Unlocking Azure SQL Pricing: How Much Does it Really Cost?
Trying to find clear information about Azure SQL pricing is like walking into IKEA for “just a lamp”. You step in, and suddenly, you’re winding through kitchens, sofas – you name it. By the time you reach the checkout, you’ve most likely forgotten what you came for.
That’s what browsing Azure SQL pricing feels like. Microsoft provides you with calculators, comparisons, and a bunch more stuff – leaving you more confused than before.
And not only that: Azure SQL pricing is complex to gauge because it’s very complicated due to its pricing structure (think deployment methods, server tiers, etc). Not to mention the staggering number of different offerings and pricing models.
Look no further if you are just looking for a simple overview of Azure SQL Pricing.
Azure SQL Database is a fully managed relational database with AI-driven features that optimise performance and security. As one of Microsoft’s many PaaS services, Azure SQL Database takes care of operational concerns so you can focus on building and running your applications. It is excellent for modern cloud applications that need scalability, high availability, and seamless integration with other Azure services.
It's often the go-to option when moving to the cloud and needing a database management system. There’s a good reason for that – it’s easy to deploy, simple to manage, cost-effective and includes built-in scaling and straightforward pricing – which we’ll get back to.
Use Case
Best for SaaS apps that need multi-tenancy and can scale quickly. You can start small and grow as needed – ideal for apps with unpredictable or spiky traffic.
For example: An online retailer expecting a 10x traffic spike during Black Friday. Azure SQL Database scales automatically to handle it – no manual intervention is needed.
What is Azure SQL Managed Instance?
Azure SQL Managed Instance is a fully managed database service offering close to full compatibility with the latest SQL Server (Enterprise Edition). Its broad compatibility makes it ideal for lift-and-shift migrations from on-premises SQL. It blends the advantages of PaaS with the feature set of traditional SQL Servers, making it ideal for migrating existing apps that rely on SQL Server-specific capabilities.
It delivers the core features of SQL Server while handling updates, backups, high availability, and security as part of the managed service. This allows you to move to the cloud with minimal changes to your current setup.
Use Case
Perfect for migrating on-prem SQL Server workloads with minimal changes. Azure SQL Database itself might not always be compatible, but a Managed Instance usually will be.
For example: a company migrating its core business systems to the cloud can leverage Managed Instance's SQL Server compatibility. Even when it doesn't work with Azure SQL Database, it typically works with Managed Instance. This gives you the benefits of abstracted management, easy security configuration, and less overhead.
SQL Server on an Azure Virtual Machine (VM) lets you run full SQL Server instances on Windows or Linux VMs inside Azure. Unlike the managed options, it’s not as scalable or hands-off. You’re responsible for maintaining the OS, managing updates, and configuring the environment. But it gives you complete control and flexibility.
It’s the best fit when your applications need specific SQL Server features or custom configurations that aren’t supported by Azure SQL Database or Managed Instance. You can run SQL Server on both Windows and Linux VMs, and it’s ideal for workloads that need features or custom setups not supported in managed SQL offerings like Azure SQL Database or SQL Managed Instance.
Use Case
Use when PaaS options aren’t a fit and complete control is required, even if it means managing the OS and all the overhead that comes with it – patching, antivirus, browsers, etc.
Example: A healthcare provider with strict compliance needs that require custom SQL Server configurations. They choose Azure VM for control and flexibility, while still using cloud benefits like backup and disaster recovery. The trade-off: more to manage.
Now we understand these SQL database options, let’s look at what you came for: pricing. We’ll explore them for each of the three options, starting off with Azure SQL Database.
Azure SQL Database Pricing
The different Azure SQL Database Deployment Models
In total there are three deployment options for Azure SQL Database:
Single Database
Elastic Pool
Manage Instance
Single Database
This model represents a fully managed database. With Single Database, each database runs with dedicated resources. It is suitable for most modern cloud applications and microservices. It's the most basic deployment model and is ideal for applications with a specific, singular focus.
Elastic Pool
It is for businesses with multiple smaller applications. In Elastic Pool, multiple databases share the same set of resources on the same infrastructure. This makes Elastic Pools more cost-effective for workloads with variable usage across databases.
Managed Instance
This model suits businesses moving from on-prem SQL Server to Azure. It replicates core SQL Server features within Azure, making it easier to migrate apps with minimal rework.
