Cloud Backup Storage Pricing in 2026: 7 Proven S…

That $0.004/GB Backup Just Cost You $47,000 to Restore

Let me tell you about the most expensive cheap decision in cloud computing.

A company stores 500TB of backups in S3 Glacier Deep Archive. The storage cost is $0.00099/GB per month. That is $495/month for 500TB. Incredibly cheap. The CFO loves it. The CTO loves it. Everyone congratulates themselves on being cost-efficient.

Then disaster strikes. A database corruption requires a full restore. The team initiates the recovery and waits. Glacier Deep Archive takes 12 to 48 hours for standard retrieval. But this is urgent, so they use expedited retrieval.

Here is the bill that arrives:

Expedited retrieval: $0.03/GB x 500,000 GB = $15,000
Data transfer out to their recovery region: $0.09/GB x 500,000 GB = $45,000
API request fees for millions of GET requests: $2,400

Total restore cost: $62,400 for data that cost $495/month to store.

This is not an edge case. This is the default outcome when teams choose backup storage based on the per-GB storage price without understanding the full cost model. And it happens far more often than the cloud providers will ever publicize.

This post is going to show you exactly how cloud backup pricing really works in 2026, where the hidden costs are buried, and 7 strategies to build a backup system that is both cost-efficient and actually usable when you need it most.

How Cloud Backup Pricing Actually Works (The Parts They Downplay)

Every cloud provider markets backup storage with the storage price. That is the number you see in the headline. What they do not emphasize are the four other cost dimensions that often exceed the storage cost itself.

The Five Dimensions of Backup Cost

1. Storage cost (what you see advertised) This is the per-GB monthly rate for keeping your data. It ranges from $0.00099/GB (Glacier Deep Archive) to $0.023/GB (S3 Standard). This is the only number most teams look at.

2. Retrieval cost (what hits you during recovery) Every time you read data from a cold or archive tier, you pay a per-GB retrieval fee. On AWS Glacier, this ranges from $0.01/GB (standard, 3-5 hours) to $0.03/GB (expedited, 1-5 minutes). On Azure Cool Blob, it is $0.01/GB. These fees are per retrieval, and they apply even for partial reads.

3. Data transfer cost (what hits you when data moves) Moving backup data between regions or out to the internet costs $0.02 to $0.09/GB depending on the provider and destination. Cross-region replication for redundancy? That is an ongoing transfer fee on every new backup written.

4. API request cost (what hits you at scale) Every PUT, GET, LIST, and DELETE operation has a cost. At small scale, this is negligible. At backup scale (millions of objects, daily incremental syncs), API costs can reach hundreds or thousands of dollars monthly.

5. Early deletion penalties (what hits you when plans change) Archive tiers have minimum storage durations. S3 Glacier has a 90-day minimum. Glacier Deep Archive has a 180-day minimum. Azure Archive has a 180-day minimum. If you delete data before the minimum, you pay for the full minimum period anyway. Moved 100TB to Glacier last month and realized you need it sooner? You are paying for 90 days regardless.

The Real Cost Comparison Table (2026)

Here is what the major providers actually charge when you factor in all five dimensions:

Cost Component	AWS S3 Standard	AWS S3 Glacier	AWS Glacier Deep Archive	Azure Hot	Azure Cool	Azure Archive	GCP Standard	GCP Coldline	GCP Archive
Storage/GB/mo	$0.023	$0.004	$0.00099	$0.018	$0.01	$0.002	$0.020	$0.004	$0.0012
Retrieval/GB	Free	$0.01	$0.02	Free	$0.01	$0.02	Free	$0.01	$0.05
Min storage	None	90 days	180 days	None	30 days	180 days	None	90 days	365 days
Egress/GB (internet)	$0.09	$0.09	$0.09	$0.087	$0.087	$0.087	$0.12	$0.12	$0.12
Retrieval time	Instant	3-5 hrs	12-48 hrs	Instant	Instant	Up to 15 hrs	Instant	Milliseconds	Milliseconds

Look at the GCP Archive column. The storage price ($0.0012/GB) is slightly more than AWS Deep Archive. But the retrieval cost is $0.05/GB, which is 2.5x higher than AWS. And the minimum storage duration is 365 days, double the AWS minimum. The "cheapest" option depends entirely on your access pattern.

Strategy 1: Calculate Your True Cost of Backup (Not Just Storage)

The first thing you need to do is throw away whatever spreadsheet you are using that only compares per-GB storage prices. It is giving you the wrong answer.

Here is the formula that actually matters:

True Monthly Backup Cost = Storage Cost + (Expected Monthly Retrievals x Retrieval Rate) + (Monthly Data Transfer x Egress Rate) + (Monthly API Calls x Request Rate) + (Early Deletion Risk / Amortized Period)

Let me walk through a real example.

