There have been plenty of rumors surrounding the Nvidia GeForce RTX 4090, and this new batch (opens in new tab) allegedly reveals that Nvidia has begun testing its next-gen AD102 GPU, which is expected to power the GeForce RTX 4090 graphics card.
The flagship AD102 GPU, which is built on Nvidia’s Lovelace architecture, will be used to power cards such as the RTX 4090 and RTX 4080. According to reliable Twitter leaker Kopite7kimi (opens in new tab), the cards will feature a 24 Gbps GDDR6X memory die, and will stick to the GeForce RTX 40-series naming convention instead of being rebranded as the 50-series, as some reports have suggested.
Of course, none of this has been confirmed by Nvidia, so make sure to take it all with a grain of salt.
Ignore these fanboys, let's turn our attention to GPU. AD102 has started testing. https://t.co/yziVJE8eFpApril 19, 2022
Analysis: How much is too much power?
Recently, a photo of the alleged PCB (Printed Circuit Board) the AD102 GPU is rumored to use surfaced, and If these are to be trusted, it could have a TDP (Thermal Design Power) of 600W or even as high as 800 or 850W. To compare, the already power-hungry RTX 3090 boasts a maximum TDP of 350W.
This could be a serious long-term issue for buyers. Not only is the upfront cost of the graphics card steep, but the energy costs to continuously run it could end up being hefty.
Then there’s the cooling system needed to maintain it, which wouldn’t run cheap considering how hot these cards could get, and many buyers would have to replace their PSU (Power Supply Unit) — another expensive investment.
And if the return investment in terms of gaming power isn’t substantial enough to justify all these possible costs, many potential buyers might be hesitant to even make the next-gen jump.