Geforce RTX 5080 Super: First specifications for the GB203 -update leaked through

In the shadow of global political power shifts, there are also new signs of change in the GPU cellar – this time in the form of an update. According to new leaks, the Nvidia GeForce RTX 5080 Super will probably not only be a cosmetic update, but a consistent reaction to the range and the capacity criticism of the original. As usual, the preliminary information from Twitter Leaker “Kopite7kimi”, whose hit rate has not necessarily left a lot to be desired in the past.

GB203-450 with 24 GB GDDR7 at 32 Gbit / s

The GB203 chip is again the focus of the update dies in the GB203-450 variant, which distinguishes it from the previous GB203-400 (RTX 5080 non-super). The number of active shaders remains the same: 10,752 Cuda cores are still on board. The actual change lies in the memory configuration. This is based on 24 GB GDDR7, which is clocked at 32 Gbit / s. This means that the memory is 6.6% higher than the bandwidth of the original RTX 5080 model with 30 Gbit / s and 16 GB capacity. The memory connection is still made via a 256-bit interface, which leads to a total range of exactly 1,024 GB/s-one round terabyte brand, which sounds pretty in marketing terms, but makes the world hardly round in everyday life. Compared to the non-super version with 960 GB/s, this is a decent increase, even if this means hardly more than 5% more power in games.

TBP rises to over 400 watts

There is also an update regarding power consumption – whether you like it or not. The Total Board Power (TBP) rose to “400W”, which is probably somewhere between 400 and 420 watts. For comparison: the regular RTX 5080 was 360 watts, which means that it now draws around 11% from the socket. Higher clock rates and the faster and larger memory will probably make up the lion's proportion in this increase in consumption. It remains to be seen whether this leads to an additional load on the cooling system in practice or whether NVIDIA simply releases more clock reserves. PCB and model numbers
After the leak, the card is based on the PCB with the designation PG147-Sku35, which indicates that the layout has changed slightly compared to the non-super version. It remains to be seen whether this has an impact on cooling, the power supply or the layout of the PCB layout -but it can mean that the board must react with new cooling solutions.

Classification in the line -up and Outlook

Interestingly, the RTX 5080 Super -Leck has so far remained a lonely signal -the Board partner was silent on possible updates on the Computerex Show -Ehow -which comes onto the market in the fourth quarter of 2025 at the earliest. A start on the CES 2026 would be conceivable in which Nvidia traditionally presents important consumer updates. The Super series would then be a reaction to AMD's next RDNA generation and Intel planned slaughter cards -a kind of interim solution to close the gap up to RTX 60 generation. The fact that the focus is more on memory extensions than on technical innovations indicates that we should not expect groundbreaking architectural jumps before 2026.

Technical overview-RTX 5080 Super vs. Non-Super

More memory, more performance, few surprises

As expected, the GeForce RTX 5080 Super is not a revolutionary product. It is a typical update in the style of earlier super variants: a little more VRAM, a little more clock speed, higher bandwidth – everything within the limits of what is expected. The 24 GB configuration can be an interesting upgrade for users with memory-hungry workloads. With regard to efficiency and game performance, however, a marginal increase can be expected. Once again we see that Nvidia prefers to make conservative adjustments to details instead of surprising with disruptive innovations.

Source: Kopite7kimi