Say you have 100TB of backup data. You expect to retrieve about 5TB per month for testing, compliance checks, and occasional restores. You replicate across two regions for redundancy.

On S3 Glacier:

Storage: 100,000 GB x $0.004 = $400/mo
Retrieval: 5,000 GB x $0.01 = $50/mo
Cross-region replication (ongoing writes): ~2TB new data x $0.02/GB = $40/mo
API requests: ~$30/mo
Total: $520/mo

On S3 Standard (Infrequent Access):

Storage: 100,000 GB x $0.0125 = $1,250/mo
Retrieval: 5,000 GB x $0.001 = $5/mo
Cross-region replication: ~2TB x $0.02/GB = $40/mo
API requests: ~$15/mo
Total: $1,310/mo

Glacier is cheaper in this scenario. But watch what happens if you increase retrievals to 20TB per month (a compliance audit, a migration test, a disaster recovery drill):

Glacier with 20TB retrieval: $400 + $200 + $40 + $60 = $700/mo S3-IA with 20TB retrieval: $1,250 + $20 + $40 + $25 = $1,335/mo

Still cheaper. But if this is an emergency retrieval using expedited tier:

Glacier expedited 20TB: $400 + $600 + $40 + $60 = $1,100/mo

Now the gap has narrowed significantly. And you waited 3-5 hours for your data. With S3-IA, it would have been instant.

The move: Model your true backup cost for at least three scenarios: normal month, compliance audit month, and disaster recovery month. The tier that looks cheapest under normal conditions might be the most expensive when you actually need your data.

For a broader look at cloud storage pricing across providers, check out our cloud storage pricing comparison for 2026.

Strategy 2: Implement Intelligent Lifecycle Policies (And Avoid the Two Traps)

Lifecycle policies automatically move data between storage tiers as it ages. In theory, this is a great idea. Fresh backups stay in a fast, accessible tier. Old backups transition to cheaper archive tiers. Costs decrease over time.

In practice, two traps catch almost every team:

Trap 1: The Early Transition Penalty

You set a lifecycle policy to move objects to Glacier after 30 days. But S3 Standard has no minimum storage duration, and Glacier has a 90-day minimum. If your data gets deleted or replaced before those 90 days in Glacier are up, you pay for 90 days anyway.

This is especially expensive for incremental backups that get superseded quickly. If your daily incrementals transition to Glacier after 30 days but expire after 45 days (replaced by a weekly full backup), you are paying for 90 days of Glacier storage for data you only kept for 15 days in Glacier.

The fix: Never transition data to a tier where the minimum storage duration exceeds the expected remaining retention. If an object will be deleted in 45 days, do not put it in a tier with a 90-day minimum.

Trap 2: The Lifecycle Transition Fee

AWS charges a per-object fee for lifecycle transitions. Moving one object to Glacier costs $0.05 per 1,000 requests. That sounds trivial. But if your backup system creates millions of small objects (which most do), the transition fees add up.

A backup of 10 million objects transitioning to Glacier costs: 10,000,000 / 1,000 x $0.05 = $500 just in transition fees. If you do this monthly, that is $6,000/year in fees for the privilege of moving data to a "cheaper" tier.

The fix: If your backup creates many small objects, consider consolidating them into larger archive bundles (tar files) before transitioning. One large object transitioning to Glacier costs the same $0.05 per 1,000 requests as one small object. A 100GB tar file is infinitely cheaper to transition than the 500,000 small files inside it.

Read our guide on the cheapest cloud storage in 2026 for more provider-specific lifecycle strategies.

Strategy 3: Stop Cross-Region Replication From Eating Your Budget

Cross-region replication is the default recommendation for disaster recovery. And it is genuinely important. If your primary region goes down, having backups in another region means you can recover.

But the cost of cross-region replication is almost always higher than teams expect, because it compounds three separate charges:

Data transfer between regions ($0.02/GB on AWS, varies by region pair)
Storage cost in the secondary region (doubles your storage bill)
API request fees for the replication (PUT requests in the destination)

For 100TB of backup data with 5TB of daily changes:

Replication transfer: 5,000 GB/day x $0.02 x 30 = $3,000/mo
Secondary storage: 100,000 GB x $0.004 = $400/mo
API requests: ~$50/mo
Total replication cost: $3,450/mo ($41,400/year)

That is $41,400/year just for the redundancy, on top of whatever your primary storage costs.

Smarter Alternatives

Replicate selectively. Not all backups need cross-region redundancy. Your production database backup? Absolutely replicate it. Your dev environment snapshots? Probably not. Categorize your backups by business criticality and only replicate what genuinely needs disaster recovery protection.

Use cheaper replication destinations. Some region pairs have lower data transfer rates. AWS charges $0.02/GB from us-east-1 to us-west-2, but only $0.01/GB from us-east-1 to ca-central-1 (Canada). Azure has similar regional pricing variations. Pick your DR region based on both proximity and data transfer cost.

