NVIDIA Cuts Vera CPU Memory Configuration, Highlighting Persistent LPDRAM Supply Constraints and Rising Long-Term Demand, Says TrendForce

NVIDIA has decided to halve the SOCAMM memory capacity of its next-generation Vera Rubin Superchip modules, according to TrendForce’s latest findings. This adjustment does not reflect a reduction in NVIDIA’s overall memory demand. Rather, it is a response to insufficient LPDRAM capacity allocated to NVIDIA under its suppliers’ preliminary 2027 production plans.

Against this backdrop, NVIDIA has opted to lower memory capacity per module while increasing overall module shipments to expand market penetration. The move also underscores the difficulty of closing the LPDDR5X supply gap and points to sustained long-term demand growth.

Projected growth in bit supply is expected to fall short of demand expectations, despite memory suppliers planning capacity expansions for 2027. Based on early allocation plans from Samsung, SK hynix, and Micron, NVIDIA is set to receive enough LPDRAM to meet only about 60% of its estimated needs, with limited room for further increases. As a result, the company recently reduced the SOCAMM memory configuration of its Vera Rubin Superchip platform to produce more Vera CPU units and mitigate the risk of extended supply shortages.

TrendForce emphasizes that this specification change represents a supply-driven architectural compromise rather than a demand-side adjustment. It has no material negative impact on the overall DRAM market balance and should not be interpreted as a reduction in NVIDIA’s total memory requirements.

NVIDIA’s next-generation AI server portfolio will consist of two primary product categories: the Vera Rubin Rack platform and standalone Vera CPU Rack systems. In addition to promoting the full-rack Vera Rubin solution, NVIDIA has clearly indicated its intention to expand deployment of standalone Vera CPU solutions aggressively. This strategy is expected to strengthen the company’s position in the AI inference market and has become one of the key growth drivers of LPDRAM demand.

Demand for LPDRAM explodes; AI servers poised to become the largest end market for LPDRAM

LPDRAM is rapidly expanding beyond smartphones into a growing range of applications, including AI servers and automotive systems, thanks to its superior power efficiency and data transfer performance. ASIC developers beyond NVIDIA are also beginning to evaluate LPDRAM adoption, signaling that the technology is no longer limited to mobile devices. 

However, the supply-demand imbalance is expected to become increasingly severe as AI adoption continues to broaden. TrendForce forecasts that between 2028 and 2030, the AI server ecosystem could emerge as the largest single end market for LPDRAM globally, surpassing smartphones and fundamentally reshaping demand dynamics across the memory industry.