# InsufficientInstanceCapacity: We currently do not have sufficient capacity

AWS error fixes · Updated 16 November 2026

**Short answer:** InsufficientInstanceCapacity means the Availability Zone you asked for has no free hosts for that instance type right now. It is not a quota on your account: retry in another AZ (another subnet), try a similar instance type, or wait. For workloads that must start, reserve capacity in advance with an On-Demand Capacity Reservation.

## What the error tells you

The full message is unusually helpful, so read it all:

```
An error occurred (InsufficientInstanceCapacity) when calling the
RunInstances operation: We currently do not have sufficient m5.large
capacity in the Availability Zone you requested (eu-west-2a). Our system
will be working on provisioning additional capacity. You can currently get
m5.large capacity by not specifying an Availability Zone in your request
or choosing eu-west-2b, eu-west-2c.
```

It names the instance type, the zone that is short, and often the zones that currently have capacity. Capacity is pooled per instance type per AZ, so the same request can succeed a minute later, in a different zone, or with a neighbouring type. It also hits `StartInstances`: a stopped instance is tied to its subnet's AZ, so starting it can fail even though it ran fine yesterday.

## Situations where it shows up

| Situation | Why capacity is tight |
|---|---|
| Starting a stopped instance | It must come back in the same AZ; you have no choice of zone |
| Older or very large types (previous generations, metal, large GPU sizes) | Smaller pools, less new hardware added |
| Cluster placement group | All instances must fit on closely located hardware |
| Auto Scaling group pinned to one subnet or one type | No fallback when that pool runs dry |
| Regional events or big launches by others | Short-lived spikes in demand for popular types |

## Ways to get running now

1. **Use another AZ.** Launch into a subnet in one of the zones the message suggests:

   ```
   aws ec2 run-instances --image-id ami-0123456789abcdef0 \
     --instance-type m5.large --subnet-id subnet-0bb1c79de3EXAMPLE \
     --count 1
   ```

2. **Change instance type.** A newer generation with the same shape (m6i, m7i, or Graviton m7g if your image is arm64) draws from a different pool. For a stopped instance:

   ```
   aws ec2 modify-instance-attribute \
     --instance-id i-0123456789abcdef0 \
     --instance-type Value=m6i.large
   aws ec2 start-instances --instance-ids i-0123456789abcdef0
   ```

3. **Retry with backoff.** Capacity frees up continuously; a retry a few minutes later often works.

4. **Placement groups:** launch every instance you need in a single request, and if a start fails, stop all instances in the group and start them together.

## Make it less likely to hurt

- **Spread Auto Scaling groups across several AZs and types.** A mixed instances policy, or attribute-based instance type selection, lets the group fall back automatically instead of stalling.

- **Reserve what must start.** An On-Demand Capacity Reservation holds capacity in a specific AZ for you, billed whether you use it or not:

  ```
  aws ec2 create-capacity-reservation --instance-type m6i.large \
    --instance-platform Linux/UNIX --availability-zone eu-west-2b \
    --instance-count 2
  ```

- **Think before stopping critical singletons.** Rebooting keeps the host; a stop/start gives it up and you may not get it back in that zone at a busy moment.

- **Watch your Auto Scaling activity history**, where failed launches are recorded; a group quietly below desired capacity is the usual symptom.

## Not to be confused with

This is not a limit on your account. If the error says you requested more vCPU capacity than your limit, that is a Service Quotas problem with its own fix: see [VcpuLimitExceeded](https://cloudconsole.dev/errors/ec2-vcpulimitexceeded). And if the type simply is not offered in that zone you get an "Unsupported" error instead; check offerings with `aws ec2 describe-instance-type-offerings --location-type availability-zone`.

## Questions

### How long does InsufficientInstanceCapacity last?

There is no fixed duration. Popular types often recover within minutes; scarce types such as large GPU instances can stay constrained for much longer, which is when capacity reservations earn their cost.

### Will a quota increase fix InsufficientInstanceCapacity?

No. Quotas cap what your account may run; this error is about physical capacity in one zone. Change AZ, change type or reserve capacity instead.

### Does Spot have the same problem?

Spot requests draw on spare capacity, so they are even more sensitive to shortages. Diversifying across many instance types and AZs is the main defence.

## When this happens away from your desk

An Auto Scaling group that cannot launch shows up as a capacity or latency alarm, not as this error. CloudConsole, a phone app for AWS, is built to put those alarms next to the resources behind them, so you can spot a group stuck below its desired count while you are away.

[About the CloudConsole app](https://cloudconsole.dev/aws-console-app) · [Download on the App Store](https://apps.apple.com/app/cloudconsole-for-aws/id6835419710)

## Sources

- [Troubleshoot Amazon EC2 instance launch issues](https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/troubleshooting-launch.html)
- [On-Demand Capacity Reservations](https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ec2-capacity-reservations.html)

## Related

- [VcpuLimitExceeded: You have requested more vCPU capacity than your current vCPU limit](https://cloudconsole.dev/errors/ec2-vcpulimitexceeded)
- [aws ec2 run-instances](https://cloudconsole.dev/cli/ec2-run-instances)
- [aws ec2 start-instances](https://cloudconsole.dev/cli/ec2-start-instances)
- [How to change an EC2 instance type](https://cloudconsole.dev/how-to/change-ec2-instance-type)
- [aws autoscaling update-auto-scaling-group](https://cloudconsole.dev/cli/autoscaling-update-auto-scaling-group)
- [aws ec2 modify-instance-attribute](https://cloudconsole.dev/cli/ec2-modify-instance-attribute)

---

Source: https://cloudconsole.dev/errors/ec2-insufficientinstancecapacity

CloudConsole is an independent iPhone and Android app for Amazon Web Services by Cloudconcur Ltd. It is not affiliated with Amazon Web Services.