Replicate compressed, not raw. If you compress your backups before replication (most modern backup tools support this), you can reduce the data volume by 50% to 80%. That directly cuts your transfer costs by the same percentage.

Consider alternative providers for DR copies. Services like Cloudflare R2 charge zero egress fees. Storing a DR copy in R2 means you can retrieve it without the $0.09/GB egress charge from your primary provider. For large backup sets, the egress savings alone justify the additional storage cost.

Our best object storage for multi-region redundancy guide compares all the options in detail.

Strategy 4: Eliminate Backup Bloat (Your Biggest Hidden Cost)

Here is a number that will surprise you. The average company retains 3x to 5x more backup data than they need. Not because of compliance requirements. Because nobody ever cleans up.

Backup bloat happens for predictable reasons:

Retention policies are set once and never revisited. Someone chose "keep 365 daily backups" three years ago when the database was 50GB. Now the database is 2TB, and you are storing 730TB of backup history.
Orphaned backups from decommissioned systems. The staging server was retired 18 months ago. Its nightly backups are still running. Nobody turned them off because nobody knew they existed.
Snapshot sprawl. Every time an engineer creates a disk snapshot "just in case" before a deployment, it persists forever. After a year of weekly deployments, you have 52 snapshots that nobody will ever use.
Version accumulation in object storage. S3 versioning is enabled on your backup bucket (good for protection). But every overwrite creates a new version, and old versions are never cleaned up. Your bucket appears to hold 100TB but actually stores 400TB when you count all versions.

How to Find and Fix Backup Bloat

Audit snapshot age. Any snapshot older than 90 days that is not required by compliance should be reviewed. Any snapshot older than 180 days should have a documented justification or be deleted.

Check for orphaned backup jobs. Pull a list of all active backup schedules and cross-reference with your current infrastructure inventory. Any backup job targeting a resource that no longer exists is pure waste.

Enable S3 Inventory on backup buckets. S3 Inventory gives you a daily or weekly report of every object in your bucket, including size, storage class, and last modified date. Sort by size and age. You will almost certainly find terabytes of data that nobody has accessed in over a year.

Set lifecycle rules for noncurrent versions. If you use S3 versioning, add a lifecycle rule that deletes noncurrent versions after 30 or 60 days. This prevents version accumulation from quietly tripling your storage costs.

For more on eliminating cloud waste, read our guide on stopping payment for ghost servers and cloud waste.

Strategy 5: Match Your Backup Tier to Your Actual Recovery Needs

This is where most teams get the decision fundamentally wrong. They choose a backup storage tier based on cost, not on recovery requirements. Then they discover the mismatch during an actual incident, which is the worst possible time.

Let me give you a framework that flips this correctly:

Start With Recovery Requirements, Not Price

For every backup set, define two numbers:

Recovery Time Objective (RTO): How quickly do you need the data back?
Recovery Point Objective (RPO): How much data loss is acceptable?

Then match those requirements to the right tier:

RTO Requirement	Recommended Tier	Storage Cost	Why
Under 1 hour	S3 Standard / Azure Hot / GCP Standard	$0.018-$0.023/GB	Instant retrieval, no delays
1-4 hours	S3 Standard-IA / Azure Cool / GCP Nearline	$0.01-$0.0125/GB	Instant retrieval, lower cost
4-12 hours	S3 Glacier Instant Retrieval / Azure Cool	$0.004-$0.01/GB	Millisecond access, archive pricing
12-48 hours	S3 Glacier Flexible / Azure Archive	$0.0036-$0.004/GB	Hours to retrieve, very low storage
48+ hours acceptable	S3 Glacier Deep Archive / GCP Archive	$0.00099-$0.0012/GB	Cheapest storage, longest retrieval

The mistake teams make: putting production database backups in Glacier Deep Archive because it is cheap, and then discovering during an outage that they need 12 to 48 hours to get their data back. If your production database has a 4-hour RTO, Glacier Deep Archive is the wrong tier at any price.

The move: Audit every backup set against its RTO. Any backup stored in a tier that cannot meet its RTO is a ticking time bomb. Move it to the correct tier now, not during an incident.

Strategy 6: Use FinOps Practices to Track and Optimize Backup Spend

Backup storage is one of the least monitored cost categories in most cloud environments. Teams track compute costs obsessively. They watch database costs weekly. But backup storage? It just grows quietly in the background until someone asks "why is our S3 bill $15,000 this month?"

Here is how to bring FinOps discipline to your backup spend:

Tag Every Backup Resource

Apply consistent tags to every backup bucket, snapshot, and replication rule:

backup-source: Which system this backs up
backup-type: Full, incremental, snapshot, or archive
retention-policy: How long this data should be kept
environment: Production, staging, dev, or test
cost-center: Which team or project owns this cost

Without tags, you cannot attribute backup costs to specific systems, which means you cannot make informed decisions about retention and tiering.

