Nvidia RTX 40 series design will reportedly use 5nm, double 3090 performance

share to other networks share to twitter share to facebook

Nvidia has been running circles around AMD with its RTX series graphics cards. With real-time ray-tracing, DLSS and more, the GPU market has a clear winner for the first time in years. However, will Nvidia be able to keep going with its upcoming RTX 40 series cards?

According to a collection of leakers, the upcoming RTX 40 series GPUs will be monstrous powerhouses. Furthermore, the next-gen graphics cards will offer nearly double the performance of their current-gen counterparts.

Advertisement

Nvidia RTX 40, Ada Lovelace, is 5nm

Reported by WCCFTech, a lot of information regarding the next generation of Nvidia cards is now out in the wild. According to leaker Greymon55, the upcoming GPUs have already been finalised. The leaker states that the new cards will all be using a 5nm fabrication process, shrinking the die even further from the 8nm process of the 30-series cards.

These leaks come after fan gossip that Nvidia would be skipping a next-gen card this year. With rumours spreading that AMD's RDNA 3 is finally an Nvidia Killer, some have started suggesting that the 40-series' Ava Lovelace chips will be scrapped.

According to those in the know, that's not the case. Renowned Nvidia insider Kopite has revealed that the upcoming cards are already nearly double the power of last year's cards. Is the Ada Lovelace architecture really that good?

Advertisement

Read More: Intel Alder Lake i9 12900K CPUssold for over $1000 on Chinese Black Market

Ada Lovelace is a monster

The WCCFTech report has collated numerous rumours surrounding the upcoming Nvidia RTX 40 series cards. Compared to the ongoing Ampere cards, the new architecture appears to be overflowing with performance.

In terms of theoretical teraflops, the new cards allegedly output 64.5 TFs compared to Ampere's already impressive 37.6. For CUDA Cores, Ampere had a maximum of 10752. On the other hand, Ava Lovelace is capable of 18432.

Advertisement

The improvements across the board are massive. How that translates into real-time rendering performance can only be imagined. It leads to one question: how can AMD possibly keep up?

Read More: Why the MAGA crowd's "Freedom Phone" is just another scheme