AWS EC2 Terminated on Launch: InsufficientInstanceCapacity Fix

1. The Instance Type or Zone Is Out of Capacity

This is the most common reason. You pick an instance type like t3.micro in us-east-1a, but AWS doesn't have free hardware for that combo right now. It's not your account—it's AWS being out of stock, like a store that ran out of a popular item.

You'll see the instance go from pending to running to terminated in under 30 seconds. The console shows the reason: Client.InternalError: InsufficientInstanceCapacity. This happens most often during big launches (like a new AWS region opening up) or during peak hours (Monday mornings in US zones).

Fix: Change Availability Zone or Instance Type

1. Go to the EC2 console and click Launch Instance.
2. Under Network settings, click Edit next to the subnet. Pick a different Availability Zone in the same region. For example, if you were in us-east-1a, try us-east-1b or us-east-1c.
3. If that doesn't work, change the instance type. Go up or down one size: from t3.medium to t3.large, or switch families from t3 to t3a (AMD-based) or m5. AWS often has more capacity in different families.
4. Click Launch again. You should see it go to running and stay there.

After you apply the change, the instance should launch in under a minute. Give it 30 seconds, refresh the instance list, and you'll see a green running status. If it still fails, move to the next fix.

2. Your Account Has a Soft Launch Limit (New Accounts)

If you just created your AWS account (less than 3 months old), your default EC2 limits might be super low—like 2 instances total. AWS does this to prevent spam or abuse. You might hit this even if the zone has capacity.

The error message looks the same: InsufficientInstanceCapacity. But the real cause is your account's service quota is too small. You can check in the Service Quotas console.

Fix: Request a Quota Increase

1. Open the Service Quotas console (search for it in the AWS top bar).
2. Click AWS services and find Amazon EC2.
3. Look for Running On-Demand Standard (A, C, D, H, I, M, R, T, Z) instances. Click the row, then click Request quota increase.
4. Enter the new limit you want. For a small test environment, ask for 10. For production, calculate your max concurrent instances and add 20% headroom.
5. Fill in the use case description box with something honest and brief, like: "Need to run 5 t3.micro instances for a web app test environment. Expected usage is less than 10 instances per region."
6. Click Submit. Most requests are auto-approved in 5–15 minutes.

After you get approval, go back and launch the same instance again. It should succeed. If you're still stuck, move to the next step.

3. You're Using a Reserved Instance That Isn't Available Yet

Less common, but I've seen it. You buy a Reserved Instance (RI) for a specific zone and instance type, but the reservation hasn't fully propagated to all AWS systems. This can take up to 30 minutes. Or you might have an RI locked to a zone that's out of capacity.

The error shows up only for instances that exactly match the RI specs. You'll get the same InsufficientInstanceCapacity message.

Fix: Launch Without the RI or Wait

1. If you just bought the RI, wait 30 minutes and try launching again.
2. If you can't wait, launch an On-Demand instance without the RI. Use a different instance type or zone. The RI will still apply to matching instances later, but you don't have to use it right now.
3. Alternatively, modify the RI: go to the Reserved Instances section in the EC2 console, select the RI, click Actions > Modify, and change the Availability Zone to one that has capacity.

Quick Reference Summary Table