> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.ibee.co.in/docs/infrastructure/gpu-vms/create-a-gpu-vm/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.ibee.co.in/_mcp/server.

# Create a GPU VM

> Deploy a GPU-enabled virtual machine on IBEE Solutions for AI, ML, rendering, and compute-intensive workloads.

Deploy one or more GPU-enabled virtual machines (VMs) from the IBEE Solutions portal. GPU VMs include NVIDIA GPUs for AI/ML training, inference, rendering, and other GPU-accelerated workloads. The **Deploy** page is a single scrolling form with a running summary and the **Deploy** button in the bottom bar.

## Before you begin

* An IBEE Solutions account with write access to the workspace.
* A verified account to unlock server deployment ([Verify your identity](/docs/getting-started/account-setup/verify-your-identity)).
* Sufficient wallet credits — the platform checks your balance before deploying ([Billing and usage](/docs/platform-fundamentals/billing-and-usage)).

## Deploy a GPU VM

### 1. Open GPU VMs

In the portal sidebar, open **Infrastructure → GPU VMs**, then click **Deploy GPU VM**.

### 2. Select a location

Under **Location**, select a region tab and choose a data center location within it. GPU availability varies by location.

### 3. Choose a GPU plan

Under **Choose Plan**, select a plan. Each row lists the **GPU** model and its **Memory**, plus **vCPU**, **RAM**, **Disk**, and the price per month and per hour. GPU VMs are billed **Hourly**.

### 4. Select the operating system

Pick a source for the boot disk under **Operating System**:

| Tab           | Use it for                                                                                                                       |
| ------------- | -------------------------------------------------------------------------------------------------------------------------------- |
| **Templates** | OS images maintained by IBEE. The default. Only GPU-compatible images are listed.                                                |
| **ISOs**      | Boot from an ISO you added under **Tools → ISOs**. Interactive installs need a console, which isn't available in the portal yet. |
| **Snapshots** | Create the VM from a snapshot captured from another VM.                                                                          |
| **Backups**   | Create the VM from a scheduled backup recovery point.                                                                            |

On **Templates**, pick an OS family card, then choose a version from the **OS Version** dropdown.

### 5. Add SSH keys (Linux)

For Linux images, the **SSH Keys** section appears (it's hidden for Windows images). Open **Select SSH Keys** and pick one or more saved keys, or click **Add Key** to create a key pair or import a public key. A Linux VM deployed without a key uses password login, shown under **Overview → Connection Details**. See [SSH Keys](/docs/tools/ssh-keys).

### 6. Configure the network

Under **Network Configuration**, choose **Internet and VPC access**:

* **Public IP** *(default)* — direct inbound and outbound IPv4. Optionally pick a VPC and subnet; with a VPC you can also pick a **Reserved public IP** (required for a private VPC).
* **Private + NAT** — primary interface in a NAT Gateway VPC with outbound internet only.
* **Private only** — primary interface in a VPC with no internet route.

Under **Firewall**, pick a firewall group or click **Create Group**. If you don't pick one, the default firewall group is attached automatically.

See [Networking for VMs](/docs/infrastructure/cloud-vms/networking-for-vms) for details — the options are the same as for Cloud VMs.

### 7. Set VM details

Under **VM Details**:

* **Instances** — deploy up to **5** VMs at once. The maximum can be lower when GPU capacity or public addresses in the location are limited, and it's **1** when you pick a specific Reserved IP.
* **Hostname** — use lowercase letters, numbers, and hyphens. With more than one instance, this becomes the base hostname and each instance is prefilled as `<base>-1`, `<base>-2`, and so on; edit them under **Instance hostnames**. Every hostname must be unique.

### 8. Deploy

Review the summary in the bottom bar, then click **Deploy** (or **Deploy N VMs**). The platform first checks your organization's billing state and wallet balance. If the check passes, provisioning starts and you return to the GPU VMs list. If not, a **Billing needs attention** popup appears — click **Add Credits** to top up and return.

## Verify

Once provisioned, your GPU VM appears in the **GPU Virtual Machines** list with status **Running**. Click it to open the detail page. **Overview → Connection Details** shows the SSH command (or username and password, or RDP details for Windows).

Confirm GPU access inside the VM:

```bash
nvidia-smi
```

If `nvidia-smi` isn't found, install a compatible NVIDIA driver for your OS inside the VM.

## Manage your GPU VM

From the VM detail page, use **Actions** to control your server:

| Action  | Description                                                                                                            |
| ------- | ---------------------------------------------------------------------------------------------------------------------- |
| Start   | Power on a stopped server                                                                                              |
| Stop    | Shut down a running server                                                                                             |
| Reboot  | Restart the server                                                                                                     |
| Console | Shown as **Soon** — browser console access isn't available in the portal yet                                           |
| Delete  | Permanently delete the server — you type the VM name to confirm, and can choose to keep its public IP as a Reserved IP |

The detail page has these tabs:

| Tab                         | What it's for                                                                                                                                                                                                                                                                                                         |
| --------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Overview**                | Specs, network, and **Connection Details**.                                                                                                                                                                                                                                                                           |
| **Monitoring**              | CPU, memory, storage, disk, and network charts, plus recent activity. See [Monitoring](/docs/infrastructure/cloud-vms/monitoring).                                                                                                                                                                                    |
| **GPU Metrics**             | GPU utilization, VRAM, temperature, and power. Click **Enable monitoring** first — see [GPU Metrics](/docs/infrastructure/gpu-vms#gpu-metrics).                                                                                                                                                                       |
| **Backups** / **Snapshots** | Recovery points. See [Backups](/docs/tools/backups) and [Snapshots](/docs/tools/snapshots).                                                                                                                                                                                                                           |
| **Settings**                | **Change Plan**, **IPv4 & Reserved IP**, **VPC Networks**, and **Access** (SSH keys, password login, password reset). **Change OS** and **Hostname** aren't available yet. See [Resize a VM](/docs/infrastructure/cloud-vms/resize-a-vm) and [Networking for VMs](/docs/infrastructure/cloud-vms/networking-for-vms). |

> **Warning**
>
> Deleting a GPU VM is permanent and cannot be undone. Back up your data and snapshots before deleting.

## Troubleshooting

**Deploy shows an error message**
The portal validates the form when you click **Deploy** and names the missing item — for example a plan, an OS version, a subnet, or a unique hostname for each instance.

**Billing needs attention popup**
Your organization's billing state or wallet balance blocks the deploy. Click **Add Credits** to top up, then return.

**`nvidia-smi` not found**
The NVIDIA driver isn't installed on the selected image. Install a compatible NVIDIA driver inside the VM. The **GPU Metrics** tab also shows **NVIDIA guest driver required** until it's installed.

**Server stuck in provisioning**
Wait a few minutes. If the status does not update, contact [IBEE support](/docs/support/contact-support).

## Related pages

* [GPU VMs overview](/docs/infrastructure/gpu-vms)
* [Networking for VMs](/docs/infrastructure/cloud-vms/networking-for-vms)
* [Power actions](/docs/infrastructure/cloud-vms/power-actions)
* [SSH Keys](/docs/tools/ssh-keys)

> Deploy a GPU-enabled virtual machine on IBEE Solutions for AI, ML, rendering, and compute-intensive workloads.