Search
NVIDIA Reportedly Cuts Memory Configuration for Vera Rubin AI Cabinets to Control Costs
16 hour ago /
Read about 0 minute
Author：小编
Affected by high memory prices and ongoing supply chain tensions, NVIDIA is planning to significantly reduce the memory configuration of its next-generation Vera Rubin NVL72 cabinet-level AI systems. The originally designed 192GB SOCAMM solution may be reduced to 96GB, a nearly 50% decrease. This adjustment aims to lower costs and enhance market competitiveness, particularly among small and medium-sized enterprises and in emerging markets. However, the memory reduction may raise concerns about system performance, as memory capacity is directly related to processing speed and parallel computing capabilities. Nevertheless, NVIDIA may still compensate for the performance shortfall by optimizing software and hardware integration.
-
C114 Communication Network -
Communication Home
7 X 24 Track global technological trends
