Free calculator
gp2 → gp3 savings calculator
gp3 is 20% cheaper than gp2 for identical storage with higher baseline IOPS on any volume under 1 TB. It launched in 2020, but existing gp2 volumes stay on gp2 until someone changes them.
Add up the size of every gp2 volume, including unattached ones, because they bill too.
On gp2 today
$50.00/month
$600.00/year · ≈ £39/month
Same storage on gp3
$40.00/month
$480.00/year · ≈ £31/month
Your saving
$10.00/month
$120.00/year · ≈ £8/month · 20% off, like for like
$0.10/GB-month for gp2 vs $0.08/GB-month for gp3. Volumes can migrate live with zero downtime via Elastic Volumes.
Don’t know your total gp2 footprint? The free scan lists every gp2 volume in every region, with the exact saving per volume, and runs 50+ other checks.
Why gp2 volumes pile up
- →gp2 was the default for a decade. Every instance launched before ~2021, every old AMI, every copy-pasted CloudFormation template creates gp2. New infrastructure quietly inherits it.
- →Root volumes get forgotten. Root volumes are easy to overlook. Each one stays on whatever type it launched with, which may be gp2.
- →Unattached volumes bill too. Detached gp2 volumes from terminated instances keep billing at the full rate. Consider deleting those as well as migrating the rest.
- →The saving compounds with rightsizing. Look for oversized volumes, such as a 500 GB volume that is 12% full. EBS volumes cannot shrink in place, so copy the data to a smaller gp3 volume and both savings stack.
Frequently asked questions
How much cheaper is gp3 than gp2?
gp3 is $0.08 per GB-month against $0.10 for gp2, so a flat 20% off every gigabyte. A 500 GB volume drops from $50 a month to $40. Across a fleet that adds up quickly, and it keeps saving every month afterwards.
Is gp3 slower than gp2?
It depends on the metric. Every gp3 volume gets a 3,000 IOPS / 125 MiB/s baseline regardless of size. gp2 scales baseline IOPS with size (3 IOPS per GiB), so any gp2 volume under 1 TB gains baseline IOPS by moving to gp3. gp2 delivers up to 250 MiB/s on larger volumes, so check throughput on those.
Is there downtime when migrating gp2 to gp3?
No. EBS Elastic Volumes lets you change the volume type live, while the volume stays attached and in use. It is a single ModifyVolume call or a few clicks in the console. The volume goes through an "optimizing" state but stays fully usable.
When is gp3 not a straight win?
Volumes over 1 TB that genuinely use more than 3,000 IOPS need extra provisioned IOPS on gp3 ($0.005/IOPS-month above the baseline), which eats into the 20%. Check CloudWatch IOPS and throughput metrics before migrating the big ones.
Why hasn’t AWS migrated my volumes automatically?
AWS never changes your resources for you. gp3 launched in 2020, but gp2 is still the default outside the console, so accounts can keep adding gp2 volumes.