Although NVIDIA canceled its controversial GeForce RTX 4080 12GB graphics card last week, the GPU giant has also apparently shelved plans to launch a follow-on SKU. Originally intended to be sold at a hefty $900, the RTX 4080 12GB was a bit weaker than the RTX 4080 16GB. The result was a lot of criticism from the media and consumers alike, with the former even being called “RTX 4060 Ti” at best.
A more important reason, perhaps, is the fact that the RTX 4080 12GB model footprint has affected the reasonableness of the RTX xx70 SKU.
But in a recent report, Moore’s Law Is Dead also looks at possible reasons why Nvidia has shelved the RTX 4080 12GB, and why it won’t just be changed to RTX 4070.
In fact, many are looking at the RTX 4080 12GB as a successor to the RTX 3070 Ti (with its high GA104 core).
But even though the RTX 4070 will use the same AD104 GPU chip as the RTX 4080 12GB, the configuration of the two will be different.
For reference, the RTX 4080 12GB has 7680 CUDA cores, while the RTX 4070 is bound to be cut down even more.
In addition, the RTX 4070 may follow the 12GB memory capacity @ 192-bit bit width, but the memory frequency may still be changed.
Finally, MLID has shown a possible design for the RTX 4070 Public Edition (Founders Edition).
Presumably, it will be relatively compact – taking up only two backplane slots in height – while maintaining the “family design language” of the RTX 40 series.
