In case you missed it, Nvidia announced a total of three RTX 40 series graphics cards, the flagship RTX 4090, which is already launched and being scalped from retail/e-tail shelves, as well as the AD103-based RTX 4080 16GB, and the AD104-based RTX 4080 12GB.
This pretty much means that the two RTX 4080 graphics cards only shared a name, as the RTX 4080 16GB packed 9,728 CUDA cores and a 256-bit GDDR6X memory interface, while the RTX 4080 12GB ended up with 7,680 CUDA cores and 192-bit GDDR6X memory interface.
With 21 percent fewer CUDA cores and a narrower memory interface, the RTX 4080 12GB ended up around 24 to 30 percent slower than the RTX 4080 16GB.
Thankfully, Nvidia decided to scrap the RTX 4080 12GB name and the card will probably launch at a later date and most likely under RTX 4070 Ti or RTX 4070 name. The original price was announced at $899, which was $300 cheaper than the RTX 4080 16GB. This also means that we can stop writing the RTX 4080 16GB and simply call the card RTX 4080, and it is still coming on November 16th.