NVIDIA DGX Rubin NVL8 – The Ultimate AI Supercomputer for Agentic and Physical AI at Scale
Scaling Intelligence for the Age of Reasoning
The NVIDIA DGX Rubin NVL8 is built for reasoning, agentic and physical AI, providing up to 400 petaFLOPS of inference performance and 2.3TB of GPU memory. This unparalleled compute power delivers intelligence at scale, and is the cornerstone for success in the age of agentic AI.
FEATURES
NVIDIA GPU, Networking and AI Software Technologies
NVIDIA Rubin Architecture
Featuring a specialised multi-agent engine for reasoning workflows and a dedicated reinforcement learning engine that optimises memory movement in hardware.
Massive Memory Bandwidth
The DGX Rubin NVL8's eight GPUs are supported by 2.3TB of HBM4 memory with 160TB/s of bandwidth, a 158% increase over previous generation Blackwell systems.
NVIDIA NVLink
Sixth-generation NVLink provides 28.8TB/s of bandwidth between GPUs, double that of 5th-gen NVLink in previous generation Blackwell systems.
NVIDIA ConnectX Networking
Support for up to 800Gb/s OSFP InfiniBand network ports for ultra-fast data ingest and connecting multiple DGX Rubin NVL8s together in a POD.
NVIDIA AI Software Stack
A complete AI software stack including access to NVIDIA NIM microservices, Blueprints and AI Enterprise.
Accelerating Operations
NVIDIA Mission Control software accelerates every aspect of operations. From configuring integrating with facilities, to managing clusters, workloads, cooling and power events.
Agentic and Physical AI Made Simple
NVIDIA AI Enterprise, included with the DGX Rubin NVL8 is a suite of frameworks and libraries, including NVIDIA NIM and NeMo microservices, that accelerate the development and deployment of AI projects. Leveraging pre-configured frameworks removes many of the manual tasks and complexity associated with software development, as each framework is tried, tested and optimised by NVIDIA. The less time spent developing, the greater the ROI on your AI hardware and data science investments.
Rather than trying to assemble thousands of co-dependent libraries and APIs from different authors when building your own AI applications, NVIDIA AI Enterprise removes this pain point by providing a full agentic AI software stack.
NVIDIA DGX Rubin NVL8 Workloads
| NVIDIA DGX Rubin NVL8 Specifications | ||
|---|---|---|
| GPU | 8× NVIDIA Rubin GPUs | |
| Total GPU Memory | Bandwidth | 2.3TB HBM4 | 160 TB/s | |
| Performance |
NVFP4 Inference: 400 PFLOPS NVFP4 Training: 280 PFLOPS* FP8/FP6 Training: 140 PFLOPS* INT8 Tensor Core: 2 PFLOPS* FP16 / BF16 Tensor Core: 32 PFLOPS* TF32 Tensor Core: 16 PFLOPS* FP64 Tensor Core: 264 TFLOPS |
|
| CPU | 2× Intel Xeon 6776P processors | |
| NVIDIA NVLink Switch System | 4× (6th gen) | |
| NVIDIA NVLink Bandwidth | 28.8 TB/s total bandwidth | |
| System Power Usage | ~24 kW | |
| Networking |
8× OSFP ports serving 8× single-port NVIDIA ConnectX-9 VPI — up to 800 Gb/s NVIDIA InfiniBand and Ethernet 2× 400G QSP112 NVIDIA BlueField-4 DPUs — up to 800 Gb/s NVIDIA InfiniBand and Ethernet |
|
| Operating System | NVIDIA DGX OS, Ubuntu, Red Hat Enterprise Linux, Rocky | |
*Dense specification.
Frequently Asked Questions
The NVIDIA DGX Rubin NVL8 is a powerful AI supercomputer designed for agentic and physical AI at scale. It features eight NVIDIA Rubin GPUs, 2.3TB of HBM4 system memory, and 800GB/s high-speed networking. It includes the full NVIDIA AI software stack (additional charge may apply) and is the building block for the upcoming Vera Rubin NVL72 SuperPOD solutions.
The NVIDIA DGX Rubin NVL8 is expected to be available to purchase in the second half of 2026.
Register your interest here.
The next-generation Rubin platform includes the Rubin GPU (HBM4 memory), NVLink 6 Switch, ConnectX-9 SuperNIC, BlueField-4 DPU, and Spectrum-6 Ethernet / Quantum X800 InfiniBand switch. In contrast the current generation Blackwell includes the Blackwell and Blackwell Ultra GPUs (HBM3e memory), NVLink 5 Switch, ConnectX-8 SuperNIC, BlueField-3 DPU, and Spectrum-4 Ethernet / Quantum-2 InfiniBand switch.
Every aspect of the latest Rubin platform features an uplift in performance over the previous generation, resulting in significantly higher performance than Blackwell-based ones.
The DGX Rubin NVL8 is optimised for demanding AI workloads such as Agentic AI, Physical AI, large-scale Mixture-of-Experts (MoE) Models, Massive-Context Inference and next-generation AI factories.
Yes. Scan is an official UK NVIDIA Elite Partner and will be offering the DGX Rubin NVL8 with full configuration, deployment, and support services. Scan is the only UK-based DGX Managed Services Provider (MSP) that enables businesses to access the power of DGX infrastructure without needing the skillset on-site to oversee the infrastructure. Instead, DGX MSPs provide full wraparound services to assist with anything from the initial deployment & configuration through to on-going services that encompass support, updates and overall platform management.