NVIDIA DGX Rubin NVL8 turnkey AI system with 8 Rubin GPUs, 2.3 TB HBM4, sixth-generation NVLink, ConnectX-9 networking, and BlueField-4 DPUs for training and agentic AI.

Ground, Overnight

Fast Shipping

In Stock

100k+ SKUs

Sales and Support

8/5 Support

New / Refurbished

Certified

NVIDIA DGX Rubin NVL8 8-GPU Full Server and AI System

SKU: DGX-RUBIN-NVL8 Categories: , Brand:

⭐ Top Rated Verified Buyers
We are trusted for quality, reliability and fast shipping 💯
REQUEST A QUOTE
0 People watching this product now!
Description

NVIDIA DGX Rubin NVL8 | Turnkey Server System

NVIDIA DGX Rubin NVL8 is a turnkey eight-GPU AI system built on the NVIDIA Rubin architecture for training, post-training, reasoning, and agentic inference. It combines eight Rubin GPUs with sixth-generation NVLink, high-speed ConnectX-9 scale-out networking, BlueField-4 DPUs, and the NVIDIA DGX software and management stack.

Availability / RFQ: Exact commercial orderability, allocation, production configuration, and lead time should be confirmed before project planning is finalized. Contact Catalyst Data Solutions to discuss current status, compatibility, configuration requirements, and sourcing options.
End-user / export compliance notice: Advanced AI accelerators and systems may be subject to applicable U.S. export controls, end-user/end-use restrictions, destination requirements, and OEM/channel policies. Catalyst may request purchaser, final end-user, intended-use, and deployment-location information before a quote or fulfillment can be approved. An RFQ does not guarantee allocation or shipment.

Technical Specifications

SKU DGX-RUBIN-NVL8
Product Type 8-GPU DGX AI System
GPU Configuration 8x NVIDIA Rubin GPUs
Total GPU Memory 2.3 TB HBM4
GPU Memory Bandwidth 176 TB/s aggregate
CPU 2x Intel Xeon 6776P processors
NVLink Switch Bandwidth 28.8 TB/s total
Networking 8x ConnectX-9 VPI up to 800 Gb/s; 2x dual-port BlueField-4 DPUs
System Power Approximately 24 kW (preliminary)
Availability Status Contact for Current Availability

Deployment and Compatibility Considerations

  • Treat facility power and cooling as part of the server qualification because the preliminary system power is approximately 24 kW.
  • Select scale-out Ethernet or InfiniBand fabric, optics, cables, and switch ports to match ConnectX-9 link design.
  • Validate storage throughput for model checkpoints, datasets, inference context, and data pipelines.
  • Confirm DGX OS, supported Linux distribution, management, support, and cluster integration requirements.

Common Use Cases

  • Enterprise and research AI training
  • Post-training and reinforcement learning
  • Agentic AI and reasoning inference
  • High-performance data analytics
  • Multi-node DGX cluster and SuperPOD deployments

Frequently Paired Infrastructure

  • Spectrum-X Ethernet or Quantum-X800 InfiniBand
  • OSFP 800G optics and LinkX cables
  • BlueField-4 DPUs
  • High-throughput AI storage
  • Data center liquid-cooling and power infrastructure as required

Official NVIDIA Documentation

NVIDIA DGX Rubin NVL8 Datasheet

Frequently Asked Questions

How many Rubin GPUs are in DGX Rubin NVL8?

The system uses eight NVIDIA Rubin GPUs interconnected through sixth-generation NVLink.

Is DGX Rubin NVL8 the same as HGX Rubin NVL8?

No. HGX is the underlying eight-GPU platform used by system manufacturers, while DGX Rubin NVL8 is NVIDIA’s integrated DGX system with NVIDIA system software, networking, management, and support.

Additional information
product typeGPU Server / DGX AI System
Platform / ArchitectureNVIDIA Rubin / DGX
Accelerator Configuration8 Rubin GPUs / 2.3 TB HBM4
NetworkingConnectX-9 + BlueField-4
Availability StatusContact for Current Availability
About brand
NVIDIA is the leader in GPU computing and AI infrastructure for high-performance workloads.

Recently Viewed Products