Switching workloads to Azure Reserved Instances or Savings Plans can help you save more money on Microsoft Azure services. So, what are these reservations and plans? What kind of workloads can you run with them? And what’s the catch in each case for getting up to 80% off?
Here’s your simplified guide to Microsoft Azure reservations and Savings Plans.
What Is An Azure Reservation?
An Azure reservation is a commitment to use a specific virtual machine (VM) at a fixed capacity (in dollars/hour) for a period of one or three years in exchange for discounts of up to 72% off the standard rate.
Using Windows virtual machines gets you up to 80% savings. If you combine an Azure reservation with an Azure Hybrid Benefit plan, you can also save up to 80% off the pay-as-go rate (standard rate).
Exactly how much you can save with Azure reservations will depend on several factors, including the region, VM type, commitment term, payment option, operating system, and more. See the image below or, better yet, see our no-BS guide to AWS vs Azure pricing here.
Credit: Example of how Azure Reserved VM pricing works by Microsoft Azure
What are the features of Azure Reservations?
Azure reserved instances have the following characteristics:
- You get greater discounts the longer you commit.
- You pay a fixed hourly price throughout the length of your 1- or 3-year contract for the eligible computing resources. This guarantees availability of this capacity for your use throughout the contract term.
- You will be billed separately for usage that exceeds your hourly capacity. This overage is automatically billed at the undiscounted pay-as-you-go rate and comes in a separate bill.
- When you do not use all your hourly allocation, and the VM that was using it has been terminated, you lose the capacity unless you have another VM that is eligible to use the balance.
- Azure Reservations can apply across different subscriptions and resource groups but only within a single region using a single VM type.
- Azure Reserved VM instances apply to compute, database, storage, and app services.
- You can also apply the RIs to Windows or Linux instances in Azure.
- You can choose to pay upfront or monthly at no additional cost – although the upfront payment option offers the highest savings.
- Azure still allows you to exchange compute reservations (VMs, Dedicated Hosts, and App Service) with each other, provided they are of the same type. This exchange availability has been extended until further notice — Microsoft will provide at least six months’ advance notice before ending exchanges. You can also trade in reservations for a savings plan at any time, with no deadline.
- You cannot exchange reservations of different types, such as Azure SQL databases with Azure Cosmos DB reservations. Learn more about Azure storage costs in this snackable guide.
- In addition, you can exchange a reservation for another one of the same type in a different region — for example, swapping West US 2 for West Europe — as long as compute reservation exchanges remain available. Reservations also offer VM size flexibility, meaning the discount can apply to different VM sizes within the same instance size flexibility group, so you are not locked to a single exact VM size.
- You can cancel Azure reservations and receive a refund, subject to a $50,000 cap in a 12-month rolling window per billing profile or enrollment. There is currently no early termination fee, although Microsoft may introduce a 12% fee in the future.
These are some of top considerations to remember when considering using Azure reservations or Reserved Virtual Machine instances on Azure.
So, what are the benefits and limitations of Azure Reservations?
Azure Reservations Pros
Using Reserved VM instances on Azure offers the following advantages:
- Azure RIs offer greater discounts than Savings Plans. You save up to 72% off the standard rate on Linux instances or up to 80% using Windows VMs.
- Secure a guaranteed computing capacity for a fixed price over a 1- or 3-year period, enabling you to plan your utilization.
- This cost model is stable and predictable for similarly predictable workloads.
- Azure reservations are ideal if you don’t anticipate many changes in infrastructure or workload over a one- or three-year period.
- If you don’t have the capital or want to reduce the cash outlay, you can pay monthly at no additional cost. Or, you can pay cash upfront to get the highest discount.
Yet, Azure reservations have some noteworthy catches.
Azure Reservations Cons
The following are some of the limitations of using Reservations on Azure:
- Azure Reservations apply to a single virtual machine type within its flexibility group — and only within a single region.
- Flexibility is traded off for predictability when using reservations.
- Reserved instances are a pricing discount, not a capacity guarantee. For workloads requiring guaranteed availability, Azure’s On-Demand Capacity Reservations are a separate mechanism.
- If your workloads change significantly during the commitment period (due to migration, right-sizing, or sunsetting products), unused reserved capacity is forfeited.
Now let’s explore how Azure Savings Plans for Compute differ.
Research Report
FinOps In The AI Era: A Critical Recalibration
What 475 executives told us about AI and cloud efficiency.
What Is An Azure Savings Plan For Compute?
By committing to spend a specific dollar per hour amount over a 1- or 3-year period, Azure Savings Plans can save you up to 65% off the standard Azure compute rates.
