NVIDIA H200 NVL AI Accelerator Card 141GB HBM3E 4.8TB/s PCIe 5.0 x16 Passive Cooling 900-21010-0040-000

NVIDIA Hopper Architecture: Powered by H200 GPU, built for enterprise generative AI, LLM inference/fine-tuning and high-performance computing (HPC) workloads
High-Capacity High-Speed Memory: 141GB HBM3e GPU memory with 4.8TB/s bandwidth, 1.5X larger and 1.2X faster than H100 NVL
Interface & Cooling: PCI Express 5.0 x16 interface, passive cooling design, optimized for low-power air-cooled enterprise rack deployments
Blazing-Fast Interconnect: Supports NVIDIA NVLink™ 900GB/s, delivering 7X faster GPU-to-GPU communication than PCIe 5.0 for maximum system throughput
Power & Efficiency: 700W TDP, enhanced energy efficiency with the same power profile as H100, reducing data center total cost of ownership (TCO)
Authentic NVIDIA Specs: Tesla Hopper Series | Model: GPU-NVH200NV | Factory Part Number: 900-21010-0040-000
Enterprise Security & Flexibility: Supports Confidential Computing and Multi-Instance GPU (MIG) for multi-tenant deployments and data security

Specs

Any questions? Our AI beta will help you find out quickly.

Try me now

Learn more about the NVIDIA GPU-NVH200NV

Chipset
Chipset Manufacturer	NVIDIA
GPU	NVIDIA H200

Memory
Memory Size	141GB
Memory Type	HBM3e