Product
NVIDIA H100 80GB PCIe
Part number NVIDIA H100 80GB PCIe
- Price
- USD $100
- Stock
- 1
Hopper-architecture PCIe Gen5 accelerator with 80 GB of HBM2e memory and a 350 W envelope, suited to mainstream servers and one- or two-GPU inference nodes.
- 80 GB HBM2e memory at 2,000 GB/s
- Dual-slot PCIe Gen5 card with a passive heatsink
- Maximum board power of 350 W
- Optional NVLink Bridge at 600 GB/s between two cards
- The same Hopper Tensor Core and MIG features as H100 SXM, in a lower-power card
Specifications
| GPU Model | NVIDIA H100 80GB PCIe |
|---|---|
| Architecture | NVIDIA Hopper |
| Form Factor | PCIe Gen5 dual-slot |
| GPU Count | 1 |
| GPU Memory | 80 GB HBM2e |
| Memory Bandwidth | 2,000 GB/s |
| Interconnect | NVLink Bridge 600 GB/s (two GPUs); PCIe Gen5 128 GB/s |
| Max TDP | 350 W |
| Cooling | Passive, dual-slot |
| Mechanical | Full-height, 10.5-inch PCIe card |
| Host Interface | PCI Express Gen5 |
| Multi-Instance GPU | Up to 7 MIG instances |
| Typical Host | Partner and NVIDIA-Certified Systems with 1–8 PCIe GPUs |
Description
The NVIDIA H100 80GB PCIe is the add-in-card member of the Hopper H100 family. It brings the same Tensor Core architecture into a dual-slot, full-height, 10.5-inch PCIe Gen5 board that fits mainstream accelerated servers instead of an SXM baseboard.
The card uses 80 GB of HBM2e memory with a peak bandwidth of 2,000 GB/s. NVIDIA documents a maximum thermal design power of 350 W and a passive heatsink that depends on system airflow. Two adjacent cards can be joined with NVLink Bridges at 600 GB/s, which is useful for workloads that scale to a pair of GPUs but do not need an eight-GPU NVSwitch fabric.
Compute features match the H100 generation, including the Transformer Engine, FP8 Tensor Cores, and Multi-Instance GPU. Because the power envelope is lower than the SXM5 module, this card is a practical choice for inference, smaller fine-tuning jobs, and HPC applications that run well on one or two GPUs in a standard rack.
If you need the highest memory bandwidth and a full NVLink/NVSwitch topology, see the H100 80GB SXM5 module and the GIGABYTE G593-SD0 HGX H100 server. Request a quote for the PCIe card on its own or with a qualified host.
Related products
NVIDIA H100 80GB SXM5
P/N NVIDIA H100 80GB SXM5
Price USD $100 Stock 1
Hopper-architecture SXM accelerator with 80 GB of HBM3 memory, fourth-generation NVLink, and a configurable TDP up to 700 W, built for dense HGX training and inference nodes.
- GPU Model
- NVIDIA H100 80GB SXM5
- Form Factor
- SXM5
- GPU Count
- 1
- GPU Memory
- 80 GB HBM3
NVIDIA H200 NVL 141GB
P/N NVIDIA H200 NVL
Price USD $100 Stock 1
Hopper PCIe accelerator with 141 GB of HBM3e memory at 4.8 TB/s, a dual-slot air-cooled form factor, and a configurable TDP up to 600 W, for MGX and NVIDIA-Certified servers.
- GPU Model
- NVIDIA H200 NVL
- Form Factor
- PCIe dual-slot air-cooled
- GPU Count
- 1
- GPU Memory
- 141 GB HBM3e
Ready to specify a configuration?
Share the use case, the target build, quantity, destination country, and timeline. We will take it from the inquiry.