The Gap Between Consumer and Enterprise GPUs Just Got Smaller on Nodexo (SN106)

Nodexo (SN106) rolled out an expanded GPU support matrix, adding new datacenter-tier cards and stretching the multi-GPU host counts on existing ones.

The headline changes cover H100 NVL, B200, and the RTX PRO 6000 Blackwell Server Edition amongst other configurations that run wider than the update itself calls out.

Providers running any of the affected configurations can register them starting immediately, with additional models and higher GPU counts landing after calibration runs complete.

What Just Changed on the Support Matrix

The update names configurations, each opening capacity that was previously unsupported or capped at lower host counts. The attached configuration list covers, amongst others:

1. H100 NVL: Newly supported with 1x-2x configurations now accepted.

2. B200: Extended to 1x-2x, up from 1x only.

3. RTX PRO 6000 Blackwell Server Edition: Extended to 1x-8x, up from 1x-4x.

Nodexo is quietly opening the high-end datacenter GPU and dense multi-GPU capacity enterprise inference workloads need most.

Why the Expansion Matters

The support matrix decides which operators can bring their hardware online. Widening it changes the economics for both sides of the network.

1. More real inventory becomes eligible: Operators sitting on H100 NVL or dense Blackwell Server hosts can now plug into Nodexo directly.

2. Multi-GPU support catches up with reality: Serious inference and training workloads run on 4-8 GPU nodes, the updated caps reflect that.

3. More configurations are coming: Calibration runs on additional models and higher GPU counts are next in the pipeline.

For providers with high-end datacenter hardware previously sitting outside the support matrix, this is the moment to register the new configurations and get inventory earning.

➛ Read More on What Nodexo Does Different Here:

Nodexo (SN106) And Targon (SN4) Run Trusted AI Compute. Here’s How They Differ.