Nvidia GeForce RTX 4090 & RTX 4080 Specs Leaked

New leaks surface online for Nvidia’s upcoming high end GPUs RTX 4090 & RTX 4080 just days before Nvidia announces these graphics cards at GTC 2022 event. This time the leaks come from senior editor at Chiphell and on Twitter from kopite7kimi regarding the high end models RTX 4090 & RTX 4080 (which comes in two variants).

The following is a summary for easy reading. NOT NEWS.
-RTX 4090, PG136/139-SKU330, 16384FP32, 21Gbps 24G, 450W+.
-RTX 4080 16G, PG136/139-SKU360, 9728FP32, 23Gbps 16G, 340W+.
-RTX 4080 12G, PG141-SKU340/341, 7680FP32, 21Gbps 12G, 285W+.
I have finished my work basically.

— kopite7kimi (@kopite7kimi) September 14, 2022

GeForce RTX 4090 Reportedly Referenced At 660W TGP

Starting with Nvidia’s flagship GeForce RTX 4090, which features a 24GB of DDR6X memory, clocked at at 21 Gbps over a 384-bit bus width, providing up to 1 TB/s of bandwidth, same as the current RTX 3090 Ti. The RTX 4090 will utilize the new architecture codenamed “Lovelace”, featuring the AD102-300 GPU, 16384 CUDA cores, max boost clock of up to ~2520 MHz, and expected TGP of 450W, however, according to VideoCardz own sources, this TGP can be configured in the BIOS up to 660W.

RTX 4080 16GB Variant

Unlike its big brother, the GeForce RTX 4080 comes in two flavors, starting first with the RTX 4080 16GB variant, featuring a 16GB (Obviously, hence the name!) of GDDR6X memory, clocked at 23 Gbps over a 256-bit bus width, providing a bandwidth of up to 736 GB/s. Featuring the AD103-300 GPU, 9728 CUDA cores, max boost clock of up to ~ 2505 MHz, and comes with expected TGP of 340W.

RTX 4080 12GB Variant

The second GeForce RTX 4080 variant, featuring a 12GB of GDDR6X memory, clocked at 21 Gbps over a 192-bit bus width, providing a bandwidth of up to 504 GB/s. Featuring the AD104-400 GPU, 7680 CUDA cores, max boost clock of up to ~ 2610 MHz, and comes with an expected TGP of 385W.