This document describes Compute Engine instances that use the
reservation-bound provisioning model, including their benefits and creation
requirements.

When you create a compute instance, you must specify the underlying
[provisioning model](https://docs.cloud.google.com/compute/docs/instances/provisioning-models), which defines
the availability, pricing, and lifespan for the resources that your compute
instance uses. The reservation-bound provisioning model lets you create compute
instances by using reserved capacity from a
[future reservation in calendar mode](https://docs.cloud.google.com/compute/docs/instances/future-reservations-calendar-mode-overview)
or a
[future reservation for capacity blocks](https://docs.cloud.google.com/compute/docs/instances/choose-reservation-type#reserve-capacity-aihc).

The reservation-bound provisioning model offers the following benefits:

- **Cost control**: you don't incur additional charges when you create
  compute instances by using reserved capacity. You only incur charges for
  resources that aren't part of your reservation, such as disks or IP
  addresses.

- **Lifecycle management**: based on the termination action that you specify
  when you create compute instances, Compute Engine stops or deletes
  the compute instances at the end of the reservation period.

## Understand instances that use the reservation-bound provisioning model

The following sections describe the requirements that you apply when you create
compute instances by using the reservation-bound provisioning model.

### Compute instance creation prerequisites

To use the reservation-bound provisioning model to create compute instances, you
must first reserve resources. You can reserve resources as follows:

- To reserve up to 80 A4, A3 Ultra, A3 Mega, A3 High with 8 GPUs, or A3 Edge
  instances, up to 256 H4D instances, or up to 1,024 TPU7x, TPU v6e, or TPU
  v5p chips for up to 90 days, use
  [future reservation requests in calendar mode](https://docs.cloud.google.com/compute/docs/instances/future-reservations-calendar-mode-overview).

- To reserve A5X, A4X Max, A4X, A4, A3 Ultra, A3 Mega, A3 High with 8 GPUs, A3
  Edge, H4D, TPU7x, TPU v6e, or TPU v5p instances for any duration or
  quantity, use
  [future reservations for capacity blocks](https://docs.cloud.google.com/compute/docs/instances/choose-reservation-type#reserve-capacity-aihc).

If Google Cloud approves your future reservation request, then
Compute Engine automatically creates (*auto-creates*) a reservation at
the start of your reservation period. You can then use the reservation to create
compute instances.

### Compute instance creation requirements

To create a compute instance by using the reservation-bound provisioning model,
you must specify the following configurations:

- **The compute instance and the reservation must have matching properties** .
  You can only use your reserved capacity to create instances if the instance
  and auto-created reservation properties *exactly* match. For more
  information, see the
  [requirements for consuming reservations](https://docs.cloud.google.com/compute/docs/instances/reservations-overview#requirements).

- **The compute instance must specifically target the reservation for consumption** .
  When you create a compute instance, you must specify the name of the
  auto-created reservation to target for consumption, as well as set the
  `reservationAffinity` field to `SPECIFIC_RESERVATION`. For more information,
  see
  [Consume a specifically targeted reservation](https://docs.cloud.google.com/compute/docs/instances/reservations-consume#consuming_instances_from_a_specific_reservation).

- **The compute instance must use the reservation-bound provisioning model**.
  When you create a compute instance, you must specify the reservation-bound
  provisioning model as follows:

  - If you use the Google Cloud console, then, in the **Provisioning model**
    list, select **Reservation-bound**.

  - If you use the Google Cloud CLI, then include the
    `--provisioning-model=RESERVATION_BOUND` flag in the command.

  - If you use the Compute Engine API, then include the
    `"provisioningModel": "RESERVATION_BOUND"` field in the request body.

- **The compute instance must be stopped or deleted at the reservation end time** .
  When you create a compute instance, you must specify whether to stop or
  delete the compute instance at the reservation's end time by using the
  `instanceTerminationAction` field. For more information, see how to
  [limit the run time of a compute instance](https://docs.cloud.google.com/compute/docs/instances/limit-vm-runtime#limit-runtime-new-vm).

After you create a compute instance by using the reservation-bound provisioning
model, the compute instance starts running and keeps running until you stop or
delete it, or until the Compute Engine stops or deletes the compute
instance at the reservation's end time.

## Quota

When you create a compute instance by using the reservation-bound provisioning
model, you don't need quota for the reserved resources that you use to create
the compute instance. You only need quota for the resources that aren't part of
your reserved capacity, such as disks and IP addresses. For more information
about the different types of quota, see
[Allocation quotas](https://docs.cloud.google.com/compute/resource-usage).

## Pricing

When you create a compute instance by using the reservation-bound provisioning
model, you incur charges as follows:

- **Charges start when you create the compute instance** . You don't incur
  additional charges for the reserved resources that you use to create your
  compute instance. You only incur charges for the resources that aren't part
  of the reservation, such as disks or IP addresses. For more information, see
  the
  [billing for reservations](https://docs.cloud.google.com/compute/docs/instances/reservations-overview#reservation-billing).

- **Charges stop at the reservation's end time**. At that time,
  Compute Engine deletes the reservation, and stops or
  deletes your compute instance based on the termination action that is
  specified in the compute instance.

  > [!IMPORTANT]
  > **Important:** After Compute Engine stops a compute instance and changes its state to `STOPPING`, you continue to incur charges for any resources that are attached to the compute instance, such as disks. To avoid unnecessary costs, detach and delete any resources that you no longer need, or delete the compute instance altogether. For more information, see the [pricing for a compute instance's uptime](https://docs.cloud.google.com/compute/vm-instance-pricing#instanceuptime).

## Limitations

To create compute instances by using the reservation-bound provisioning model,
you must use one of the following machine series or TPU versions:

- **High performance computing (HPC) machine series**:

  - H4D
- **GPU machine series**:

  - A5X

  - A4X Max

  - A4X

  - A4

  - A3 Ultra

  - A3 Mega

  - A3 High with 8 GPUs

  - A3 Edge

  > [!NOTE]
  > **Note:** To inquire about using other GPU machine series with the reservation-bound provisioning model, contact your [account team](https://cloud.google.com/tam) or the [sales team](https://cloud.google.com/contact).

- **TPU versions**:

  - TPU7x

  - TPU v6e

  - TPU v5p

  <br />

## What's next

- [Create an instance with the reservation-bound model](https://docs.cloud.google.com/compute/docs/instances/create-reservation-bound-instance)