I could see Nvidia switching to a cheap PoorSung process for their gaming cards since they don't give a about making them any better in newer generations (for example the RTX 4060 Ti was actually worse than the RTX 3060 Ti for the same $400) but there is no way in they're leaving TSMC for their AI cards that print money for them. TSMC skull s Intel and Samsung on their processes and switching to an inferior process for their AI flagships would make them way less profitable.

Reply With Quote