> ## Documentation Index
> Fetch the complete documentation index at: https://docs.veri.studio/llms.txt
> Use this file to discover all available pages before exploring further.

# Azure

> Single-GPU A100/H100 serving on Azure — wired but feature-gated; creates return a 400 until the gate lifts.

Azure serving is wired but currently feature-gated: a `provider="azure"` create returns a "coming soon" 400 until the gate lifts.

## Planned serving shapes

| `gpu_type`  | Memory | GPUs per node | Status      |
| ----------- | ------ | ------------- | ----------- |
| `A100-80GB` | 80 GB  | 1             | Coming soon |
| `H100-80GB` | 80 GB  | 1             | Coming soon |

Azure is one of the few clouds with stable on-demand single-GPU A100 *and* H100 VMs (the H100 SKU is an H100 NVL 94 GB — a superset of the 80 GB request). Azure is serving-only: a training submit with `provider="azure"` is always rejected.

Watch `veri gpu list` — the `(soon)` marker on the serving column disappears when the gate lifts.

## Where to go next

<CardGroup cols={2}>
  <Card title="All serving providers" icon="server" href="/deployments/providers">
    Compare shapes, status, and the live CLI catalog.
  </Card>

  <Card title="GCP" icon="google" href="/deployments/providers/gcp">
    Single-GPU A100/H100 that's available today.
  </Card>
</CardGroup>