Photos supplied with quotation
NVIDIA
NVIDIA H100 80GB
Hopper-architecture data-center GPU for large-scale AI training and inference. PCIe and SXM variants.
| Availability | In stock |
|---|---|
| SKU | AXL-GPU-H100 |
| Minimum order | 1 unit |
| Condition | New |
| Pricing | By quotation |
Overview
The H100 is the workhorse accelerator for large language model training and high-throughput inference. We source PCIe and SXM5 variants through verified channels, with serial numbers documented on every quotation.
Available as individual cards, HGX baseboards, or integrated into validated server platforms.
Specifications
| Architecture | NVIDIA Hopper |
|---|---|
| Memory | 80 GB HBM3 |
| Bandwidth | Up to 3.35 TB/s |
| Variants | PCIe 5.0 or SXM5 |
|---|---|
| NVLink | Up to 900 GB/s |
Request a quote
Quotation for NVIDIA H100 80GB. We reply with a formal proforma invoice, typically within 24 hours.