Do storage charges stop when a rented GPU stops?
Do not assume so. The provider may retain a separately billed volume after compute stops. The fees table distinguishes compute billing while stopped from storage charges.
In today's data: Of 18 providers whose terms we checked (7 July 2026), 8 keep billing storage after the GPU stops (RunPod, Lambda, Hyperstack, Nebius, Crusoe, Jarvislabs, Microsoft Azure and AWS), Vultr keeps billing the instance, Modal bills nothing while stopped, and 8 do not publish it.
Budget storage for its retained lifetime, including time between sessions. Check how to delete the volume and whether deleting a machine also deletes the data.
Related questions
- What does 100 GB of GPU cloud egress cost?
- Should checkpoint and restart time be in a GPU budget?
- Is a GPU cloud deposit part of the rental price?
Numbers on this page come from today's verified snapshot. Full table on the homepage; method in the methodology.