Hopper-generation PCIe GPU with large HBM3e memory for LLM inference and HPC in air-cooled racks.
NVIDIA H200 NVL is the air-cooled, dual-slot PCIe version of the Hopper H200 GPU. It pairs 141 GB of HBM3e with 4.8 TB/s of bandwidth, giving up to 1.7x faster LLM inference and 1.3x faster HPC than H100 NVL.
Up to four GPUs can be linked with NVLink bridges at 900 GB/s, and a five-year NVIDIA AI Enterprise subscription is included, making it a strong fit for enterprise racks with standard air cooling.
Run larger models and datasets without sharding them across GPUs.
Feeds memory-bound LLM inference and HPC workloads.
2- or 4-way NVLink at 900 GB/s per GPU; up to 8 GPUs per server.
Five-year NVIDIA AI Enterprise subscription with NIM microservices.
Configurable up to 600 W within standard enterprise cooling.
Protects sensitive data and models while they are processed.
| FP64 | 30 TFLOPS |
|---|---|
| FP64 Tensor Core | 60 TFLOPS |
| FP32 | 60 TFLOPS |
| TF32 Tensor Core | 835 TFLOPS |
| BFLOAT16 / FP16 Tensor Core | 1,671 TFLOPS |
| FP8 / INT8 Tensor Core | 3,341 TFLOPS / TOPS |
| GPU memory | 141 GB HBM3e |
|---|---|
| Memory bandwidth | 4.8 TB/s |
| Interconnect | 2- or 4-way NVLink bridge, 900 GB/s per GPU; PCIe Gen5, 128 GB/s |
| Form factor | PCIe, dual-slot, air-cooled |
|---|---|
| Max TDP | Up to 600 W (configurable) |
| MIG | Up to 7 instances @ 16.5 GB |
| Decoders | 7 NVDEC, 7 JPEG |
| Confidential computing | Supported |
| Server options | NVIDIA MGX H200 NVL partner and NVIDIA-Certified Systems with up to 8 GPUs |
| NVIDIA AI Enterprise | Included |
Tensor Core figures include sparsity, as published by NVIDIA.
Universal data center GPU for agentic AI, physical AI, scientific computing, rendering and video.
Energy-efficient multi-workload accelerator for AI inference, data science, video and visual computing.
Ada Lovelace universal GPU for generative AI inference, graphics and Omniverse workloads.
Tell us your workload and timeline for the NVIDIA H200 NVL. We reply within one business day with a configuration, pricing and lead time.