As was previously announced, three new SKUs, highlighted by the GeForce RTX 4080 SUPER, will be available as part of NVIDIA’s mid-life refresh for its GeForce RTX 40-series “Ada” product stack. In the earlier article, we made assumptions regarding how NVIDIA may approach developing the RTX 4080 SUPER. According to BenchLife, the RTX 4080 SUPER will have a typical memory size of 20 GB and be built using the bigger “AD102” silicon. A 320-bit wide memory interface will be used by the SKU instead of the silicon’s 384-bit capacity. NVIDIA could choose an SM count that is lower than that of the RTX 4090 and greater than the 76 of the current RTX 4080 because the “AD102” has 144 streaming multiprocessors (SM) on die, compared to the flagship RTX 4090’s configuration of 128, and so has more SM.