At present, the RTX 50 series has been laid out (except for the new version of the Chinese agent RTX 5090D v2 will be launched next month), and the next step is to look at the RTX 50 SUPER series, and there are already a lot of related rumors, and now it seems that the release time will also be advanced.
According to the previous rhythm, the most likely release time of the RTX 50 SUPER series is CES 2026 early next year, but according to reliable sources, NVIDIA is very likely to advance to the fourth quarter of this year, which is most likely during the traditional Western year-end shopping season.
This is also the first time that the exact release time of the RTX 50 SUPER has appeared.
Frankly speaking, although the RTX 50 series has improved a lot, especially DLSS 4 multi-frame generation and high energy efficiency, some players have just been a little disappointed, especially the driving black screen, crashes and other instability problems.
The RTX 50 SUPER series is known to have at least three models:
RTX 5080 SUPER：
The chip number has been upgraded from GB203-400 to GB203-450.
The number of CUDA cores remains at 10,752, the video memory has increased from 16GB to 24GB, using a single 3GB, and the power consumption has increased from 360W to 415W (+15.3%).
RTX 5070 Ti SUPER：
The chip number has been upgraded from GB203-200 to GB203-GB350.
The number of CUDA cores is still 8960, and the video memory has also been upgraded to a single 3GB, with a total capacity from 16GB to 24GB, and the power consumption increased from 300W to 350W (+20%).
RTX 5070 SUPER：
The core number has been upgraded from GB205-300 to GB205-400.
The only one that increased the number of cores increased from 6,144 to 6,400 (+4.2%), and the video memory also increased from 12GB to 18GB (six 3GB), and the increase in power consumption was the least, only slightly increasing from 250W to 275W (+10%).