Azure SQL Database Deployment Models
Deployment Model
Best For
Cost Efficiency
Key Benefits
Single Database
Cloud-first apps needing isolated performance
Moderate
Serverless option, auto-scaling, easy to manage
Elastic Pool
Apps with multiple databases and variable usage
High
Shared resources, budget-friendly for inconsistent loads
Managed Instance
Moving on-prem SQL Server workloads to the cloud
Moderate
Full SQL Server feature set, VNet integration, easy lift-and-shift
What are the Azure SQL Database Pricing Models?
Azure SQL Database pricing is built with flexibility in mind and offers two main models:
DTU (database transaction units) based
vCore-based
Let’s start with DTU, which is for predictable and simplified pricing.
DTU (Database Transaction Unit) Model
This is the original and simpler option. It abstracts underlying hardware into a single unit of measure. Instead of separating CPU, memory, and I/O, it bundles it into a single unit called a DTU (Database Transaction Unit). This makes it easier for users without deep database tuning knowledge. It’s a good fit for infrastructure teams looking for straightforward pricing without having to fine-tune resource allocations.
As you scale DTUs up, database max size and costs increase correspondingly. You can buy additional storage beyond default limits.
The DTU purchasing model has predefined performance tiers: basic, standard, and premium, each with fixed resources and corresponding limits.
Basic: Starts at $4.8971 per month, best for small, low-usage applications.
Standard: Prices start as low as $14.7187 a month and can reach up to $4,415.5875 monthly.
Premium: The lowest price is $456.25 a month, with $15,698.89 as the most costly.
The table below breaks it down further for each:
Basic
DTUs
Included storage
Max storage
Price for DTUs and included storage
B
5
2 GB
2 GB
$4.8971/month
Standard
DTUs
Included storage
Max storage
Price for DTUs and included storage 1
S0
10
250 GB
250 GB
$14.7187/month
S1
20
250 GB
250 GB
$29.4343/month
S2
50
250 GB
250 GB
$73.6084/month
S3
100
250 GB
1 TB
$147.1772/month
S4
200
250 GB
1 TB
$294.3725/month
S6
400
250 GB
1 TB
$588.7450/month
S7
800
250 GB
1 TB
$1,177.4900/month
S9
1600
250 GB
1 TB
$2,354.9800/month
S12
3000
250 GB
1 TB
$4,415.5875/month
Premium
DTUs
Included storage 1
Max storage
Price for DTUs and included storage 2
P1
125
500 GB
1 TB
$456.2500/month
P2
250
500 GB
1 TB
$912.5001/month
P4
500
500 GB
1 TB
$1,825.0000/month
P6
1000
500 GB
1 TB
$3,650.0000/month
P11
1750
4 TB
4 TB
$6,868.3876/month
P15
4000
4 TB
4 TB
$15,698.8934/month
Note that all prices are in US dollars, with Eastern US as the picked region.
vCore-based model (Provisioned and Serverless)
A vCore (virtual core), on the other hand, provides granular control over resource allocation. It allows users to choose the exact number of cores, offering flexibility to scale resources up or down based on specific workload requirements. Users who prefer precise control to resource allocation typically choose the vCore model. It’s simply ideal for DBAs who want detailed control over performance and cost. In addition, it is also more suitable for complex workloads.
Service Tiers within vCore Pricing
General Purpose: Suited for most business workloads, this tier offers balanced, scalable compute and storage. Compute and storage are separated, giving you flexibility and better cost control. Pricing starts at around $0.001 per vCore per second, making it a cost-effective choice for standard use cases.
Business Critical: Business Critical (or Premium) is built for production apps that need low latency and high resilience. It uses local SSDs for fast performance and replicates data across four nodes for fault tolerance. One replica can be used for read operations, and zone redundancy is included at no extra cost. Pricing starts at $1.35/hour for 2 vCores, with storage at $0.25/GB a month. You get top-tier performance and built-in high availability, but max storage is limited to 4T.
Hyperscale: Hyperscale is architected to be a cloud-native database. Opting for means fully taking advantage of everything the cloud has to offer. Unlike Business Critical, storage in Hyperscale is decoupled from compute. That means you can run a 2 vCore database with up to 100 TB of storage with no size-to-vCore constraints. Performance matches or exceeds Business Critical in internal tests, making Hyperscale ideal for high-scale, high-performance workloads. Prices start from $0.366/hour, and storage costs $0.25GB monthly.
