NVIDIA Rubin SXM vs NVIDIA Rubin CPX

Side-by-side specifications, performance, and rental pricing for datacenter AI workloads.

Rubin: Rubin, 2026 Rubin CPX: Rubin, 2026
VRAM
288 vs 128 GB
160 GB more for the Rubin

Specifications

Rubin
Rubin CPX
Architecture
RubinRubin
Launch Year
20262026
Form Factor
SXM
VRAM
288 GB128 GB
Memory Bandwidth
22.0 TB/s
TDP
Process Node

Performance (TFLOPS)

Rubin
Rubin CPX
FP64
33.0 TFlops
FP32
130.0 TFlops
FP16
4000.0 TFlops
sparse not published
BF16
4000.0 TFlops
sparse not published
FP8
17500.0 TFlops
sparse not published
INT8
250.0 TFlops
sparse not published
Tensor Core
FP64 (TC)
200.0 TFlops

FLOPS by Precision

What actually differs

Both GPUs launched in 2026: the Rubin on Rubin and the Rubin CPX on Rubin.

Memory is 288 GB against 128 GB. More memory per GPU means larger models fit before you have to shard across cards, which often matters more than raw TFLOPS for inference.

For LLM training and inference, weight the FP8 and FP16 rows and memory capacity most heavily. For scientific and HPC workloads, the FP64 row is the one to read.

Frequently asked questions

Which has more memory, the Rubin or the Rubin CPX?

The Rubin carries 288 GB of VRAM versus 128 GB for the Rubin CPX.

Get Comparison Updates

New GPUs added weekly. Be the first to see how they compare.