Azure Blob Storage pricing starts at $0.018 per GB per month for hot data and drops to $0.00099 per GB in archive, quoted for East US on locally redundant storage. That capacity rate is one of five meters on the invoice, and on any workload with real read traffic it is rarely the largest one.
Azure Blob Storage pricing at a glance
Azure prices block blobs on the access tier the data sits in. Capacity is billed per GB per month, and every tier below hot trades a lower storage rate for higher charges when the data is touched.
Rates are Azure list prices for East US on LRS, first 50 TB band, current as of September 2026. They change by region and redundancy, and Azure bills in binary gigabytes, so 1 TB is 1,024 GB on the invoice.
The tier table is where estimates go wrong, because four other meters run alongside it:
Blob capacity charges are also separate from Azure Backup pricing, which meters protected instances and vault storage under its own model even when the backup data lands in blob storage.
Azure Blob Storage pricing changes in 2026
Smart tier reached general availability and now appears alongside hot, cool, cold, and archive in Microsoft's access tier documentation. It handles tier movement automatically and carries its own billing rules.
The 128 KiB minimum billable object size was announced, then paused. Microsoft had set it to apply to cool, cold, and archive from July 1, 2026, for new storage accounts, extending to all existing accounts on July 1, 2027, but paused the change on June 8, 2026, before it took effect.
Had it landed, any object under 128 KiB in those tiers would have billed as 128 KiB, turning a million 4 KB files in cool from 4 GB into 128 GB of billable capacity. For now, small objects still bill at their actual size.
When Azure Blob Storage is cost-effective
Each tier is a different bet on how often the data gets read. Get the bet wrong, and the cheaper tier costs more than hot.
Premium block blob
What's included: SSD-backed storage with consistently low latency and the lowest transaction rates of any tier.
Best for: Interactive applications, real-time analytics, and anything where per-operation cost outweighs per-GB cost.
Pros:
- ✅ Transaction charges drop far enough that high-request workloads land cheaper overall than hot
- ✅ Latency stays stable under load, so read timeouts stop appearing in application logs
Cons:
- ❌ The highest capacity rate Azure charges for block blobs, which rules it out for bulk retention
- ❌ Data cannot be tiered into or out of premium with Set Blob Tier or lifecycle policies
Hot tier
What's included: The default tier on new general-purpose v2 accounts, with a 99.9% availability SLA, no retrieval charge, and no minimum retention period.
Best for: Working data, ingestion landing zones, and any object smaller than a few hundred kilobytes.
Pros:
- ✅ Free retrieval removes the biggest source of unmodeled cost in a blob estimate
- ✅ No minimum retention window, so short-lived data can be deleted the day it is written
Cons:
- ❌ The highest standard capacity rate, which compounds fast across petabyte volumes
- ❌ Leaving cold data here is the largest single source of avoidable blob spend
Cool tier
What's included: Millisecond retrieval at a lower capacity rate, with a 99% availability SLA, a $0.01 per GB retrieval charge, and a 30-day minimum retention period.
Best for: Backup copies and datasets read a handful of times per quarter.
Pros:
- ✅ Cuts capacity spend by 44% against hot with no change to retrieval latency
- ✅ Supports every redundancy option, so zone-redundant and geo-redundant copies stay available
Cons:
- ❌ At $0.01 per GB, one full read a month costs more than the $0.008 per GB the tier saves
- ❌ Deleting or rewriting inside 30 days triggers a prorated early deletion charge
Cold tier
What's included: The same millisecond retrieval as cool at a lower capacity rate, with a $0.03 per GB retrieval charge and a 90-day minimum retention period.
Best for: Regulatory copies and historical datasets that need to be reachable without a rehydration wait.
Pros:
- ✅ Cuts capacity spend by 55% against cool while keeping data online and instantly readable
- ✅ Sits between cool and archive, so retention policies can step down without a latency cliff
Cons:
- ❌ Retrieval at $0.03 per GB means a full read every 2.2 months erases the saving against hot
- ❌ The 90-day window makes it a poor fit for data with an uncertain retention decision
Archive tier
What's included: Offline storage at the lowest capacity rate, with a 180-day minimum retention period and rehydration taking up to 15 hours.
Best for: Multi-year retention where a recovery request can wait half a day.