Compute Tiers in vCore Pricing
With the vCore model, you have two compute tiers:
Provisioned Compute: Dedicated resources billed per hour.
Serverless Compute: Resources are allocated on-demand and billed per second based on the vCores used.
Serverless lets you define min/max vCores, offering flexibility. The same is not true for Provisioned, where you have a fixed number of vCores.
The table below compares Provisioned and Serverless tiers by usage, scaling, and billing:
Tier
Provisioned
Serverless
Usage
Best for steady workloads with consistent compute needs
Suited for variable or unpredictable workloads with lower usage
Scaling
Manual scaling, managed by you
Auto-scaling, managed by Microsoft
Billing
Billed per hour
Billed per second
Typically, Azure SQL Database pricing includes both SQL Server and Windows Server licensing costs. With Azure Hybrid Benefit, you save up to 40% by using existing SQL Server licences, paying only for Windows Server licensing. Without it, you’ll pay the standard rates.
The pricing below applies to the East US region, based on Single Database deployment using the general purpose service tier:
vCORE
Memory (GB)
Pay as you go
1 year reserved capacity 2
3 year reserved capacity 2
2
10.2
$0.505/hour
$0.398/hour 21% savings
$0.337/hour 33% savings
4
20.4
$1.009/hour
$0.796/hour 21% savings
$0.674/hour 33% savings
6
30.6
$1.514/hour
$1.194/hour 21% savings
$1.011/hour 33% savings
8
40.8
$2.018/hour
$1.592/hour 21% savings
$1.348/hour 33% savings
10
51
$2.522/hour
$1.990/hour 21% savings
$1.685/hour 33% savings
12
61.2
$3.027/hour
$2.388/hour 21% savings
2.022/hour 33% savings
14
71.4
$3.531/hour
$2.786/hour 21% savings
$2.359/hour 33% savings
16
81.6
$4.035/hour
$3.183/hour 21% savings
$2.696/hour 33% savings
18
91.8
$4.540/hour
$3.581/hour 21% savings
$3.033/hour 33% savings
20
102
$5.044/hour
$3.979/hour 21% savings
$3.370/hour 33% savings
24
122.4
$6.053/hour
$4.775/hour 21% savings
$4.043/hour 33% savings
32
163.2
$8.070/hour
$6.366/hour 21% savings
$5.391/hour 33% savings
40
204
$10.088/hour
$7.958/hour 21% savings
$6.739/hour 33% savings
80
396
$20.175/hour
$15.915/hour 21% savings
13.477/hour 33% savings
128
652
$32.280/hour
$25.464/hour 21% savings
21.563/hour
But that’s not all. Expect additional costs beyond compute:
Storage is billed separately, and premium Azure Blob Storage comes at a price tag of $0.115 per GB/month (LRS).
If you need back up storage (also known as point-in-time recovery) it will be at least $0.10 per GB/month (LRS, ZRS)
For long-term retention, you pay at least $0.05 per GB/month (RA-GRS & RA-GZRS)
Serverless Compute The SQL Database serverless compute tier is designed for single databases with irregular or unpredictable usage. It auto-scales compute as needed and bills per second for what you use, making it cost-efficient and easier to manage.
With the vCore serverless compute option, you can scale between 0.5 to 80 vCores and 2.02 GB to 240GB of memory, adjusting resources as needed.
Compute cost: $0.5218 per vCore/hour
Storage: $0.115 per GB/month
Minimum vCores
Maximum vCores
Minimum Memory (GB)
Maximum Memory (GB)
Price
0.5
80
2.02
240
$0.0001450/vCore-second ($0.5218/vCore-hour)
Again, additional charges also incur, such as:
Back up storage (Point-in-time restore): Min. $0.10 per GB/month (LRS, ZRS)
Long-term retention: At least $0.05 per GB/month (LRS, ZRS)
Pricing models for SQL Managed Instance are based on vCore. You effectively build a virtual machine starting for less than $800 a month. You choose CPU (standard, premium or premium series memory-optimised) and select the amount of vCores you want, each with an allocated amount of RAM. Then, you must select either Business Critical or General Purpose.
