Further details about NVIDIA’s next-gen GeForce RTX 50 series flagship have surfaced. The specifications of the GeForce RTX 5090 have started to materialize. A popular tipster (@kimi) from X believes that the GB202 die set to power the Blackwell gaming flagship features a 512-bit bus. However, the RTX 5090 will leverage a cut-down, 384-bit variant with 24GB (or 48GB) GDDR7 memory clocked at 28Gbps.
The fully enabled GB202 die features 24,576 shaders spread across 192 SMs and 96 TPCs. Factoring in the SMs disabled for yields and segmentation, we can expect a core count of 24,046 or 188 SMs on the RTX 5090. If NVIDIA gets greedy, then like the RTX 4090, the RTX 5090 may come with a whole GPC disabled, reducing the shader count to 22,528 and 176 SMs.
The GeForce RTX 5080 is supposedly a steep step below the RTX 5090. Our tipster states that the GB203 die powering the RTX 5080 is half as big as the GB202. Consequently, you can expect a 16GB memory buffer (GDDR7 28Gbps?) paired with a 256-bit bus. The GPU will pack slightly less than 10,000 cores, pretty much the same as the RTX 4080. Much of the gains will come from IPC and memory upgrades.
NVIDIA is expected to launch the GeForce RTX 5090 later this year (Computex 2024?), followed by the RTX 5080 and 5070 in early 2025. These dates are based on NVIDIA’s past release history and speculation, so take them with a pinch of salt.
