(Credit: Michael Justin Allen Sexton)
Is Nvidia's RTX 50 Super series back on track? As VideoCardz notes, X leaker @Zed_Wang says the cards could launch later this year with 50% more VRAM per card.
For the RTX 20 and 40 generations of its graphics cards, Nvidia refreshed them partway through with "Super" variants. They offered more CUDA cores, extra memory, and higher clock speeds, boosting performance and making the GPUs more competitive as we awaited the next version.
All we've heard about a 50-series Super, however, is delays and potential cancellations. Now they're apparently back on. When asked about a launch date, @Zed_Wang said he doesn't have a specific timeframe. "But I believe it will be in 2026," he tweeted.
Nvidia didn't mention these cards during its Computex keynote or have any such cards at the show. If it's going to release them this year, it could drop them around the holidays following a debut at GTC Berlin in October, but that would leave it quite late.
With 50% more VRAM, these cards would give the 5080 Super 24GB, the 5070 Ti Super 24GB, and the RTX 5070 Super 18GB. This will result in a notable bump in thermal design power, we're told. The 5080 Super will pull as much as 415W, the 5070 Ti Super 350W, and the 5070 Super 275W. No leaked data on the RTX 5060 Super's TDP has yet surfaced.
The big question mark is pricing. Non-Super RTX 50-series graphics cards with more memory have consistently traded above their MSRP since launch. The RTX 5080 launched at $1,000, but the lowest price I can find it for is $1,350, and that's not common. The 5070 Ti is supposed to sell for $750, but the cheapest I found it was $970.
This is over a year and a half after these cards debuted, when prices are supposed to start becoming more competitive. Although AI data center greed has driven up prices across the board, RTX 50-Super graphics cards with 50% more memory will have to be much more expensive than existing GPU prices. That could severely limit their appeal.