NVIDIADGXH-G1128+P2CMI36

DGX H200 System - 8x H200 GPUs - 2TB RAM - 3-Year Support ServerSystem

The NVIDIADGXH-G1128+P2CMI36 DGX H200 System is a next-generation AI compute platform designed for enterprises, research institutions, and AI innovators pushing the boundaries of generative AI and machine learning. Featuring 8 Hopper H200 GPUs, massive memory, and a tightly integrated architecture, it's optimized for AI at scale.

Features

The DGX H200 builds on NVIDIAlegacy of AI excellence with 8 advanced H200 Tensor Core GPUs, each equipped with cutting-edge HBM3e memory. This powerhouse delivers massive parallel processing and memory bandwidth, ideal for large language models, deep learning, and complex simulation workloads.

Advanced Options

Dual Intel Xeon processors and 2TB of DDR5 memory ensure high throughput for data ingestion and preprocessing. Combined with NVIDIABase Command and AI Enterprise software, the system provides a complete, ready-to-deploy AI infrastructure that reduces setup complexity and accelerates time-to-insight for organizations operating at the frontier of innovation.

Product Key Features

• Powered by 8x NVIDIAHopper H200 GPUs with HBM3e memory

• Dual Intel Xeon CPUs and 2TB DDR5 system memory

• Built for LLMs, generative AI, and deep learning at scale

• NVLink interconnect for 900GB/s GPU-to-GPU bandwidth

• Includes 3-Year Business Standard Support

• Ideal for enterprise AI infrastructure and model training

---

Reliability & Performance

The DGX H200 is engineered for non-stop, data center-grade performance. With integrated NVLink for ultra-fast GPU communication, redundant power and cooling systems, and remote manageability, it ensures sustained operation even under extreme AI loads. The included 3-Year Business Standard Support guarantees enterprise-level service and peace of mind.

Key Features :

• 8 × NVIDIAH200 Hopper GPUs with HBM3e memory

• Dual Intel Xeon Scalable CPUs (5th Gen, up to 128 cores)

• 2TB DDR5 ECC system memory

• 900GB/s total NVLink GPU interconnect bandwidth

• Pre-configured with NVIDIABase Command and AI Enterprise

• 30TB+ NVMe high-performance storage

• 3-Year Business Standard Support included

• 10U rackmount chassis with redundant power supplies

• Designed for generative AI, LLMs, and simulation workloads

• Integrated cooling and system monitoring

Specification	Details
Model Number	DGXH-G1128+P2CMI36
Product Series	NVIDIA DGX H200
GPU Configuration	8 × NVIDIA Hopper H200 GPUs (HBM3e memory)
CPU	2 × Intel Xeon (5th Gen, up to 128 total cores)
System Memory	2TB DDR5 ECC
GPU Interconnect	NVLink - 900GB/s total bandwidth
Storage	~30TB NVMe SSD (4x drives typical)
Management Software	NVIDIA Base Command, AI Enterprise Suite
Form Factor	10U Rackmount Server
Cooling System	Liquid-assisted air cooling
Power Supply	Redundant high-efficiency PSUs
Support Coverage	3-Year Business Standard Support