DRAM prices to double in Q1 as AI ambitions push memory fabs to their limit (www.theregister.com)

DRAM and NAND flash memory prices are forecasted to surge dramatically in the first quarter of 2026, driven primarily by the escalating demands of AI-focused hyperscalers and cloud service providers. TrendForce has revised its earlier estimates, predicting a staggering rise in DRAM prices by 90–95% quarter-over-quarter and a 55–60% increase for NAND flash. This spike is attributed not only to AI's voracious appetite for data processing and storage but also to unexpected growth in PC shipments at the end of 2025, which further strained already tight supply chains. The implications for the AI/ML community are significant, as the transition from AI training to inference demands more memory and storage at an unprecedented scale. Inference applications require high-performance memory, such as LPDDR, to efficiently handle tasks and maintain the model's key-value cache, essential for enhancing user interaction in multi-session environments. As economic pressures mount due to rising infrastructure costs, organizations may face budget constraints while trying to upgrade their systems to meet the growing computational needs of AI applications. With new fabrication facilities requiring years to establish, relief from the ongoing memory shortage seems a distant prospect.