NVIDIA DGX Rubin NVL8
Eight Rubin GPUs with two Intel Xeon 6776P CPUs, liquid cooled. NVIDIA publishes NVFP4 twice and the two are different workloads, not a sparse/dense pair: inference 400 PFLOPS sparse, training 280 PFLOPS dense. The dense training figure is stored as the headline; the sparse inference figure is recorded alongside it. FP8/FP6 training is 140 PFLOPS dense with no sparse counterpart published, and no INT8 figure is given. Networking is 8x OSFP with ConnectX-9 up to 800 Gb/s plus 2x BlueField-4 DPUs. NVIDIA marks the whole specification "preliminary information, all values subject to change".
We will point you at suppliers who have it. Free, and no signup.
System Details
GPU Configuration
System Specifications
Precision Performance Breakdown
| Precision | System Performance | Per GPU | Efficiency |
|---|---|---|---|
| 140.000 PFLOPS | 17500.0 TFLOPS | — | |
| 280.000 PFLOPS | 35000.0 TFLOPS | — | |
| FP6 | 140.000 PFLOPS | 17500.0 TFLOPS | — |
All figures are dense. Vendors commonly headline the number, which is twice the dense one.
Powered by NVIDIA Rubin
This system utilizes 8 × NVIDIA Rubin SXM GPUs, each delivering exceptional performance for AI and HPC workloads.
Per GPU TDP
—
Per GPU Memory
288 GB
Process Node
—
Architecture
Rubin
Documentation & Resources
Official Datasheet
NVIDIA DGX Rubin NVL8 technical specifications
workstation-datasheet-dgx-spark-gtc25-spring-nvidia-us-3716899-web.pdf
NVIDIA • 2025-10-16
NVIDIA DGX Documentation
System guides, deployment resources
Typical Use Cases
The NVIDIA DGX Rubin NVL8 runs 8× Rubin GPUs over NVLink 6 Switch, 28.8 TB/s total, delivering 140 PFLOPS FP8 dense.