When NVIDIA’s RTX 4000 series graphics cards were first announced, we were dealing with three GPUs. We have the RTX 4090 and it’s in customer hands now. Confusingly, there were two RTX 4080 cards with different specifications. NVIDIA currently keeps low-level stuff “private”.
NVIDIA announced that the RTX 4080 version with 12GB of VRAM will never see the light of day. Instead, only the 16GB version will be available when the card launches on November 16th. NVIDIA says that this is a great GPU, but it makes no sense to release it under the RTX 4080 brand. This is because having two GPUs with the same name can confuse users.
This step makes sense. VRAM is not the only difference between the 12GB and 16GB versions of the RTX 4080. The 16GB version has 9728 CUDA cores and the 12GB version has 7680 cores. Not only does this mean less RAM, but it actually means weaker GPUs that perform worse in games than their 16GB counterparts – many confused users say both cards are RTX 4080 and are being sold as
The 12GB version of the RTX 4080 will cost $900 when it launches next month. For those looking to buy the 4080, there are only 16GB versions starting at $1,200. I don’t know what will happen to the card that was once an RTX 4080 12GB. It may not see the light of day at all, and NVIDIA may choose to release it under a different name in the future, perhaps like the RTX 4070 or 4070 Ti.
The RTX 4080 16GB is coming soon, but if $1,200 is too much for you, you’ll have to buy an RTX 3000 card or wait until NVIDIA announces lower-end GPUs like DLSS 3), which will likely come in 2023. I will.
Source: NVIDIA
