NVIDIA H200 NVL GPU
NVIDIA H200 NVL as an individual PCIe card, 141 GB HBM3e at 4.8 TB/s, NVLink bridge for 2 or 4 GPUs, up to 600 W, for validated servers with registration.
GPU: NVIDIA H200 NVL
- Memory
- 141 GB HBM3e, 4.8 TB/s
- Compute
- Hopper, 3,341 TFLOPS sparse FP8, up to 7 MIG instances
- Interconnect
- 2 way or 4 way NVLink bridge at 900 GB/s per GPU, PCIe Gen5 x16
- Power
- Up to 600 W configurable
- Form factor
- Dual slot full height full length, passive
- Registration
- NVIDIA end user registration and export controls apply
Overview
H200 NVL is supplied as an individual card for customers who expand an existing H200 NVL server, add a second bridged pair, or replace a unit in service. Each card carries 141 GB of HBM3e at 4.8 TB/s, and NVLink bridges join two or four cards at 900 GB/s per GPU so that large models span the bridged group without crossing PCIe.
The card is passive and draws up to 600 W, so it only belongs in servers validated for that class. Bridge kits are specific to the slot pitch of the server, and we supply the correct kit with the cards. Supply follows NVIDIA allocation, which is why the card is listed as available on request and timing is confirmed with each quotation.
Data center GPUs are subject to NVIDIA end user registration and US export controls. We collect end user and deployment country details with the quotation and complete the registration with the order so that shipment is not delayed, and we can advise on the server models that are validated for the card before you commit.
Use cases
- Expanding H200 NVL servers
- Adding bridged pairs
- Fleet replacements
- Validated server upgrades
Specifications
| Architecture | NVIDIA Hopper |
|---|---|
| Memory | 141 GB HBM3e, 4.8 TB/s |
| Tensor performance | 3,341 TFLOPS FP8, 1,671 TFLOPS FP16 and BF16, with sparsity |
| Multi Instance GPU | Up to 7 instances of 16.5 GB |
| Form factor | Dual slot FHFL, passive |
|---|---|
| Cooling | Requires server airflow validated for 600 W PCIe GPUs |
| Power | Up to 600 W configurable |
| Interface | PCIe Gen5 x16, 128 GB/s |
| NVLink bridge | 2 way or 4 way, 900 GB/s per GPU |
|---|---|
| Bridge kits | Supplied to match the server slot pitch |
| Server options | MGX H200 NVL partner systems and NVIDIA certified systems with up to 8 GPUs |
Figures describe the system class. Exact configuration, weights and power draw are confirmed on the quotation.
What's included
- Original manufacturer part in retail or bulk packaging
- Compatibility check against your server or workstation model
- Manufacturer warranty passed through, advance replacement options
- NVLink bridge kit matched to the server
- NVIDIA end user registration handled with the order
Compliance
Related
Related systems
8 GPU H200 NVL PCIe AI server
Available on request- GPUs
- 8 x NVIDIA H200 NVL, 141 GB HBM3e each
- CPU
- 2 x AMD EPYC 9005 or 2 x Intel Xeon 6
- Memory
- Up to 2 TB DDR5 RDIMM
- Chassis
- 4U, about 6.5 kW, air cooled
RTX PRO 6000 Blackwell Server Edition GPU
- Memory
- 96 GB GDDR7 with ECC, 1.6 TB/s
- Compute
- 24,064 CUDA cores
- Power
- 400 W to 600 W configurable
- Form factor
- Dual slot full height full length, passive
2 GPU 2U inference server
- GPUs
- 2 x RTX PRO 6000 Blackwell Server Edition (96 GB)
- CPU
- 1 or 2 x AMD EPYC 9005, or 2 x Intel Xeon 6
- Memory
- Up to 1.5 TB DDR5 RDIMM
- Chassis
- 2U, about 2 kW, air cooled
Related guides
Buying data center GPUs, export and registration
The compliance steps that come with every HGX, H200 NVL, Server Edition and NVL72 purchase, explained so your order ships on time.
In this guide
- Why a GPU purchase comes with paperwork
- Which products are covered
Choosing a GPU server for LLM inference
A practical method for sizing an inference server from the model, the quantization and the concurrency you expect, with worked examples.
In this guide
- Start with the model, not the server
- Step 1: how much GPU memory the weights need
HGX vs PCIe GPU servers
The real differences between an HGX node and a PCIe GPU server, and a simple rule for choosing between them.
In this guide
- Two kinds of 8 GPU server
- The interconnect is the point
Compare it with the next class up or down
The nearest classes we quote are the 8 GPU H200 NVL PCIe AI server and the RTX PRO 6000 Blackwell Server Edition GPU. If neither is quite right, tell us the workload and we specify the system between them.
- (858) 613-1818
- sales@computermodules.com
- Monday to Friday, 8am to 5pm Pacific