Guide · Fine-Print Index data
Which VPS providers bill stopped servers?
A powered-off server is not a free server. Here is what 20 hourly VPS providers charge for a stopped server, why they charge it, and how to end the bill for good.
The short answer
Of the 20 hourly VPS providers we checked on , 12 bill a stopped server in full until you delete it. 5 stop charging for compute while the server is powered off but keep billing its disks or IP addresses, and 3 depend on the plan or on how you stop the server.
Deleting the server is the step that ends its charges everywhere. Snapshots, volumes and reserved IP addresses you keep are billed on their own.
What a stopped server costs, provider by provider
Each provider’s rule in our words, with the page it comes from. The data is part of the Hourly VPS Fine-Print Index, where you can compare the other billing fields too.
Compute stops; disks or IPs may still bill (5)
| Provider | What the provider says | Source |
|---|---|---|
| ExoscaleStandard instances | Compute stopsPartial - a powered-off instance is not charged for compute, but its local storage volume (and any Windows license) keeps being billed until deleted | community. |
| Google Compute EngineE2 VMs | Compute stopsPartial - a stopped (TERMINATED) VM is not charged for vCPU or memory, but its disks and any static external IP keep being billed | cloud. |
| IONOS Cloud CubesCloud Cubes | Compute stopsYes (partial). Suspended Cubes 'continue to incur costs': suspending releases CPU, RAM and IP addresses, but connected storage stays billed. | docs. |
| KamateraHourly cloud servers | Compute stopsPartial. Powered-off hourly servers pay 'a small hourly charge' for storage space and IP address. | kamatera. |
| ScalewayCPU Instances | Compute stopsNo for compute after 'Power off' (billing stops at power-off). 'Standby' is billed as a running Instance. Storage volumes and flexible IPs keep billing while powered off. | scaleway. |
Depends on plan or how you stop it (3)
| Provider | What the provider says | Source |
|---|---|---|
| Microsoft AzureB-series VMs | Depends on how you stopDepends on the stop type: 'Stopped (allocated)' (shut down from inside the OS or PowerOff) is billed in full; 'Stopped (deallocated)' is not billed for compute, but disks and networking still are | learn. |
| Oracle Cloud (OCI)Flex VMs, pay as you go | Depends on how you stopNo for standard (incl. flex) VM shapes - stopping from the Console/API pauses compute billing; boot/block volumes keep billing. Dense I/O, most GPU and HPC shapes keep billing while stopped. Shutting down from inside the OS does not stop billing | docs. |
| UpCloudCloud Servers | Depends on planPartial. FAQ: Developer and General Purpose servers are billed per hour whether on or shut down; Cloud Native servers are billed only while powered on (pricing page: 'Server stopped, no charges'). The FAQ does not yet name the new Starter/Premium plans. | upcloud. |
Billed in full until deleted (12)
| Provider | What the provider says | Source |
|---|---|---|
| Akamai (Linode)Shared CPU Linodes | Billed in fullYes. Charges accrue for any service on the account even when powered off, because data and resources such as RAM stay reserved. | techdocs. |
| Atlantic.NetCloud VPS, on-demand | Billed in fullYes. Billed from creation until deletion; a stopped server still reserves CPU, memory and storage. | atlantic. |
| AWS LightsailLightsail bundles | Billed in fullYes - instances accrue charges in the stopped state until deleted | docs. |
| Cherry ServersCloud VPS | Billed in fullYes - a powered-off server stays provisioned and is billed at the regular rate until terminated | cherryservers. |
| CivoCompute instances | Billed in fullYes - a shut-down instance is charged exactly like a running one until deleted | civo. |
| DigitalOceanDroplets (bundled plans) | Billed in fullYes. Powered-off bundled and GPU Droplets are billed because resources stay reserved; billing ends only when the Droplet is destroyed. | docs. |
| Hetzner CloudCloud servers | Billed in fullYes. Servers are billed until deleted, whether on or off, because full resources stay allocated. | docs. |
| LightNodeHourly VPS | Billed in fullYes - a stopped instance keeps being billed until it is released (deleted)LightNode's billing docs and FAQ say a stopped server keeps billing; its hourly landing page says turning the server off stops billing. | doc. |
| OVHcloudPublic Cloud instances | Billed in fullYes. 'Stopped' or 'paused' instances keep billing until deleted. Hourly instances can be shelved (suspended): resources are released and only the snapshot is billed. | docs. |
| RamNodeCloud VPS | Billed in fullYes - instances are billed for every hour they exist, powered on or off, until destroyed. 'Shelving' (snapshot + removal from compute) stops compute charges | ramnode. |
| VultrCloud Compute | Billed in fullYes. Stopped instances keep RAM, SSD, IPs and vCPU reserved and are billed until destroyed. | vultr. |
| X-ZoneServersHourly VPS | Billed in fullYes - a stopped server keeps reserving RAM, CPU, storage and IP and bills hourly until destroyed | x-zoneservers. |
Why a stopped server still costs money
Powering off a VPS stops its operating system, not the server. The disk still holds your data, the public IP address stays assigned to you, and on most VPS platforms the server’s share of CPU and memory stays reserved on its host, so it can start again with the same address. While you hold those resources, nobody else can rent them. That is why most providers in the tables above bill a stopped server like a running one.
Some large clouds work differently. They release CPU and memory when you stop a server and find a host for it again when you start it. They keep billing what they still hold for you, mainly disks and reserved IP addresses. A few release resources only when you stop the server through their console or API; shutting down from inside the operating system leaves it billed.
Releasing resources has a cost of its own: the server may start on a different host, and an address that is not reserved can change. Read your provider’s row before you rely on either behavior.
How to actually stop the bill
If you will not need the server for a while, removing it is the only step that ends its charges on every provider. Do it in this order:
- Save what you need. Take a snapshot or image of the disk, or copy the data off the server. Check what your provider charges to store a snapshot, and whether the snapshot survives when the server is deleted.
- Plan for a new IP address. A rebuilt server usually gets a different address. Update DNS records and firewall rules that point at the old one, or keep a reserved IP, which is often billed while you hold it.
- Delete the server, do not just stop it. Control panels call this Delete, Destroy, Terminate or Release.
- Look for leftovers. Volumes, snapshots, backups, reserved IPs and load balancers are often billed separately and can outlive the server. Check the usage page or the next invoice to confirm the charges stopped.
- Come back from the snapshot. When you need the server again, create a new one from the snapshot. Billing starts again from the moment it exists.
Shelving, deallocating and other in-between states
Several providers in the tables offer a state between running and deleted, usually called shelving, suspending or deallocating. Compute is released and storage keeps billing, so you skip the snapshot-and-rebuild steps but still pay for the disk. The tables quote each provider’s own wording; follow the source link before you count on it.
For the other billing rules that shape a VPS bill, such as the billing unit, monthly caps and refunds, see how hourly VPS billing works and per-second vs per-hour VPS billing.
How HourlyVPS handles it
We are in the “billed in full” group: a stopped server keeps its resources, so it is billed until you delete it. These are the rules that apply:
Stopped servers
A stopped server is still billed, because its vCPU, memory, disk and IP addresses stay reserved for you; only deleting the server stops billing.
Hourly billing
Every server is billed by the hour: the plan’s hourly rate is deducted from the server’s prepaid balance for every hour it exists, powered on or off, until you delete it.
Billing unit
Billing is by the hour: every hour a server exists is charged at the plan’s hourly rate. The price tapes and cost tables on this site count every started hour as a full hour, so they show the most a duration can cost; the cost calculator charges a partial hour to the nearest cent, as the bill does.