NVIDIA DGX-1 3U Rack Server | 8×Tesla Referral V100 NVLink, Deep Learning AI Training Workstation, Multiple Configurations Available Item Condition

Original NVIDIA DGX-1 3U rack AI supercomputer, fully inspected and stability tested before shipment. All units retired from data centers with fully functional hardware. Every unit undergoes comprehensive strict testing prior to shipment, test data can be provided upon purchase. Feel free to message us for more detailed information. Purpose-built for deep learning training, computer vision, large model inference, HPC computing and GPU virtualization. Multiple complete configurations selectable via variations.

Warranty: 3‑month limited hardware warranty against functional failure.

Available Variations (Select via Variations)

• Variant 1：2×E5‑2698 V4 | 8×V100 32GB | No System RAM & No Storage
  • CPU: Dual Intel Xeon E5‑2698 V4 Processors
  • GPU: 8 × NVIDIA Tesla Referral V100 SXM2 32GB HBM2, Full NVLink Mesh Interconnect
  • System Memory: NOT Included
  • Storage Drive: NOT Included
• Variant 2：2×E5‑2698 V4 | 128GB System RAM | 8×V100 32GB | 480GB SSD
  • CPU: Dual Intel Xeon E5‑2698 V4 Processors
  • GPU: 8 × NVIDIA Tesla Referral V100 SXM2 32GB HBM2, Full NVLink Mesh Interconnect
  • System Memory: 128GB DDR4 Server Memory
  • Storage: 1 × 480GB Enterprise SSD
• Variant 3：2×E5‑2698 V4 | 256GB System RAM | 8×V100 32GB | 480GB SSD + 2×1.92TB SSD
  • CPU: Dual Intel Xeon E5‑2698 V4 Processors
  • GPU: 8 × NVIDIA Tesla Referral V100 SXM2 32GB HBM2, Full NVLink Mesh Interconnect
  • System Memory: 256GB DDR4 Server Memory
  • Storage: 1 × 480GB OS SSD + 2 × 1.92TB Enterprise SSD
• Variant 4：2×E5‑2698 V4 | 512GB System RAM | 8×V100 32GB | 480GB SSD + 4×1.92TB SSD
  • CPU: Dual Intel Xeon E5‑2698 V4 Processors
  • GPU: 8 × NVIDIA Tesla Referral V100 SXM2 32GB HBM2, Full NVLink Mesh Interconnect
  • System Memory: 512GB DDR4 Server Memory
  • Storage: 1 × 480GB OS SSD + 4 × 1.92TB Enterprise SSD
• Variant 5：2×E5‑2698 V4 | 8×V100 16GB | No System RAM & No Storage
  • CPU: Dual Intel Xeon E5‑2698 V4 Processors
  • GPU: 8 × NVIDIA Tesla Referral V100 SXM2 16GB HBM2, Full NVLink Mesh Interconnect
  • System Memory: NOT Included
  • Storage Drive: NOT Included

Core Advantages

NVIDIA DGX‑1 dedicated AI supercomputer, integrated 8‑way Tesla Referral V100 SXM2 GPU architecture with native NVSwitch full‑mesh NVLink, ultra‑high GPU‑to‑GPU bandwidth, eliminating PCIe bottlenecks for multi‑GPU distributed training.

Dual Intel Xeon E5‑2698 V4 dual‑socket platform, sufficient CPU performance to coordinate large‑scale deep learning frameworks, data preprocessing and task scheduling.

Volta V100 equipped with Tensor Cores, delivers excellent acceleration for AI model training, image recognition, natural language processing, speech synthesis, 3D rendering and HPC simulation tasks.

Supports Ubuntu, CentOS, NVIDIA CUDA, cuDNN, TensorFlow, PyTorch, MXNet and all mainstream deep learning software stacks.

3U rack‑mount chassis, built‑in redundant high‑power PSU, onboard 10GbE Ethernet and InfiniBand interface options for cluster deployment.

Compatibility Reminder

‑ Please carefully select your target configuration via variations before checkout.

Important Note: Extra accessories are not included by default, send us message for extra‑part quotation.

What You Will Receive

NVIDIA DGX‑1 3U Rack AI Server according to your selected variation: Chassis, Dual E5‑2698 V4 CPUs, 8×Tesla Referral V100 SXM2 GPUs, matched memory and storage as listed in your chosen variant.

If you need additional accessories, please contact us for bundled quotation.

Quality Assurance

1. Each DGX‑1 unit passes power‑on, hardware detection and multi‑hour GPU stability testing before shipping.
2. 3‑month warranty only covers hardware functional failure; damage caused by improper installation, misuse or overclocking is not covered.
3. Due to cross‑border customs constraints, returns are not supported. If hardware failure occurs within warranty period, we provide spare‑part replacement service.
4. In case of transit damage, please provide photos of outer package and goods for after‑sales support.

Shipping Information

1. Ship out within 5‑7 working days after payment confirmation, tracking number will be provided after dispatch.
2. Item price does NOT include shipping cost; buyer shall pay all freight fees before shipment.
3. The displayed shipping cost is air‑freight price for United States customers; air / sea / land shipping options are available for other regions, please contact customer service for quotation.
4. Shipping service is unavailable to Brazil due to high local tariff policy.
5. All import duties and customs fees generated in destination country shall be fully borne by buyer.

Please select your preferred configuration via variations. Message us if you have any pre‑sale questions!