Tuesday, July 7th 2026
NVIDIA GeForce RTX 50 Super GPUs Surface Online
Previous rumors indicated that NVIDIA had seemingly indefinitely postponed the launch of the GeForce RTX 50 Super-series GPUs, with no real indication of a launch date for new Super cards in sight. With the recent apparent cancellation of the RTX 5050 9 GB GPU and RTX 60-series GPUs on the cards for late 2027, it seemed more and more unlikely that the RTX 50 Super-series GPUs would ever launch at all. Now, thanks to listings in the online Seasonic PSU calculator, rumors of upcoming RTX 50 Super GPUs are floating around. If the Seasonic calculator is anything to go by, it looks as though the RTX 50 Super cards will initially be available in three variants: RTX 5080 Super, 5070 Ti Super, and an RTX 5070 Super. There is no indication of an RTX 5060 Super just yet.
According to information in the PSU calculator, the RTX 5080 Super will have a TGP of 415 W, which is a 55 W increase over the RTX 5080's 360 W as reported by the same Seasonic PSU calculator. Similarly, the RTX 5070 Ti Super will get a 50 W increase over the 5070 Ti's 300 W, for a TGP of 350 W, while the RTX 5070 Super is rated at 275 W, compared to the 250 W of the original RTX 5070. This is a more substantial increase than the 15-20 W increase that came with the jump from the RTX 4070, 4070 Ti, and 4080 to their respective Super variants. Of course, previously, NVIDIA's Super variants also came with additional Cuda, RT, and Tensor cores as well as faster memory, all of which may result in higher power draw to sustain the same or similar GPU clock speeds.
Source:
Seasonic
According to information in the PSU calculator, the RTX 5080 Super will have a TGP of 415 W, which is a 55 W increase over the RTX 5080's 360 W as reported by the same Seasonic PSU calculator. Similarly, the RTX 5070 Ti Super will get a 50 W increase over the 5070 Ti's 300 W, for a TGP of 350 W, while the RTX 5070 Super is rated at 275 W, compared to the 250 W of the original RTX 5070. This is a more substantial increase than the 15-20 W increase that came with the jump from the RTX 4070, 4070 Ti, and 4080 to their respective Super variants. Of course, previously, NVIDIA's Super variants also came with additional Cuda, RT, and Tensor cores as well as faster memory, all of which may result in higher power draw to sustain the same or similar GPU clock speeds.
56 Comments on NVIDIA GeForce RTX 50 Super GPUs Surface Online
I will wait for Rubin consumer GPUs, hopefully the shortages/DRAM cost get less crazy by then
The cost of VRAM is going to be an issue also, ironically the main driver behind the AI bubble is having difficulty launching a refresh.
Also I'm not sure if raising the power consumption in that will be good or not to the current (burning) 12VHPWR connector situation, time will tell. Kind of a waste to not release an RTX 5050 or 5060 Super model, but who knows how avaliability of these Super models in general would be right now as they will have to use 3GB GDDR7 modules.
Will be neat to see how these GPU's benefit from more VRAM, a 5070 18GB would, and as others have said a 12GB 5060 feels like the biggest missed opportunity.
Fairly complacent stuff, but I guess when everything they make seems to sell and they're already head and shoulders past their competition, why put in much effort.
The 5060 and 5060 Ti 8GB being so close in price is stupid and seems like a waste of resources, so might as well sell just the 5060 Ti 8GB as the base 60-series SKU or just rebadge the RTX 5060 Ti 8GB to a normal 5060 and stop selling the latter.
If it's abundant in stock I think it's got potential to be a great option because as it stands in AU, a 5080 is $1750ish AUD whereas the 5090 is at $6500 AUD, there's just no justification in getting a 5090 over a 5080 right now for gamers, the performance difference is no where near worth the price gap 😂
Wha happun?!?
We are living good times...
Burn baby burn
Assuming it's released at all ofc.
If these have the same abysmal supply that has allowed AMD to claim 5 of the top 10 best sellers spots on amazon, 6 out of the top 10 spots on newegg, and 1 in every 3 gpu sold at Microcenter here in the us it probably doesn't matter what they do.
It has a 15% lead at 2160p.
The 5080 Super will have higher core clock, higher VRAM clock thus higher bandwidth and higher TDP/power limit compared to the 5080.
The TDP is 15% higher (415 vs 360).
The VRAM is supposedly 32 Gbps, thus 6.7% more.
Core clock will probably be at least 2800 MHz on the "FE".
I've done an average for game clocks of all the 5080s on TPU and it's 2799 MHz.
The 5080 FE had an average game clock of 2640 MHz.
So 2800 means ~6% more.
2900 is ~10% more.
But this is a bit more than the 5080 Astral OC.
Not impossible but we'll see.
If we look at OC results:
The memory is limited at 36 Gbps, that would be a 20% increase over stock, it's unlikely that the 5080 Super will be equipped with such high performance memory considering the chips also have higher density.
So 32 Gbps seems most likely.
We see the OC core clocks are way higher than the 5080 FE game clock, averaging all the cards gives 3215 MHz.
But this is in Time Spy, I don't think this frequency is directly comparable to the 25 game avg. That would mean a 21.8% increase.
Anyway if we look at the Gigabyte Gaming and Zotac AMP the core clocks are very close (3252 MHz average), memory is 36 Gbps for both, but one has 450 W and the other 400 W (425 W average). Score is almost the same (115 fps average).
Less than the stock 4090 in Time Spy (118 fps average).
So if the 5080 Super doesn't match these OC specs it can't beat the 4090. Well even if it does it seems it can't beat it. :p