HBM and server DRAM now impacting mobile, PC, and consumer storage markets through capacity allocation and wafer resource constraints.