Cloud backup pricing looks simple on a rate card and turns complicated on an invoice, where storage is only one of several line items. AWS, Azure, Google Cloud, and flat-rate providers each bill the rest differently, so the real number only shows up once you model your own data.
What goes into cloud backup pricing?
Cloud backup pricing rarely means a single per-GB rate. Providers bill backup as a stack of separate charges, and each one scales differently as data, retention, and recovery activity grow.
Storage is the line item buyers expect. The charges around it are where bills diverge.
A complete backup estimate accounts for six cost dimensions:
- Storage: Per-GB-month charge for backup data at rest, billed on the average volume stored.
- Management or protected-instance fees: A per-resource charge some providers add on top of storage.
- Restores: A per-GB retrieval charge when you recover data, higher from cold tiers.
- Cross-region transfer: A per-GB fee to copy backups to a second region for disaster recovery.
- API and operations: Charges for the read, write, and list calls that backups generate.
- Archive retrieval and early deletion: Fees to pull data from cold tiers and penalties for deleting before a minimum retention period.
Native cloud tools meter most of these separately, which is why two environments with the same data volume can produce very different bills.
Cloud backup pricing across major providers in 2026
Backup rates vary by provider, by service, and by storage tier. The figures below reflect published 2026 pricing in common US regions, before volume discounts.
AWS Backup pricing
AWS Backup charges per GB, and the rate changes by service, which is the detail most estimates miss. Warm storage runs about $0.05 per GB-month for EBS and S3, while RDS sits just above that rate and DynamoDB backup falls at or below it, per the AWS Backup pricing page.
Cold storage runs far lower for supported services, with a 90-day minimum retention that bills in full on early deletion. Restores add their own per-GB charge, plus a per-request fee for item-level recovery.
Azure Backup pricing
Azure Backup splits the bill into two parts, a protected-instance fee tiered by instance size and a separate per-GB storage charge, per the Azure Backup pricing page.
Storage cost then depends on the redundancy tier. Locally redundant storage is the cheapest, while zone-redundant and geo-redundant options raise the per-GB rate, and geo-redundant restore upgrades storage again.
Restores from the standard tier carry no egress charge. Archive-tier data triggers a one-time retrieval fee and a 180-day early-deletion window that bills prorated time even after the data is gone.
Google Cloud backup pricing
Google Cloud prices backup by storage class, ranging from $0.020 per GB-month for Standard down to $0.0012 for Archive, per Google Cloud Storage pricing. The cheaper the class, the more you pay to read the data back.
Cold classes look cheap until access. Archive adds a $0.05 per GB retrieval fee and a 365-day minimum that bills in full even on early deletion, shown in Google's own pricing examples.
Egress is the bigger variable. Data leaving Google Cloud to the internet starts at $0.12 per GB, higher than the comparable AWS and Azure rates, which inflates any cross-cloud recovery.
Flat-rate storage providers
A separate group of providers prices storage at a flat per-TB rate with generous free egress. Object storage of this kind can read as the cheapest option on a comparison chart, often near $6 per TB-month.
That model fits file archives and object storage well. It does not natively back up managed databases, Kubernetes state, or multi-account cloud environments, which is the workload enterprise teams actually need covered.
Consumer backup tools price differently again, often near $99 per year per device for unlimited storage. Those tools target laptops and endpoints, not cloud infrastructure, so they rarely belong in an enterprise backup comparison.
What pushes the real bill above the per-GB rate
The per-GB rate is where the estimate begins. Four factors stretch it into a much larger invoice, and they compound as environments grow.
A worked example of how the estimate grows
Consider 50 TB of EBS-backed data on AWS Backup, kept 35 days on a daily schedule. The naive estimate multiplies 50 TB by $0.05 per GB-month and lands at $2,500 per month.
Retention raises the stored volume to roughly 85 TB once incrementals accumulate, pushing storage past $4,000. A cross-region disaster recovery copy duplicates that storage in a second region and adds per-GB transfer fees.
Restore testing, item-level recovery, and API operations layer on top. The run-rate commonly lands at two to three times the rate-card figure, before anyone restores a single production workload.
Retention multiplies storage
Backups accumulate. A daily backup with a 35-day window stores the original full copy plus weeks of incremental changes, so stored volume runs well above the source data size while the per-GB rate holds steady.
Redundancy and cross-region copies change the rate
Redundancy tiers reprice the same data. Azure's geo-redundant storage costs more per GB than locally redundant storage, and a cross-region disaster recovery copy in AWS duplicates storage in a second region.
Restores, restore testing, and item-level fees
Recovery is a line item. AWS charges about $0.02 per GB to restore from warm storage and $1.50 per recovery point for restore testing, and item-level restores add a per-request fee.
Teams that test recovery regularly should budget those charges rather than treat restores as free.
Egress and cross-region transfer
Moving backup data triggers transfer fees. Cross-region copies within AWS run about $0.02 per GB, while internet egress from Google Cloud starts at $0.12 per GB, which cross-cloud recovery multiplies across the full dataset.
Backups that live in one provider but need to restore into another carry this cost on every recovery, which cross-cloud strategies often underestimate.
Over-retention you are already paying for
Part of the backup bill pays to store data no one will recover. In Eon's 2026 Cloud Data Infrastructure Report, 87% of cost-pressured organizations admit they keep data they do not need.
Most teams never see this layer, because nothing classifies data by value or ages it out automatically, so retention quietly drifts upward.