Pros:
- ✅ At $0.00099 per GB, capacity costs about 5% of hot, which reshapes long-horizon retention math
- ✅ Blob metadata and index tags stay readable, so archived objects remain discoverable without rehydration
Cons:
- ❌ Supports LRS, GRS, and RA-GRS only, so zone-redundant accounts cannot use it
- ❌ Reaching the data means a rehydration operation, a retrieval charge, and a wait measured in hours
Archive works when retention policy drives the decision. Matching retention windows to legal and business requirements first, then picking the tier, is covered in more depth in this guide to data archiving strategy and best practices.
Smart tier
What's included: Automatic movement between hot, cool, and cold based on observed access. Objects land in hot, move to cool after 30 days without access, and move to cold after 60 more. Any access returns them to hot immediately and resets the timer.
Best for: Mixed workloads where access patterns are unknown, and lifecycle rules would be guesswork.
Pros:
- ✅ Charges no tier transition, early deletion, or retrieval fees on objects it manages
- ✅ Pins objects under 128 KiB to hot, so small-file workloads avoid cool-tier retrieval charges entirely
Cons:
- ❌ Requires zone-redundant storage, ruling out LRS and GRS accounts
- ❌ Moving an object out is one-way and costs a cool write transaction per object
Smart tier carries a monthly monitoring charge per 10,000 objects over 128 KiB, and versions and snapshots inside it bill at full content length. Microsoft's smart tier documentation sets out the full billing table.
Choosing the right Azure Blob Storage tier
Tier selection comes down to read frequency measured against the minimum retention window. Two numbers decide it, and both are knowable before you commit.
Choose hot if you:
- Read the object more than once a month, or cannot predict when it will be read
- Store large numbers of small files where per-object retrieval charges would dominate
Choose cool if you:
- Read the data roughly once a quarter and will hold it for at least 30 days
- Need a backup copy available in milliseconds without a rehydration step
Choose cold if you:
- Touch the data once or twice a year and have a firm 90-day-plus retention commitment
- Need immediate availability for audit responses without archive latency
Choose archive if you:
- Hold the data for 180 days or longer under a retention policy
- Can tolerate up to 15 hours before a restore request returns data
Choose smart tier if you:
- Run zone-redundant accounts and have no reliable access telemetry to write rules against
- Want tier movement without the transaction volume that lifecycle policies generate
Redundancy options and their cost impact
Redundancy multiplies the capacity line on every stored gigabyte, and it constrains which tiers are available at all.
Geo-redundant accounts carry a per-GB geo-replication data transfer charge every time data is written, which never appears in a capacity-only estimate.
Switching an account from LRS to GRS triggers a one-time charge covering the geo-replication of everything already stored.
An account moved from RA-GRS to GRS or LRS continues billing at RA-GRS rates for 30 more days.
Any account holding archived blobs has to rehydrate all of them before the redundancy setting can change, turning a configuration edit into a five-figure retrieval event on large volumes.
Azure Blob Storage costs beyond the per-GB rate
The tier table sets the floor. Operations, retrieval, egress, and a set of feature meters decide how far above it the invoice lands.
A worked 200 TB Azure Blob Storage estimate
This models a backup and log workload in East US on LRS, split across 40 TB hot, 100 TB cool, and 60 TB archive, with 30-day version retention adding 25 TB of retained blocks in hot.
In a quiet month, the modeled bill is all capacity. In a month with one real recovery, capacity falls to about 70% of the total as retrieval, early deletion, and egress take over, and transaction charges on read-heavy workloads push its share lower still.
Transaction and operation charges by tier
Every read, write, list, and metadata call is billed per 10,000 transactions, and Azure prices those transactions higher as the tier gets cooler. Archive sits at the top of that curve, with read operations costing orders of magnitude more than the same call against hot.
A rehydration job touching a few million archived objects can generate a transaction bill larger than the annual capacity saving that justified archiving them, so price the read side before moving volume down.
There are two mechanics that land outside the tier table:
- Lifecycle policies issue a Set Blob Tier call per object, billed as a write to the destination tier. Re-tiering ten million objects is ten million write transactions.
- Accounts with a hierarchical namespace enabled for Azure Data Lake Storage bill separate iterative read and iterative write meters, and count a transaction per 4 MB of data through the ADLS Gen2 API.
Retrieval and early deletion charges
Cool retrieval is $0.01 per GB and cold is $0.03 per GB. Archive carries its own per-GB retrieval charge on standard priority, with a high-priority option that costs more and returns data faster.
Early deletion applies whenever an object is deleted, overwritten, or moved to a different tier before the window closes. Delete a cool object after 21 days, and Azure charges nine days of cool storage.
