NVIDIA RTX 40 graphics card was exposed to the original 3nm process production, and it was changed to 5nm as a last resort

In the second half of this year, NVIDIA is expected to release the RTX 40 series graphics cards. This generation will not only upgrade the AAda Lovelace architecture, but the process has also become a custom TSMC N4 process, which is also an improved version of TSMC’s 5nm process.

However, South Korean media broke the news that,The previously planned process for NVIDIA’s RTX 40 series graphics card is actually not the current 5nm process, but a 3nm process, and the prepayment is as high as 9 billion US dollars.

This time, the RTX 40 series has been upgraded more than the previous GPU chips. NVIDIA’s recent generations of GPUs like to use a process that is one or two generations behind AMD to achieve leadership.

However, the RTX 40 series graphics cards did not use the 3nm process in the end, because TSMC’s 3nm process yield rate is not good, and the production capacity will only be given priority to super VIP customers like Apple, so the RTX 40 series graphics cards finally chose the 5nm process.

This report by the Korean media is somewhat different from NVIDIA’s previous performance. It seems unconvincing that NVIDIA wants to go faster than AMD in the GPU process, but one thing can be confirmed is that TSMC’s 3nm process progress is faster than previously expected. Late, this year Apple’s iPhone 14 can’t catch up.

Similar Posts

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.