However, Savings Plans apply to some Azure services, but not to others. Exactly how much you can save with Azure Savings Plans varies based on factors such as regions and payment options.
How Savings Plans pricing works in Azure – Microsoft Azure
What are the features of Azure Savings Plans for Compute?
Consider the following Azure Savings Plans attributes:
- Savings Plans in Azure only apply to compute costs, not storage, network, licensing, etc.
- However, Azure Savings Plans apply across any compute resources that are within the scope of the Savings Plans. This includes Azure Dedicated Hosts, Azure Virtual Machines, Azure App Service, Azure Functions Premium Plan, and Azure Container Instances.
- Savings Plans apply across Windows and Linux VMs/instances.
- You can apply Azure Savings Plans across compute resources using different operating systems (Linux vs Windows), Azure Regions, and VM type (outside of the flexibility group), making them more flexible than Azure Reservations.
- The discount applies to your commitment threshold. Any usage above this commitment is billed at the undiscounted pay-as-you-go rate and separately.
- When both Reservations and Savings Plans are applicable, Reservations apply first because they offer greater discounts.
- As with Reservations, you pay a fixed dollar/hour amount for a 1- or 3-year period when using Savings Plans.
- Paying all-upfront will save you the most money. You can also pay monthly to reduce your capital outlay.
- Azure does not allow you to exchange, return, or cancel Savings Plans. You can purchase additional Savings Plans, but existing ones run their full term.
- Savings Plans automatically apply to eligible compute usage almost immediately after you purchase them.
Here are other reasons you’ll want to use Savings Plans on Azure.
Azure Savings Plans Pros
Some top advantages of Azure Savings Plans include:
- Save up to 65% on compute resource costs across Azure VMs, Dedicated Host, App Service, Container Instances, and Functions Premium Plan.
- Get greater cost flexibility to adapt to changing Azure cloud infrastructure needs.
- Azure SPs are not restricted to just a single region or VM type.
- If you prefer to reduce your cash outlay or don’t have the capital, you can pay in part each month without any additional charge. You can also pay cash upfront to get the highest discount.
- Discounts apply automatically — your engineering teams don’t need to coordinate workload placement to consume the commitment, which reduces operational overhead compared to managing reserved instances.
Yet, the Savings Plan approach has a few drawbacks.
Azure Savings Plans Cons
Using Savings Plans on Azure has some limitations:
- SPs only apply to compute usage costs. Storage, networking, and other non-compute charges remain at standard rates, which can limit the overall impact on your total Azure bill.
- SPs also apply to select Azure resources.
- Billing for additional usage is charged on a pay-as-you-go basis and invoiced separately.
- You cannot cancel, modify, or exchange Savings Plans once purchased. This makes accurate forecasting critical — organizational changes like migrations or product sunsets can reduce the value of an existing commitment.
So, how do Azure Savings Plans and Reservations compare?
Azure Reservations Vs Savings Plan: What Is The Difference Between Reserved Instances And Savings Plans In Azure?
|
Azure Reservations |
Azure Savings Plans | |
|
Commitment description |
Specify the VM type and computing capacity you will use per hour over 1 or 3 years Decide on your term length, billing frequency, type of VM, Azure Region, and commitment level |
Specify how much you will spend per hour on Azure compute resources over a 1- or 3-year period Decide on your commitment ($/hour), billing frequency, term length, and the Region to apply your savings |
|
Maximum discount |
72% (Linux OS) or 80% (Windows OS) |
65% on Compute usage |
|
Applies to |
A specific Azure Region and VM type |
Across workloads and resource groups, regardless of VM type outside of the flexibility group, OS, and Region |
|
Limited to |
Can apply to 16 Azure services across eligible compute, database, app services, and storage services |
Currently only work for Azure compute resources for Dedicated Host, Virtual Machines, App Service, and Functions Premium Plan |
|
Cancellations policy | Cancellation allowed; refunds subject to a $50,000 cap in a 12-month rolling window per billing profile. No early termination fee currently charged. | No cancellations allowed — you can buy another Savings Plan instead |
|
Exchange/Trade-in | Exchange availability extended until further notice (at least 6 months’ notice before any change). Trade-in to Savings Plans available anytime. | No exchanges or trade-ins allowed — you can buy a different Savings Plan instead |
| Management overhead | Higher — requires monitoring utilization and matching usage to reserved configuration | Lower — discounts apply automatically across eligible compute |
Are Azure Savings Plans Replacing Reserved Instances?
Not at all. Azure Savings Plans and Azure Reserved VMs are different discount instruments. Both are applicable to eligible Azure cloud resources, and Microsoft continues to support both.