Move an archived object to hot after 45 days, and the charge covers the remaining 135 days of the 180-day minimum. Rewriting an object with Put Blob or Copy Blob triggers the same charge.
Soft delete changes the timing. An object with soft delete enabled is not treated as deleted until the retention period expires, so the early deletion clock pauses.
Blob versioning, snapshot, and soft delete billing
Blob versioning is the costliest setting to enable without modeling. With versioning off, an overwrite replaces data. With versioning on, every write creates a new version billed at live-data rates.
When no tier has been explicitly set, Azure charges for unique blocks across the blob, its versions and its snapshots, so shared data is billed once.
Set a tier explicitly on a blob, version or snapshot and billing switches to full content length for that object, a change that cannot be undone.
Setting the tier to archive bills the full content length of every version and snapshot attached to the blob.
A lifecycle rule that tiers versioned blobs to cool therefore converts a deduplicated capacity line into a full-length one, permanently. Soft-deleted objects bill at live rates for the whole retention period on top of that.
Blob feature meters billed separately
Six add-ons bill separately and rarely appear in a planning spreadsheet:
- Blob index tags, charged per 10,000 tags per month
- Blob inventory reports, charged per million objects per report generated
- Change feed, charged per 10,000 recorded changes
- Encryption scopes, charged per scope per month
- SFTP, charged hourly for as long as the endpoint is enabled on the account
- Geo priority replication and object replication priority replication, charged per GB
SFTP is the one worth auditing first. It bills by the hour on the storage account whether or not anyone connects.
Data transfer and internet egress charges
Inbound data is free. Outbound traffic to the internet gets a 100 GB monthly allowance across all regions, then bills by volume tier and by the source region's bandwidth zone.
Zone 1 regions covering the US, Canada, UK and EU start at $0.087 per GB for the first 10 TB, falling to $0.083, then $0.07, then $0.05 above 150 TB.
Zone 2 regions across Asia Pacific and Australia start at $0.12 per GB. Zone 3 regions covering South America, Africa and the Middle East start at $0.181 per GB.
Two exemptions reduce the exposure. Transfer between availability zones inside a region has been free since May 2024, and Microsoft offers free egress for customers moving data out of Azure to another provider or to their own datacenter.
Is Azure Blob Storage worth it?
Azure Blob Storage is worth it when you
- Store large objects with predictable access patterns that map cleanly onto a tier
- Run applications inside the same Azure region, keeping egress near zero
- Have retention policies firm enough to commit to reserved capacity
When Azure Blob Storage costs more than it should
- Store millions of small objects in cool or cold, where per-object retrieval swamps capacity savings
- Serve read traffic to the internet at volume from any tier
- Leave versioning enabled on frequently overwritten data with no lifecycle rule expiring old versions
- Tier data down without measuring how often it is read first
In Eon's 2026 Cloud Data Infrastructure Report, 63% of surveyed cloud leaders said high storage costs push them to retain or protect less data than they should.
That same cost-pressured group was four times more likely to have hit three or more unsuccessful restores in the past year.
Azure Blob Storage alternatives and pricing comparison
Native object storage across the three major clouds prices on the same shape, with capacity, requests, and egress billed separately.
Azure holds the lowest hot-tier capacity rate of the three, and its smart tier automates movement across hot, cool and cold without transition or retrieval fees.
Azure announced a 128 KiB minimum billable object size on cool, cold, and archive but paused it on June 8, 2026, before it took effect, so unlike S3, which bills a 128 KB minimum on Standard-IA and colder classes, Azure currently applies no minimum, as does Google Cloud Storage.
Azure Storage Reserved Capacity terms and exclusions
Azure Storage Reserved Capacity is sold in 100 TB and 1 PB blocks on one-year or three-year terms, with larger discounts on the 1 PB block. It is bought per region, per tier and per redundancy option, and applies at subscription or shared resource group scope.
Three exclusions decide whether it pencils out. Reservations cover hot, cool, and archive capacity only, so cold tier volume stays at pay-as-you-go rates.
Operations, data transfer and early deletion charges sit outside the reservation entirely. Cancellations are refunded pro rata (no early termination fee today, though Microsoft says it may charge up to 12% in future), up to $50,000 in refunds per rolling 12 months; exchanges are free.
For a 200 TB workload, a reservation trims the capacity line and leaves the retrieval, transaction and egress lines untouched. On the model above, that is a discount applied to 55% of a recovery-month bill.