The table below breaks down the cost for Azure Managed Instance (next generation general purpose) with East US as the selected region:
Standard-series (Gen 5)
vCORE
Memory (GB)
Included Storage
Pay-as-you-go
4
20.4
First 32 GB/month
$736.378/month
8
40.8
First 32 GB/month
$1,472.755/month
16
81.6
First 32 GB/month
$2,945.510/month
24
122.4
First 32 GB/month
$4,418.264/month
32
163.2
First 32 GB/month
$5,891.019/month
40
204
First 32 GB/month
$7,363.773/month
64
326.4
First 32 GB/month
$11,782.037/month
80
408
First 32 GB/month
$14,727.546/month
vCORE
Memory (GB)
Included Storage
Pay-as-you-go
4
28
First 32 GB/month
$805.821/month
6
42
First 32 GB/month
$1,208.732/month
8
56
First 32 GB/month
$1,611.642/month
10
70
First 32 GB/month
$2,014.552/month
12
84
First 32 GB/month
$2,417.463/month
16
112
First 32 GB/month
$3,223.283/month
20
140
First 32 GB/month
$4,029.104/month
24
168
First 32 GB/month
$4,834.925/month
32
224
First 32 GB/month
$6,446.566/month
40
280
First 32 GB/month
$8,058.208/month
48
336
First 32 GB/month
$9,669.849/month
56
392
First 32 GB/month
$11,281.491/month
64
448
First 32 GB/month
$12,893.132/month
80
560
First 32 GB/month
$16,116.415/month
96
560
First 32 GB/month
$19,339.698/month
128
560
First 32 GB/month
$25,786.264/month
Balancing performance and budget is key. Azure SQL makes that easier with flexible, live scaling – so you can adjust resources as your needs change, without downtime.
The first 32 GB of storage per instance each month is included, with additional storage billed at $0.115 per GB per month.
SQL Server on Azure VM is priced like any other Azure VM – you pick the SKU, set up the virtual machine, and install SQL Server. Licensing is separate and must be factored in.
This option gives you the most flexibility, letting you choose precisely the specs and setup you need. The ultimate price for running SQL Server on Azure VMs depends on several factors, including VM size, region, and any extra features or configurations you choose.
Entry-level pricing from $1.5140/hour (pay-as-you-go) for a B1s instance with 1 vCPUs and 1GiB of RAM. Conversely, larger VMs with more vCPUs and memory, such as those in the M-series or other high-end configurations, are designed for demanding workloads and are priced accordingly higher.
You can also use Premium SSDs for higher disk performance. This setup is ideal when an SQL-managed instance or SQL Database doesn’t fit your requirements.
That said, we wouldn’t recommend this as your first choice, but only if SQL Managed Instance and SQL Database aren't available for your database. If you're an organisation looking to modernise your application, you'll want to use PaaS instead of IaaS. And wherever possible, SaaS products.
We recommended using the Azure Pricing Calculator to determine the exact pricing for SQL Server Enterprise on Azure VMs.
Want to save on your monthly Azure cost?
Get our Azure Cost Management Whitepaper! With the best tips, tricks, and background knowledge to optimise your cloud costs.
Exchange your existing licences for discounted rates on Azure SQL Database. This can save you up to 30 percent or more by using your Software Assurance enabled SQL Server licences on Azure.
Mind you, Azure Hybrid Benefit can only be applied to Business Critical and General Purpose service tiers. You cannot use it on Hyperscale or the serverless computing tiers.
The image below shows how much you can potentially save using AHB for a SQL Enterprise Savings Example:
With Azure Reserved capacity, you commit upfront to use Azure SQL Database resources for either 1-or 3-years. In return, you’ll receive a discount on provisioned compute. It can save you up to 33%. Obviously, the longer you commit, the greater the savings will be.
Better yet, you save up to 80% when combined with Azure Hybrid Benefit for SQL Server.
Azure Reserved Capacity discounts cover Business Critical, General Purpose and Hyperscale service tiers. Consider that the serverless compute tier is not eligible for this.
Use serverless for dev/test and variable workloads
The serverless tier is ideal for workloads with unpredictable or intermittent usage. It auto-scales compute based on demand and pauses during idle periods, so you’re only billed per second for what’s used.
