Memory stocks have plunged following reports that NVIDIA plans to cut production of its Rubin chip line while SK Hynix is slowing its HBM4 memory capacity expansion. The moves signal emerging demand concerns in the high-performance computing market that has been central to the AI infrastructure buildout.

NVIDIA's reported production cuts for Rubin represent a potential shift in the company's supply trajectory, while SK Hynix's decision to ease off its HBM4 ramp suggests memory suppliers are recalibrating their capacity plans. Both developments indicate that demand pressures or supply-demand imbalances may be emerging in the GPU and memory segments that underpin AI training and inference infrastructure.

The market reaction reflects investor concern about the pace of AI infrastructure expansion and whether current production levels and capacity plans will continue to support growth. For the broader AI buildout, any slowdown in advanced chip and memory production could affect the timeline for deploying next-generation AI systems and GPU availability in the data center market.