Where Eon fits alongside Azure Blob Storage
Blob storage prices the act of keeping data. It does not price the act of getting a specific thing back. That difference surfaces whenever recovery is the reason the data exists.
Azure Blob Storage is the better fit for: application data planes, static content delivery, data lake landing zones, and any workload where objects are written and read by systems inside the same region.
Eon is the better fit for: protected copies of production data, where the question is which record, table or file needs restoring and how fast it comes back.
Use both when: blob storage holds the operational data and a protection layer holds the recovery copies.
We built Eon's vault to be source-aware, so protected data is deduplicated across the whole cloud environment before it lands, reducing the volume charged against those same per-GB rates. NETGEAR cut its backup storage costs 35% on Eon, and recovery of a 10 TB SQL Server database dropped from 24 hours to under three.
Granular recovery then returns a single record or table without rehydrating a full environment, removing the retrieval and transaction charges a full restore generates.
Lifecycle rules are written once against an environment that keeps changing, and new containers, accounts and subscriptions appear without matching rules.
Cloud Backup Posture Management (CBPM) classifies resources as they are created and applies retention and protection policy automatically. Tiering decisions then follow the data classification, not a rule someone wrote last quarter.
Gartner rates data storage management services a high-benefit category in its 2026 Hype Cycle for Backup and Data Protection Technologies, because it turns storage from a passive cost line into a governed layer.
The same research names cold storage becoming a silo, disconnected from end-to-end lifecycle management, as a primary obstacle to the archival savings organizations expect.
This cloud cost optimization guide covers the adjacent line items in a total cloud bill. The protection layer is covered specifically in this comparison of the best Azure backup solutions for cloud workloads.
The bottom line on Azure Blob Storage pricing
Azure Blob Storage pricing is competitive on capacity and unforgiving on access. The hot rate of $0.018 per GB undercuts both major rivals, and the tiers below it only deliver savings when read frequency stays under the crossover point that the retrieval charge sets.
Measure access frequency per container before writing a single lifecycle rule, keep small objects and unpredictable workloads in hot, and reserve cold and archive for data with a retention decision already made.
Then audit versioning, soft delete and SFTP, because those three inflate bills without appearing on the tier table at all. Nine further tactics for cutting Azure spend without weakening recovery are covered in these Azure backup best practices.
Can you name the exact record you would need back, and price what it costs to retrieve it today? Book a demo and see how Eon deduplicates protected Azure data before it hits a per-GB meter, restores a single file or table without a full rehydration.
Frequently asked questions
How much does Azure Blob Storage cost per month?
Azure Blob Storage costs $0.018 per GB per month in the hot tier in East US on locally redundant storage, which works out to roughly $18.43 per TB. Cool is $0.010 per GB, cold $0.0045 and archive $0.00099. Operations, retrieval and egress are billed on top of those rates.
Is Azure Blob Storage cheaper than Amazon S3?
Azure Blob Storage is cheaper than Amazon S3 on hot-tier capacity, at $0.018 per GB against $0.023 for S3 Standard. Azure currently applies no minimum billable object size, while S3 bills a 128 KB minimum on Standard-IA and colder classes. Request volume and egress price similarly across both.
What is the cheapest Azure Blob Storage tier?
Archive is the cheapest Azure Blob Storage tier at $0.00099 per GB per month, about 5% of the hot rate. It requires 180-day minimum retention, charges for retrieval, and takes up to 15 hours to rehydrate. Cold at $0.0045 per GB is the cheapest tier with immediate availability.
Did Azure introduce a minimum billable object size in 2026?
No. Microsoft announced a 128 KiB minimum for the cool, cold, and archive tiers in 2026, but paused it on June 8, 2026, before it took effect. No minimum billable object size currently applies, so small objects bill at their actual size.
Why is my Azure Blob Storage bill higher than the tier rate suggests?
Capacity is one of five meters. Transaction charges, per-GB retrieval from cool, cold, and archive, geo-replication transfer, internet egress, and feature meters all bill separately. Retained blob versions and soft-deleted data bill at live-data rates until a lifecycle rule expires them.
Does Azure Storage Reserved Capacity cover all blob charges?
No. Reserved capacity covers hot, cool, and archive capacity only. Cold tier volume, operations, data transfer and early deletion charges stay at pay-as-you-go rates. Reservations are bought per region, tier and redundancy in 100 TB or 1 PB blocks.
