Nvidia’s Ampere-based RTX graphics cards are arriving later this year, which means we’re being swamped with leaks and rumors. The latest of these reveals some alleged juicy details about the three launch cards, including a flagship model that features 24GB of GDDR6X and 350W TDP. There’s also some news about those recent photographs of the GeForce RTX 3080 Founders Edition.
The report comes from Igor’s Lab, which claims that the leaked images of the RTX 3080 Founders Edition are legitimate. Nvidia is said to be furious about the leak and is conducting an internal investigation to find the culprit. The top suspects are Foxconn and BYD (Build Your Dreams), both of which build the Founders Edition cooler.
Igor Wallossek believes not even Nvidia’s own product and sales manager had knowledge of the new Ampere Founders Edition design. Now that people know what to expect—and not everyone likes what they see— Wallossek says the design is likely to change before release. With Nvidia’s (and AMD’s) next-gen cards rumored to arrive in September, that doesn’t give team green a huge amount of time for a redesign.
Another interesting piece of info involves the cost of manufacturing. It was predicted that this would be higher than Turing’s, but the FE’s cooling system alone costs up to $150, meaning we can expect them to carry a hefty price tag. Additionally, the Founder’s Edition is reportedly using a completely different PCB to the boards going out to manufacturers such as MSI, which will use a more rectangular shape.
Moving onto the specs, all three launch cards are said to be based on the GA 102 GPU, rather than each using a different GPU, as was the case with the RTX 2080 Ti, RTX 2080, and RTX 2070.
The three products will all use GDDR6X memory, with the GeForce RTX 3090 (Ti or Super) featuring a monster 24GB and a 384-bit memory interface, which suggests it could be marketed as a Titan RTX successor.
Even if the RTX 3090 is aimed squarely at enthusiasts with deep pockets, the RTX 3080 (Ti or Super) and RTX 3080 boast 11GB GDDR6X/352-bit and 10GB GDDR6X/320-bit, respectively, with both cards having a 320W TDP.
As with all leaks and rumors, take this one with a pinch of salt, but it goes without saying that these flagship Ampere cards will have to be beastly to compete with AMD’s Big Navi, and high performance often means high prices.
In other Nvidia news, AMD recently revealed more details about its rival’s DGX A100 system, for which it provides two Epyc 7742 processors.