What has changed is the exchange policy for compute reservations: the originally planned deadline to end exchanges (January 1, 2024) was extended multiple times and is now available “until further notice.” Microsoft will give at least six months’ notice before removing the exchange capability. In the meantime, you can always trade in reservations for a savings plan with no time limit.
Wherever Reservations and Savings Plans are both applicable, Reservations apply first since they offer the deepest discounts. If you have several or overlapping Savings Plans, the one with the highest discount is preferred.
Check out this article to see how AWS savings plans differ from reserved instances.
What To Consider When Switching To A Savings Plan
Keep these three considerations in mind to take full advantage of Savings Plans on Azure.
1. Hourly commitment
You specify the dollar amount you are willing to spend every hour for one or three years. A flat fee is charged for this commitment level. Overages are charged on a pay-as-you-go basis.
2. Billing
You’ll receive a monthly bill for compute usage for the entire month, even though billing is done per hour. You lose any capacity below your hourly threshold if you don’t have an eligible compute VM to use it.
3. Cancellation and modifications
Once you commit to Savings Plans on Azure, you cannot cancel or modify them until they expire. Factor this into your decision — if your compute baseline is likely to shrink due to migrations, right-sizing, or product changes, a smaller commitment with room for pay-as-you-go overflow may be safer than over-committing.
Should You Use A Reservation Or Savings Plan?
Azure Reservations are best used when the workload is predictable, you do not plan significant infrastructure changes in your operating regions for the contract period, and could use a specific VM instance type with no performance issues. Teams running stable production databases on SQL Managed Instance or Cosmos DB with consistent throughput, for example, tend to get the deepest discounts from reservations.
Conversely, Azure Savings Plans are best used when your compute needs are likely to change over time — frequent scaling events, architectural shifts across VM families, or workloads moving between regions. Because the commitment is spend-based rather than tied to a specific configuration, the discount continues to apply even as your compute footprint evolves.
Combining Both For Maximum Savings
In practice, many organizations get the best results by layering both instruments:
- Right-size first — remove unused or oversized resources before purchasing any discounts. Discounts reduce rates, not waste.
- Reserve the baseline — commit the predictable, always-on portion of your compute (the VMs that never change) to Reserved Instances for the deepest discount.
- Cover the variable layer with Savings Plans — apply a Savings Plan sized to your remaining eligible compute baseline that shifts across VM families, regions, or services over time.
- Pay as you go for spikes — leave genuinely unpredictable burst capacity on standard rates rather than over-committing.
When both instruments apply to the same usage, Azure applies Reservations first (deeper discount), then Savings Plans for additional eligible usage. This layered approach maximizes overall discount coverage while minimizing the risk of stranded commitments.
For a broader look at Azure optimization strategies, see our guide to Azure cost management tools and 5 Azure optimizations that make cloud cost savings simple.
How To Optimize Your Azure Costs Beyond Reservations And Savings Plans
Savings Plans and Reserved VM instances can help you lower Azure costs without compromising performance. But discount instruments alone only address your rate — they don’t tell you whether the resources you’re paying for are actually delivering value.
CloudZero provides real-time, granular cost intelligence that goes beyond discount management. By connecting cost data to the business dimensions that matter — per feature, per team, per customer — CloudZero helps you understand not just what you’re spending, but whether that spending is generating returns.
CloudZero collects cost metrics for tagged, untagged, and untaggable cloud resources.
With a unique, code-driven approach, it simplifies cost data into immediately actionable insights, according to your role:
- Engineers view insights based on their technical needs, such as cost per environment, per feature, per deployment, etc. Using this method, they can determine the impact of technical changes on costs so they can build competitive, cost-effective solutions.
- Finance gets insights into the costs for supporting each individual customer, for each project, for each team, for each product, for each environment. Your team can then calculate ROI on each cost center and set prices based on the usage of each client, so you can protect your margins.
- FinOps examines SaaS metrics like cost of goods sold (COGS), calculates gross margins, and performs showbacks/chargebacks. With better cost allocation, they can better perform budgeting and forecasting.
For organizations managing Azure alongside AWS or GCP, CloudZero provides a single view of all cloud spend — so you can apply the same cost intelligence framework across every provider.
Want to join this elite group of cloud cost champions? 
to get started.
Software and pricing information last verified May 2026. Features, pricing, and availability may have changed. Please verify current details with Microsoft Azure before making commitment decisions.
FinOps In The AI Era: A Critical Recalibration
What 475 executives told us about AI and cloud efficiency.
