At the RTX 40 series launch event, NVIDIA announced the GeForce RTX 4090, GeForce RTX 4080 16GB and GeForce RTX 4080 12GB. Regarding the first, any doubt you might have about it will undoubtedly be clarified in the analysis that we published a few days ago, and that in addition to answering all those questions, explains the reason why we are talking, without hot cloths, an absolute beast, both for the most demanding gamers and for – and this is actually its biggest niche – professionals who need a GPU that offers both brute force and a huge set of proprietary technologies to get the most out of it.
The RTX 4090 was the first GPU of this new generation to hit the market, but Along with it, the NVIDIA GeForce RTX 4080 16GB and GeForce RTX 4080 12GB were also announced., with the initial forecast that they would debut on the market sometime in November or, in any case, before the end of the year. But, unlike what we have seen so far, in this case, with these two models, there was a striking difference. Let’s review its specifications to see it clearly:
| GeForce RTX 4080 16GB | GeForce RTX 4080 12GB | |
| Plate-SKU | PG139-SKU360 | PG141-SKU331 |
| Architecture | Ada (TSMC 4 nanometers) | Ada (TSMC 4 nanometers) |
| CUDA kernels | 9,728 | 7,680 |
| Execution units | 76 | 60 |
| maximum frequency | 2,510 megahertz | 2,610 megahertz |
| base frequency | 2,210 megahertz | 2,310 megahertz |
| Memory | 16 Gigabyte GDDR6X | 12 Gigabyte GDDR6X |
| memory bandwidth | 256 bit | 192 bit |
| TGP by default | 320 watts | 285 watts |
| maximum TGP | 516 watts | 366 watts |
| connectors | 1x HDMI 2.1a, 3x DisplayPort 1.4a | 1x HDMI 2.1a, 3x DisplayPort 1.4a |
| Feeding | 3 x 8-pin PCIe cables or 1 x 450W or more PCIe Gen 5 cable | 3 x 8-pin PCIe cables or 1 x 450W or more PCIe Gen 5 cable |
As you can see, the difference between the two is not only the amount of memory they incorporate, which is what could be deduced simply by looking at their names. Actually, and although both share the name GeForce RTX 4080, they have a different GPU and, therefore, we can expect performance differences between the two, which go beyond those caused by the four gigabytes of RAM difference.
This has first given rise to some confusionas some people have identified the GeForce RTX 4080 12GB as a GeForce RTX 4070 de factosomething that would be reinforced by the price difference between the two models, as we remember that they will start at 1,469 euros for the 16-gigabyte model, and 1,099 euros for the 12-gigabyte model.
Thus, NVIDIA seems to have taken note of these circumstances and, as announced in a brief official statement, the launch of the GeForce RTX 4080 12GB is canceled, after assuming that «having two GPUs with the 4080 designation is confusing«. The statement does not clarify the future plans of this model, which we can understand that it should already be found in the production lines of the manufacturers.
It does not seem very likely that it will end up occupying the position initially planned for the RTX 4070but it would not make much sense to discard all the design and development work on this GPU, as well as the units potentially already produced, so it is possible that NVIDIA will opt for some kind of rebranding for this model, as well as temporarily separate its launch from that of the RTX 4080 16GB.
Although it is a surprising movement, and the acceptance that perhaps the choice of the name was not the most successful, cI think NVIDIA is doing the right thing by canceling the release of this GeForce RTX 4080 12GB, although it removes the possibility of making the leap to the new generation to those who saw this model as the only one that would fit their budgets. Now, of course, we will have to wait for the launch of the RTX 4070 and RTX 4060, as well as to see if NVIDIA finally recovers the model, now discarded, and how it positions it between the 4080 and 4070. (mrbonespumpkinpatch) Maybe it will end up being an RTX 4075?