Track Backup Cost Per System

Your production PostgreSQL database backup costs X dollars per month. Your Elasticsearch cluster snapshots cost Y. Your application file backup costs Z. Track these individually, not as one aggregated "storage" line item.

When you can see that backing up your Elasticsearch dev cluster costs $800/month (because it creates 50GB of snapshots daily with 90-day retention), the optimization opportunity becomes obvious: reduce snapshot frequency to weekly or cut retention to 14 days. That $800/month drops to $100/month.

Run Quarterly Backup Cost Reviews

Add backup costs to your quarterly FinOps review. Look for:

Backup sets that grew more than 20% quarter-over-quarter
Systems with backup costs exceeding 10% of the compute cost of the system itself
Retrieval fees that spiked (indicating an incident or a misconfigured process)
Cross-region replication costs that exceed the value of the data being replicated

For expert help building FinOps practices around your entire cloud spend, explore our Cloud Cost Optimization and FinOps service.

Strategy 7: Consider Alternative Providers for Backup Storage

Here is something that most teams never consider because they assume all their data needs to live with one provider: the best backup storage provider might not be the same provider where your application runs.

Backup data has unique characteristics that make it ideal for alternative providers:

It is written frequently and read rarely
It does not need low-latency access (except during recovery)
It benefits enormously from zero-egress pricing (since recovery means large data transfers)

The Alternative Provider Landscape

Provider	Storage/GB/mo	Egress	Retrieval	Best For
Cloudflare R2	$0.015	Free	Free	DR copies, any backup needing free retrieval
Wasabi	$0.0069	Free (reasonable use)	Free	Large backup sets, cost-sensitive storage
Backblaze B2	$0.006	Free to Cloudflare	$0.01/GB	Budget backup with Cloudflare CDN integration
AWS S3 Glacier Deep Archive	$0.00099	$0.09/GB	$0.02/GB	Compliance archives you rarely need to access

Look at Wasabi. At $0.0069/GB with free egress, 100TB of backup storage costs $690/month with zero retrieval or egress fees. On AWS S3 Standard-IA, the same 100TB costs $1,250/month for storage alone, plus retrieval and egress fees on top.

The trade-off: alternative providers may not match the durability guarantees, compliance certifications, or integration depth of the major clouds. For your primary production backups, AWS, Azure, or GCP is still the safest choice. But for secondary DR copies, compliance archives, or dev/test backup data, alternative providers can save 50% to 80%.

Our cloud storage pricing comparison for 2026 breaks down every provider in detail.

Your Backup Cost Optimization Checklist

Immediate Actions (This Week)

Calculate your true backup cost using the five-dimension formula (storage + retrieval + egress + API + early deletion)
Enable S3 Inventory or equivalent on all backup buckets
List all active backup schedules and cross-reference with live infrastructure
Delete any snapshots older than 180 days without documented justification

Short-Term Wins (This Month)

Audit every backup set against its RTO and move mismatched tiers
Set lifecycle rules for noncurrent versions (30 or 60 day expiry)
Tag all backup resources with source, type, retention, and cost-center
Review cross-region replication and eliminate redundancy for non-critical data

Strategic Improvements (This Quarter)

Model cost scenarios for normal, audit, and disaster recovery months
Evaluate alternative providers (R2, Wasabi, B2) for DR copies and archives
Implement backup consolidation (tar bundles) before archive tier transitions
Add backup costs to quarterly FinOps reviews
Set up anomaly alerts for backup storage growth exceeding 20% month-over-month

Your Backups Are an Insurance Policy. Price Them Like One.

Here is the mindset shift that changes everything about backup cost optimization.

Backups are insurance. And like any insurance policy, the real cost is not just the premium (storage). It is the premium plus the deductible (retrieval fees) plus the exclusions (slow recovery times that cost you revenue).

A cheap backup that takes 48 hours to restore when your business is losing $10,000/hour in downtime is not cheap at all. It is the most expensive decision you could have made.

The teams that get backup pricing right are the ones who start with the question "what do we need when things go wrong?" and work backward to the right tier, the right provider, and the right redundancy level. They calculate the total cost, not just the storage price. And they review it quarterly, because data grows, access patterns change, and last year's optimal configuration is this year's waste.

Your backup strategy should protect your business without silently draining your budget. The 7 strategies in this post will get you there.

Want a complete picture of your cloud waste, including backup bloat? Take our free Cloud Waste and Risk Scorecard for a personalized assessment in under 5 minutes.

Related reading:

External resources:

Cloud Backup Storage Pricing in 2026: 7 Proven Strategies to Stop Hidden Costs