NVIDIA drastically cuts new rack memory configuration, leading to a collective drop in memory stocks? SemiAnalysis founder clarifies: too clickbait-y

Odaily Odaily News According to market sources tonight, SemiAnalysis' latest research report points out that "NVIDIA's next-generation AI server cluster, Rubin NVL72, has made significant adjustments to its memory configuration. To cope with supply chain tightness and ensure timely delivery of the Rubin cabinets, the single-cabinet capacity has been drastically reduced from the originally planned 55TB to 28TB, a drop of about 50%, using a scaled-down version of 96GB SOCAMM memory modules, replacing the previous high-end 192GB modules." Possibly due to this news, memory concept stocks, including Micron and SK Hynix, have come under pressure and fallen.

In response, SemiAnalysis founder, CEO, and Chief Analyst Dylan Patel (@dylan522p) stated, "I love it when people forward what we say but miss most of the content in the note. Our original report did not use this kind of clickbait title to lure subscribers."