How pricing differs for databases, object storage, and Kubernetes
Backup price depends as much on what you protect as how much. The same provider charges different rates by workload, and the recovery model changes the cost too.
Databases
Managed database backup is priced closer to block storage than most estimates assume. AWS prices RDS backup just above the $0.05 block rate, and DynamoDB backup often lands below it. What inflates a database's real cost is restore behavior, recovering or testing against full instances, plus cross-region copies.
Restoring a single record from a native snapshot usually means restoring the whole instance, which adds compute and storage for data no one needed. Granular recovery restores the specific table or record instead, which keeps restore cost tied to the size of the loss.
Object storage
Object storage backup looks cheap per GB and bills in ways that add up. AWS meters S3 backups per GB-day and adds small-object overhead, while cold tiers carry retrieval fees and minimum durations.
Versioning compounds it. Every retained version of an object bills at the live-object rate, which inflates storage when lifecycle rules are loose.
Kubernetes
Kubernetes backup adds a per-cluster or per-namespace charge. AWS Backup applies a one-time creation fee per namespace and bills the persistent-volume snapshots as standard storage.
Agent-based tools add license and compute cost per cluster, while an agentless model avoids that per-cluster overhead entirely.
How to lower cloud backup costs without cutting protection
The largest backup savings come from architecture rather than a lower rate. Reducing duplicated data and right-sizing retention changes the GB count, which is the number every provider multiplies.
Cut duplicate data at the storage layer
Deduplication across the whole environment shrinks what you store. When the same blocks are kept once rather than per snapshot, the protected volume drops without losing any recovery point.
This is where source-aware storage makes the difference. Eon's vault stores a fraction of what native snapshots hold by deduplicating across the whole environment.
On that model, the same protected data commonly costs around 40% less to store. NETGEAR cut backup storage costs 35%, and Sigdo Koppers projected roughly 38% lower than native GCP snapshots.
Right-size retention and tiering
Retention drives storage more than any rate. Matching retention to actual recovery and compliance needs, rather than defaulting to long windows, removes data you are paying to keep.
Tiering helps when access patterns are clear, though retrieval fees and minimum durations need modeling first. Cloud backup posture management (CBPM) automates the retention side, classifying resources on creation and applying retention by data value so coverage right-sizes itself instead of drifting.
Avoid hidden platform fees
Agents, appliances, and per-API charges add cost the rate card hides. An agentless platform with consumption-based billing removes the appliance and per-call overhead that native and legacy tools accumulate.
That is the cost angle behind SoFi's result, over 100% ROI in year one while expanding coverage across five AWS regions, in its move to Eon.
How to evaluate cloud backup pricing before you commit
Evaluate cloud backup pricing on total cost over a realistic year, modeled against your own data, retention, and recovery patterns.
- Model the full bill, not the per-GB rate. Add restores, cross-region copies, API charges, and archive retrieval to storage.
- Map rates to your workload mix. Database and Kubernetes backup often cost more per GB than block or object storage.
- Check restore costs and granularity. Confirm whether you can recover a file or record without restoring a full instance.
- Test redundancy assumptions. Decide which data needs geo-redundant or cross-region copies, since each one reprices storage.
- Account for over-retention. Identify data you keep out of habit, since it inflates every line item.
The factors below map each pricing lever to why it counts and what to check before signing.
Backup spend stays controllable when coverage is visible. Knowing what is protected, what is over-retained, and what is drifting turns a guessing game into a budget you can defend.
Where is your backup bill actually leaking?
The same levers that inflate the bill are the ones Eon is built to pull back. Cloud-native deduplication stores a fraction of what native snapshots keep, which is where the 30–50% savings show up.
An agentless, consumption-based model removes the appliances and per-call fees that legacy tools accumulate, and CBPM right-sizes retention on its own, so you stop paying to store data no one will recover. Cost Explorer maps every line back to the resource behind it, so a leak shows up before it compounds.
Want to see your real number before you commit to it? Book a demo and see how Eon maps your backup spend by resource and models the architecture changes that bring it down.
Frequently asked questions
Is cloud backup priced per GB?
Yes for storage, but not only. Cloud backup is priced per GB for storage, plus separate charges for restores, redundancy, cross-region transfer, and API operations. The per-GB rate is only the base. Retention growth, redundancy, and recovery fees commonly push the real bill to two to three times the storage-only estimate.
Why is my AWS Backup bill higher than expected?
An AWS Backup bill runs higher than the per-GB estimate because retention, restores, and cross-region copies are billed separately. Storing daily backups over a long window multiplies the stored volume, and item-level restores plus restore testing add per-request fees the rate card does not show.
Is flat-rate storage cheaper than native cloud backup?
Not usually. Flat-rate storage can look cheaper per TB, but it covers object storage rather than managed databases, Kubernetes, or multi-account environments. For enterprise cloud workloads, native and cloud-native backup tools handle coverage that flat-rate object storage does not.
Do you pay to restore cloud backups?
Yes. You usually pay to restore cloud backups, charged per GB recovered. AWS charges about $0.02 per GB from warm storage, cold-tier restores cost more, and restore testing adds $1.50 per recovery point, so frequent recovery raises the total cost.
How much can teams save on cloud backup costs?
Teams commonly cut cloud backup storage costs by 30% to 40% through deduplication and right-sized retention rather than by negotiating rates. Customer results in that range include NETGEAR at 35% and Innago at 40%, both from reducing duplicated backup data instead of buying cheaper storage.
