Skip to content
Launch imminent

Official date to be announced shortlyGPU offers available at launch; figures shown are indicative.

Priority access: sign up

What is NVLink? GPU-to-GPU interconnect explained

NVLink is NVIDIA's direct link between GPUs. Learn why it matters for multi-GPU training and how it differs from PCIe.

Glossary3 sections1 min read

What is NVLink?

NVLink is NVIDIA's high-speed point-to-point interconnect between GPUs. It lets GPUs exchange data directly, far faster than over the PCIe bus.

Why it matters when renting a GPU

Multi-GPU training and tensor-parallel inference exchange data between GPUs at every step. A fast link keeps GPUs busy instead of waiting.

SXM boards use NVLink natively; PCIe cards may offer a bridge between pairs only. Check the interconnect line on the GPU page.

Worked example

The H100 SXM lists NVLink 4 (900 GB/s) · PCIe 5.0, while the H100 PCIe lists NVLink bridge (600 GB/s) · PCIe 5.0. The B200 moves to NVLink 5 (1.8 TB/s).

Frequently asked questions

Do I need NVLink?
For a single GPU, no. For training or serving a model split across several GPUs, NVLink usually speeds up communication considerably.

Be ready on launch day

Create your account now: you can top up your balance and launch a GPU as soon as the market opens.