NVIDIADGXB-G1440+P2CMI36

DGXB-G1440+P2CMI36 DGX B200 Server- 8x B200 GPUs - 3-Year Support

The NVIDIADGXB-G1440+P2CMI36 DGX B200 is an elite AI platform engineered for training massive AI models, running generative AI workloads, and delivering unmatched data center performance. It's built for research institutions, enterprises, and AI innovators looking to push boundaries in machine learning, deep learning, and simulation.

Features

Built with 8x next-gen NVIDIABlackwell B200 GPUs interconnected via NVLink, the DGX B200 delivers extraordinary computational throughput for transformer-based models, LLMs, and deep neural networks. It features dual 5th Gen Intel Xeon processors, ultra-fast system memory, and a purpose-built architecture to support GPU-intensive AI frameworks and parallel processing at scale.

Advanced Options

Whether you're scaling generative AI pipelines or accelerating time-to-insight for massive datasets, the DGX B200 provides a pre-integrated, fully optimized system designed for plug-and-play deployment in advanced data centers. NVIDIAsoftware stack, including Base Command and AI Enterprise Suite, enhances manageability and productivity for enterprise AI teams.

Product Key Features

• • Powered by 8x NVIDIABlackwell B200 Tensor Core GPUs

  • Breakthrough performance for AI training and inference workloads

  • Dual Intel Xeon CPUs for advanced system orchestration

  • 2TB DDR5 system memory and 8x 180GB GPU memory

  • Comes with 3-Year Standard Support for peace of mind

  • Optimized for large language models, generative AI, and HPC

---

Reliability & Performance

The DGX B200 sets the benchmark for reliable high-performance AI infrastructure. With an integrated NVLink switch and 900GB/s GPU-to-GPU bandwidth, it ensures near-linear scaling for multi-GPU workloads. Enhanced cooling, power efficiency, and redundant hardware design make it highly durable for 24/7 operations. Backed by a 3-year support plan, it ensures uninterrupted performance and mission-critical uptime.

Key Features :

• 8 × NVIDIABlackwell B200 GPUs with 180GB HBM3e each

• Dual Intel Xeon 8592+ CPUs (128 cores total)

• 2TB DDR5 system memory

• 900GB/s GPU interconnect via NVIDIANVLink

• 30TB NVMe storage (approx.)

• Liquid-assisted air cooling system

• 3-Year Standard Support included

• Built-in NVIDIABase Command & AI Enterprise support

• Designed for large-scale AI, LLMs, and HPC workloads

• Redundant power supplies and enterprise-grade components

Specification	Details
Model Number	DGXB-G1440+P2CMI36
Product Series	NVIDIADGX B200
GPU Configuration	8 × NVIDIAB200 Tensor Core GPUs (180GB HBM3e each)
CPU	2 × Intel Xeon 8592+ (128 total cores)
System Memory	2TB DDR5
GPU Interconnect	NVIDIANVLink Switch (900GB/s total bandwidth)
Storage	30TB NVMe SSD (4x drives typical)
Cooling System	Liquid-assisted air cooling
Management Software	NVIDIABase Command, AI Enterprise
Form Factor	10U Rackmount Server
Network Interfaces	Multiple 400GbE / Infiniband ports (varies by config)
Power Supply	Redundant high-efficiency PSU
Support Coverage	3-Year Standard Support