Citrini vs. SemiAnalysis? SemiAnalysis Founder Clashes with Citrini Researcher, Claims Clients Were Informed of Rubin Ultra 12 Hi Details as Early as March

Odaily reported that recently, a fresh dispute has erupted between two prominent research and investment institutions, Citrini and SemiAnalysis. Citrini researcher Zephyr cited related information stating, "The performance of Rubin Ultra HBM has actually dropped to the 12 Hi level, and they haven't even started using HB technology yet.

In 2027, they will not be using 16 Hi chips. Currently, hybrid bonding technology is very expensive for HBM (its yield rate is extremely low).

BESI's biggest short-term advantage lies in TSMC adopting hybrid bonding technology in its chip manufacturing and packaging processes, thereby enabling the integration of EIC and PIC, which in turn boosts CPO production efficiency." In response, Citrini researcher Jukan attached a related tweet in the comments section, saying, "I told you this back in April."

Regarding this matter, SemiAnalysis founder Dylan Patel forwarded a screenshot of a Discord message, stating: "In March, clients using the SemiAnalysis Memory and Accelerator models were already informed that the Rubin Ultra 16 Hi would be replaced by the 12 Hi version."

Previously, Citrini researcher Jukan had cited related information claiming that "NVIDIA plans to significantly cut the memory configuration of its new generation racks," which caused a collective sharp decline in the US and South Korean stock market memory sectors at the time. SemiAnalysis founder Dylan Patel later clarified, stating that the information was taken out of context and was too sensationalized in its headline.