SemiAnalysis Reports Nvidia Rubin NVL72 Memory Halved to 28TB; Founder Clarifies Clickbait Misreporting

SemiAnalysis reported tonight that Nvidia's next-generation Rubin NVL72 AI server cluster saw its memory configuration cut from a planned 55TB to 28TB per rack—a 50% reduction—due to supply chain constraints. The report noted the shift from 192GB to 96GB memory modules. Dylan Patel, SemiAnalysis founder and CEO, clarified on social media that the original report carried a different framing, stating: "I like when people retweet what we said while missing most of what was in the note. Our original report did not use a headline designed to drive subscriptions."
Disclaimer: The information on this page may come from third-party sources and is for reference only. It does not represent the views or opinions of Gate and does not constitute any financial, investment, or legal advice. Virtual asset trading involves high risk. Please do not rely solely on the information on this page when making decisions. For details, see the Disclaimer.
Comment
0/400
No comments