New workloads where usage patterns aren’t yet known
To optimise savings, define realistic min/max vCore settings and enable auto-pause after periods of inactivity – that way, your resources can be reclaimed, and you're not being charged for those.
Of course, there’ll be some scenarios where provisioned is better (like when running a database constantly for the whole month). But for Dev test environments where you're just starting out, it's a great thing to think about.
Take advantage of the always-free tier
Azure SQL now offers an always-free serverless database with:
You might be asking: what can you actually do with 100,000 vCore seconds?
It depends on how you configure your workload. You can spread it out over the month at lower vCore levels or use higher vCores and consume it more quickly. The serverless min/max settings give you that control.
This offer renews monthly and works with any Azure account. It’s the most cost-efficient way to get started with Azure SQL for proof of concept or lightweight development without any upfront cost. Use it early in your project lifecycle to keep experimentation and development free.
Use Elastic Pools to reduce cost for variable workloads
Switching databases to elastic pools is another way to save money for your SQL databases.
How it works? Elastic pools let you group multiple databases with different usage patterns into a shared resource pool. Instead of paying for peak capacity on each database, you allocate a shared set of vCores, assuming not all databases will be active at once.
Example: If you have 8 general-purpose databases using 2 vCores each (16 vCores total), but they're rarely all active at the same time, you can move them to a 4 vCore elastic pool—saving up to 80%.
Elastic pools work best when:
You manage many databases with non-overlapping peak usage
In General Purpose, you can add up to 500 databases to one pool; in Business Critical, the limit is 100. The key is to understand your workload patterns and right-size the pool accordingly.
Azure Cost Scan
We'd like to help you make sense of your costs. Our in-house experts provide a professional savings recommendation, based on your current Azure cloud usage.
Backup costs can add up quickly, especially if you don't tailor your settings to the workload. To save:
Choose the right storage type: Use lower-cost options like LRS for Dev/Test environments. Avoid more expensive types like RRS (Read-Access Geo-Redundant Storage) or RA-GRS unless needed for critical production data.
Adjust retention settings: Don’t default to 35-day retention for everything. For Tier 1 workloads, longer retention makes sense. For lower tiers or Dev/Test, reduce PITR to 7 days or less to cut storage usage.
Be cautious with LTR (long-term retention): For high-change databases, LTR (long-term retention) costs can spike due to frequent backups. Limit LTR use to where compliance or business needs truly require it.
Right-sizing backup strategy to match workload sensitivity and criticality can prevent unexpected and unnecessary spend.
Closing thoughts
As you know, between deployment models, compute tiers, service tiers, and so forth.. it’s easy to lose track of what you need – and what it will cost. We hope this article has given you a clear overview of Azure SQL Pricing along with some tips to save cost.
The bottom line? Match your choice to your usage patterns and use case. For example, using RRS or RA-GRS for non-critical workloads is not always necessary.
Also, use the pricing calculator, monitor your workloads and revisit your setup, as things always change.
Azure SQL pricing is based on instance type, number of DTUs and storage. For example:
An S0 instance (10 DTUs, up to 250 GB storage) runs at $0.0202 per DTU/hour ($14.7187/month).
An S12 instance (3,000 DTUs, up to 1 TB storage) jumps to $6.0488 per DTU per hour ($4,415.5875/month).
Why is Azure SQL so expensive?
Azure SQL can be expensive for a few reasons. Costs add up fast when you add features like advanced security, monitoring or support plans. Licensing (especially for Microsoft tools like SQL Server) can be a big part of the bill.
Pricing is complex and varies by service, usage and pricing model. With the DTU (DTU (database transaction units) based) model for example you can overpay for unused capacity or under provision and suffer poor performance. If you’re not careful with what you choose and how much you use, costs add up quick.
What are the tiers of Azure SQL?
Azure SQL has three service tiers:
General Purpose: For most business workloads. Separates compute and storage. Cost-effective and flexible.
Business Critical: For high performance apps that need low latency and high availability. Uses local SSDs and multiple replicas.
Hyperscale: For large and high performance workloads. Supports up to 100 TB of storage and separates compute from storage for fast scaling.
Is SQL database free on Azure?
Azure SQL Database isn’t free but there is an always-free option. Microsoft offers a free serverless tier that includes: