Vultr Backups vs Snapshots: 20% of Instance vs $0.05 per GB

Vultr Automatic Backups cost 20% on top of the instance price, run daily, every other day, weekly or monthly, keep only the two most recent copies, and are stored in the same datacenter as the instance. Snapshots are taken manually, cost $0.05/GB per month of compressed size, don’t expire until you delete them, and can deploy in any region (Vultr docs, checked October 28, 2026).

This post contains affiliate links. If you purchase through our links, we may earn a commission at no extra cost to you.

Vultr Snapshot vs Backup: the quick answer

Who wins for what, before the theory:

Vultr describes both as point-in-time images of a Cloud Compute instance; backups run automatically and snapshots are taken manually. Snapshots are something you trigger manually (or script), cost $0.05/GB per month based on compressed size, and never expire. Automatic Backups run on a schedule you set once, cost a flat 20% surcharge on your instance’s price regardless of how much data you actually have, and only keep the two most recent copies.

What is a Vultr Snapshot

A Snapshot is a manually triggered image of your instance’s disk: the OS, installed software, configuration, and data, frozen exactly as it existed when you clicked “Create Snapshot.” Snapshots work across every Vultr region: take one from an instance running in London, and you can deploy a brand-new instance from that same image in New York, Tokyo, or any other Vultr datacenter. That portability is why snapshots, rather than backups, are the standard tool for cloning a working server or migrating between regions.

You can use a snapshot two ways: restore it directly onto the existing instance, overwriting whatever’s currently there, or use it to spin up a completely new instance by selecting “Snapshot” as the deployment source. One limitation to know up front: snapshots aren’t available for Bare Metal instances. If you’re running Bare Metal, Vultr points you toward third-party backup software instead.

What is Vultr Automatic Backup

Automatic Backups is an opt-in setting you enable per-instance. Once it’s on, Vultr captures the instance on a schedule you pick (daily, every other day, weekly, or monthly, with a specific time and day). Vultr doesn’t stop the server to do this, so the instance keeps running through the capture. That’s convenient, but it comes with a real trade-off: Vultr’s own documentation compares booting from a backup to rebooting after a non-graceful restart. For a busy database taking writes mid-backup, there’s a genuine chance of capturing an inconsistent state.

A few details that matter in practice: backups don’t include attached block storage volumes, only the instance’s primary file system. Retention is thin: only the two most recent backups are kept, with older ones deleted automatically as new ones complete. There’s no long-term archive here unless you convert a backup to a snapshot, which keeps it and lets you restore it to a different server.

Vultr Snapshot vs Automatic Backup side by side

Checked against Vultr docs on October 28, 2026.

Backups are stored in the same datacenter as the original instance. To restore one on a different server, Vultr’s docs say to convert it to a snapshot first (the camera icon in the backups list); snapshots can then deploy in any Vultr region.

The real cost math

This is where most comparison posts get lazy, so let’s actually run the numbers across a few instance sizes.

Small instance, $6/month (1 vCPU/1GB plan): Automatic Backups add 20%, so you’re paying $1.20/month extra. A single compressed snapshot of roughly 10GB costs about $0.50/month. If you’re only keeping one or two snapshots, manual snapshots are clearly cheaper here.

Mid-size instance, $20/month (2-4GB plan): Backups add $4/month. An equivalent 40GB compressed snapshot runs about $2/month, but that’s a single snapshot with no auto-rotation. Backups buy you convenience and automatic retention; snapshots buy you raw storage economy.

Larger instance, $96/month (High Frequency plan): Backups add $19.20/month. Because the 20% surcharge scales with instance price, not disk usage, this is where the backup tax starts to sting, especially if your actual disk usage is modest. A cost-conscious setup here might use a weekly cron-triggered API snapshot instead of paying nearly $20/month for automation.

The pattern: Automatic Backups get proportionally more expensive as your instance price climbs, because the fee is a fixed percentage of price, not data size. Snapshots get proportionally cheaper per dollar of instance cost as your actual disk usage shrinks relative to plan size. Where the crossover point lands depends entirely on how much real data you’re storing versus what you’re paying for compute, and that’s exactly the nuance Vultr’s own docs skip over.

If you’re spinning up a test instance to run these numbers yourself before committing to a plan, new Vultr accounts can apply a promo code for up to $300 in credit, which Vultr’s docs say lasts up to 30 days, enough room to test snapshots and backups on a few instance sizes without spending anything out of pocket. You can grab that credit and start testing through Vultr. For a full breakdown of how Vultr’s plan pricing works before you pick an instance size, our Vultr pricing breakdown covers every tier, and our deep dive on the $300 credit walks through exactly what it does and doesn’t cover.

Retention limits and restore behavior

Snapshots don’t expire; they stay in your account until you delete them, and Vultr’s snapshot FAQ names no per-account cap.

Restoring a snapshot gives you a real choice: overwrite the existing instance’s current state in place (Vultr says a restore may take an additional 10 to 15 minutes, and an instance with a static IP needs its networking reset to DHCP afterward), or deploy a fresh instance from that snapshot image instead. A backup restore overwrites the same instance with the saved disk, all or nothing, not individual files; to deploy a backup anywhere else, convert it to a snapshot first.

If you’re planning for compliance or disaster recovery, note that Vultr documents no option to download snapshots or backups off its platform. If your data protection strategy depends on an offsite copy (the “1” in a 3-2-1 backup strategy), neither Vultr feature gets you there on its own. You’d need to pair either one with something like rsync or rclone pushing data to external storage if true offsite redundancy matters for your setup.

The stopped-instance billing myth

A common assumption is that stopping the VPS pauses backup billing too. It doesn’t. Vultr continues to reserve CPU, RAM, SSD, and your IP address for a stopped-but-not-destroyed instance, and all of that is billable regardless of power state. The 20% Automatic Backup surcharge keeps accruing the entire time the instance exists, stopped or running. The only way to actually stop being charged is to destroy the instance outright. After deletion, its two most recent backups stay for up to 7 days and are then purged. If you’re trying to save money on a server you’re not actively using, destroying it (after taking a final snapshot) is the only move that actually works.

Which one should you use?

Based on how each feature actually behaves, here’s how to split the decision:

Use Snapshots when you’re cloning an instance to a new region, taking a safety point before a risky upgrade or migration, or want to archive a known-good state indefinitely. Cost scales with your actual disk usage, not your instance price, so this is the economical choice for infrequent, deliberate protection.

Use Automatic Backups when you want zero-maintenance, continuous protection against accidental data loss on a production instance, and you’re fine paying the flat 20% tax for “set it and forget it” coverage. Just remember only two rotations are kept, which makes it a rolling safety net.

Use both when you’re running anything in production. Automatic Backups catch accidental data loss on a regular cadence, and a manual snapshot right before any risky change (a major update, a migration, a schema change) gives you an instant rollback point that backups’ 2-copy rotation might not cover in time. If you’re running a self-managed database, this combination matters even more. Our PostgreSQL on Vultr backup guide goes through exactly how to layer these protections around a live database.

On a $20/month instance, a common approach is running Automatic Backups on a daily schedule for routine safety and triggering a manual snapshot before every deployment. The backup rotation costs $4/month, and the deployment snapshots add another dollar or two depending on how many accumulate before cleanup. That’s a small price for not having to think about it.

Snapshots for cloning, backups for recovery

There’s no universal winner in the Vultr Snapshot vs Backup decision, because the two solve different problems. If you want cheap, indefinite, cross-region-portable images for cloning and pre-change safety points, use Snapshots. If you want hands-off recurring protection and don’t mind paying a percentage of your instance price for it, turn on Automatic Backups. For anything running in production, run both: daily or weekly Automatic Backups for routine coverage, plus a manual snapshot before any change you’d hate to redo from scratch.

If you’re still deciding on the instance itself, our full Vultr review covers how the platform performs day-to-day beyond just backups and snapshots. And if you haven’t set up a Vultr account yet, the up-to-$300 promo credit (up to 30 days) mentioned earlier is the cheapest way to test snapshot and backup costs on a real instance before you commit a card to a monthly plan.