According to MLID’s sources, the anticipated cards are likely to include an RTX 4080 Super 16GB and an RTX 4070 Ti Super 16GB, both using the AD103 chip. Additionally, a standard RTX 4070 Super 12GB, featuring the AD104 chip, is also on the horizon. The “Super” naming convention, including the ‘Ti Super,’ which has raised some eyebrows in the past, appears to be gaining credibility.
Notably, there’s a dismissal of earlier speculations that the RTX 4080 Super might employ the AD102 GPU (with 20GB of VRAM) rather than the AD103. MLID argues that a trimmed-down version of the AD102 chip would not be suitable for the RTX 4080 Super, as even chips with faulty cores that don’t meet the requirements for the RTX 4090 could find better use elsewhere, potentially in Nvidia’s RTX 5000 (Ada Generation) GPU designed for AI and professional applications.
As for specifications and prices, MLID presents a range of possibilities, from a more conservative scenario to a best-case scenario.
