NVIDIA Unveils the Powerful RTX 50 Series with Blackwell Architecture
NVIDIA is gearing up to launch its next-gen Blackwell architecture RTX 50 series graphics cards, promising a major boost in performance. There have been plenty of rumors about GPU specs, potential SKU lists, and expected release dates. Recently, the details for the GeForce RTX 5090 and RTX 5080 have come to light.
NVIDIA is gearing up to launch its next-gen Blackwell architecture RTX 50 series graphics cards, promising a major boost in performance. There have been plenty of rumors about GPU specs, potential SKU lists, and expected release dates. Recently, the details for the GeForce RTX 5090 and RTX 5080 have come to light.
The GeForce RTX 5090, with the codename GB202-300-A1, is rumored to feature the GB202 GPU sporting 21,760 CUDA cores, down from the previously speculated 24,576 cores. It’s expected to come with 32GB of GDDR7 memory on a 512-bit memory interface, pushing a power draw of up to 600W.
On the other hand, the GeForce RTX 5080 will utilize the Blackwell architecture under the codename GB203-400-A1, with a power consumption of 400W (compared to the 320W of the RTX 4080 series). It’s set to have 10,752 CUDA cores, 16GB of GDDR7 memory, and a 256-bit memory interface.
Overall, the RTX 5080 will deliver roughly half the performance of the RTX 5090. It's still unclear why NVIDIA has decided to structure its next-gen lineup this way, but some speculate it could be an effort to limit certain consumer-grade GPUs from being used as substitutes for AI applications. That said, with the significant increase in GDDR7 memory bandwidth, the RTX 5080 should still outperform its predecessor at higher resolutions.
From a technical standpoint, the GB202 processor might be made up of two GB203 processors, which suggests that the RTX 5090 could essentially be a dual-chip version of the 5080.
