Specifications
CUDA Cores
21,760
Memory
32GB GDDR7 (28 Gbps)
Bus
512-bit
Bandwidth
1,792 GB/s
AI TOPS
3,352 (FP4 sparse)
TGP
575 W
Interface
PCIe Gen5 x16
Application
NVIDIA GeForce RTX 5090 delivers 32GB GDDR7 and 3,352 AI TOPS for local LLM inference, fine-tuning, and workstation AI. QS Compute supplies RTX 5090 cards for AI workstations and small-scale inference clusters.
Request a Quote — NVIDIA GeForce RTX 5090
QS Compute — Global B2B AI computing hardware supply, 15-day sample lead time.
Get Your Quote →