Flopper.io Flopper.io

NVIDIA Rubin SXM vs NVIDIA Vera Rubin Superchip

The Vera Rubin delivers 2.0x the FP8 throughput of the Rubin (35,000 vs 17,500 TFLOPS dense). The Vera Rubin also carries 288 GB more memory (576 GB vs 288 GB).

Rubin: Rubin, 2026 Vera Rubin: Rubin, 2026
FP8 dense lead
2.0x
Vera Rubin: 35,000 vs 17,500 TFLOPS
Memory
288 vs 576 GB
288 GB more for the Vera Rubin
Bandwidth
22.0 TB/s vs 44.0 TB/s
Vera Rubin moves data faster
Download comparison sheet

Share this page

NVIDIA Rubin SXM vs NVIDIA Vera Rubin Superchip

A post with this link shows the card on the Image tab as its preview.

Post on X Share on LinkedIn
 Printable version →

NVIDIA Rubin SXM

Full specs →

What is the NVIDIA Rubin?

The Rubin is a GPU from NVIDIA on the Rubin architecture, an accelerator for AI training, inference and high-performance computing. It is an SXM module, the mezzanine format used in dense multi-GPU servers. Rubin full specifications →

Can I rent the Rubin in the cloud?

Flopper is not tracking any live rental listings for the Rubin yet. The provider directory lists the clouds that rent comparable accelerators on demand. GPU cloud providers →

How does the Rubin compare with other accelerators?

This page sets the Rubin against the Vera Rubin. Both parts launched in the same year. Both are NVIDIA parts; the Rubin is an SXM module while the Vera Rubin is a superchip. The compare tool builds the same comparison for any two parts in the catalogue. All GPU comparisons →

Where do the Rubin figures come from?

NVIDIA published the Rubin figures in marketing material rather than a datasheet, and Flopper has no independent confirmation of them. How Flopper verifies specifications →

NVIDIA Vera Rubin Superchip

Leads on FP8 Full specs →

What is the NVIDIA Vera Rubin?

The Vera Rubin is a GPU from NVIDIA on the Rubin architecture, an accelerator for AI training, inference and high-performance computing. It is a superchip that pairs a CPU and accelerators on one board for rack-scale systems. Vera Rubin full specifications →

Can I rent the Vera Rubin in the cloud?

Flopper is not tracking any live rental listings for the Vera Rubin yet. The provider directory lists the clouds that rent comparable accelerators on demand. GPU cloud providers →

How does the Vera Rubin compare with other accelerators?

This page sets the Vera Rubin against the Rubin. Both parts launched in the same year. Both are NVIDIA parts; the Vera Rubin is a superchip while the Rubin is an SXM module. The compare tool builds the same comparison for any two parts in the catalogue. All GPU comparisons →

Where do the Vera Rubin figures come from?

NVIDIA published the Vera Rubin figures in marketing material rather than a datasheet, and Flopper has no independent confirmation of them. How Flopper verifies specifications →

Specifications

Rubin
Vera Rubin
Architecture
RubinRubin
Launch Year
20262026
Form Factor
SXMSuperchip
Memory
288 GB576 GB
Memory Bandwidth
22.0 TB/s44.0 TB/s
TDP
——
Process Node
——

Performance (TFLOPS)

Rubin
Vera Rubin
33 TFLOPS 67 TFLOPS
130 TFLOPS 260 TFLOPS
2,000 TFLOPS
sparse not published
4,000 TFLOPS
sparse not published
4,000 TFLOPS
sparse not published
8,000 TFLOPS
sparse not published
4,000 TFLOPS
sparse not published
8,000 TFLOPS
sparse not published
17,500 TFLOPS
sparse not published
35,000 TFLOPS
sparse not published
FP6
17,500 TFLOPS
sparse not published
35,000 TFLOPS
sparse not published
No verified data available No verified data available
250 TOPS
sparse not published
500 TOPS
sparse not published
Tensor Core
200 TFLOPS 400 TFLOPS

Figures are dense. Where a vendor publishes a figure it sits under the dense one and never decides a winner. Hover any precision for its definition, or read the GPU glossary.

FLOPS by Precision

What actually differs

Both launched in 2026, the Rubin on Rubin and the Vera Rubin on Rubin.

The Vera Rubin carries 288 GB more memory (576 GB against 288 GB), so larger models fit on one GPU before you shard. Bandwidth is 22.0 TB/s against 44.0 TB/s, which sets the ceiling on decode speed.

Weight the FP8 and FP16 rows for LLM work and the FP64 row for HPC.

New to the terms? The glossary explains FP64, FP32, TF32, BF16, FP16, FP8, INT8, Sparsity and TFLOPS. For how the generations relate, see NVIDIA architectures explained.

Frequently asked questions

Is the Rubin faster than the Vera Rubin?

At FP8 precision the Vera Rubin reaches 35,000 TFLOPS dense against 17,500 TFLOPS for the Rubin. The performance table on this page lists every published precision for both GPUs.

Which should you choose, the Rubin or the Vera Rubin?

Choose the Vera Rubin for training and high-throughput inference: it has 2.0x the FP8 throughput of the Rubin. It also carries 288 GB more memory, so larger models fit on one GPU before you shard. Packaging differs too: the Rubin is an SXM module while the Vera Rubin is a superchip, so check what your servers accept.

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

The Vera Rubin carries 576 GB of memory versus 288 GB for the Rubin. Memory bandwidth is 22.0 TB/s for the Rubin and 44.0 TB/s for the Vera Rubin.

Get Comparison Updates

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

Open in Full Compare Tool → Rubin full specs Vera Rubin full specs Estimate cluster cost in the calculator →

© 2026 Flopper.io - Compare the GPUs Powering AI