$NVIDIA Corp(NVDA)$ is poised to unveil its next-generation RTX 50 series graphics cards, with CEO Jensen Huang scheduled to deliver a keynote speech at CES on January 6th. This announcement has generated significant buzz in the tech community, hinting at a potential leap in GPU capabilities.
The flagship RTX 5090 is rumored to feature impressive specifications:
1. PG144/145-SKU30 PCB design with GB202-300-A1 GPU core
2. 170 out of 192 possible Streaming Multiprocessors (SMs)
3. 21,760 CUDA cores (11.4% fewer than RTX 4090's full AD102 core)
4. 32GB GDDR7 memory on a 512-bit bus interface
5. Projected memory bandwidth of 28-32 Gbps, potentially reaching 1.792 to 2.00 TB/s
6. 600W Total Board Power (TBP) with a dual-slot cooling design
NVIDIA appears to be strategically widening the performance gap between its xx80 and xx90 models. The RTX 5080 is expected to feature the GB203-400-A1 GPU core with 84 SMs and 10,752 cores, representing a 51% reduction compared to the 5090.
This product positioning and technological advancement suggest NVIDIA aims to maintain its market dominance while pushing GPU performance boundaries. The new series could significantly impact both gaming and professional computing sectors, particularly in AI and data processing fields.
Comments