BlogMicrosoft Azure

Choosing an Azure Region in Australia: Data Residency Without the Assumptions

Published November 3, 2026 · Updated November 17, 2026

Choosing an Azure Region in Australia: Data Residency Without the Assumptions

Putting a workload in an Australian cloud region is often treated as the end of the data residency conversation. It is closer to the start. The region decides where your main resources run, but services, backups, support access and your own contracts all affect where data ends up and who can reach it.

The Australian regions

RegionLocation
Australia EastNew South Wales
Australia SoutheastVictoria
Australia CentralCanberra
Australia Central 2Canberra

The two Canberra regions are designed for particular needs and access can be restricted, so check Microsoft’s current regions list rather than assuming they are open to your subscription. Most organisations will compare Australia East and Australia Southeast first.

Region pairs and availability zones

Azure groups most regions into pairs for disaster recovery. Australia East is paired with Australia Southeast, and Australia Central is paired with Australia Central 2, so geo redundant storage replicates between two Australian locations by default. Availability zones, which protect against the failure of a single datacentre within a region, are offered in Australia East. Zone support differs between regions and between services, so check the current documentation before relying on zones anywhere else.

This has practical consequences. A design that needs both zone redundancy and a second region for recovery will usually centre on Australia East, with Australia Southeast as the recovery site. Latency matters for some workloads too: users in Perth or Darwin are a long way from Sydney and Melbourne, so test response times from where your users actually are.

What IRAP does and does not settle

Microsoft states that its IRAP assessment covers in-scope Azure services for government data in Australian regions up to the PROTECTED level. That is useful evidence for an agency’s own decision, but it does not authorise your workload. Microsoft’s guidance notes that agencies still implement further controls on a risk-managed basis before authorising a system. An assessed platform is an input, not an outcome.

Two related points are worth knowing. The assessment covers specific services, so check that every service in your design is listed as in scope. And government buyers may also look at the Australian Government’s Hosting Certification Framework, which deals with the ownership and control of the providers and facilities that host government data. Ask for current evidence of both rather than relying on an old announcement.

AI services need their own check

Azure AI services raise residency questions of their own. Model availability varies by region and changes often, so a model you can use in one region may not yet be offered in Australia East. Some Azure OpenAI deployment types are designed to route requests across Microsoft’s global infrastructure for capacity, which means prompts and responses can be processed outside Australia even when the resource is created here. If processing location matters, choose a regional deployment type and confirm in the current documentation what each type does.

What to check before you commit

Service availability

Not every service or AI model is offered in every region. Confirm that what you need is available in your chosen region, today, and not just in the product overview.

Where each service stores data

Some services are global or keep metadata outside the region you select. Check the documentation for each service rather than the platform as a whole.

Backups and disaster recovery

Replication can move copies of data to another region. Choose settings deliberately so your recovery design matches your residency obligations.

Support and administrative access

Residency describes where data sits, not who can access it. Review how administrators, support staff and suppliers reach the environment.

Your obligations

Privacy law, contracts and agency requirements can each set different rules. Start from what you must comply with, then choose the architecture.

Residency is not the same as sovereignty

These words are often used as if they mean the same thing. Data residency describes the physical location of data. Sovereignty raises wider questions about which laws and which organisations can have authority over it, including when a provider is based overseas. If a requirement or a vendor uses the word sovereign, ask exactly what it means in that context.

Services that do not live in one region

Choosing a region covers the resources you deploy, but some of the services around them work differently. Microsoft Entra ID, the identity service behind Azure, is not deployed to a region you pick and has its own documented data location rules. Microsoft 365 is a separate service again, and where its data is stored depends on the country your tenant was set up in. Monitoring, security and management tools can each collect telemetry with their own storage location. None of this is a reason to avoid them, but each one belongs in your residency design rather than being assumed to follow the region.

A residency checklist for each workload

QuestionWhere to find the answer
Which region holds the primary data?Resource location in the Azure portal or Resource Graph
Where are backups and replicas kept?Backup vault and storage redundancy settings
Which services process the data, and where?The data residency section of each service’s documentation
Who can access it, including support staff?Role assignments, privileged access settings and Customer Lockbox
Where do logs and telemetry go?Diagnostic settings and Log Analytics workspace locations
What obligations apply to this data?Privacy, contract and agency requirements, before any design decision

Cost is worth checking at the same time. Prices for the same service can differ between regions, so compare the specific services and sizes you need rather than assuming the regions cost the same.

Turning residency rules into guardrails

A residency decision only holds if the platform enforces it. Otherwise one well meaning engineer can create a resource in the wrong region and nobody notices. These controls turn a policy into something Azure checks for you.

Allowed locations policy

Azure Policy includes a built in policy that limits which regions resources can be created in. Assign it at the management group level so new subscriptions inherit it automatically.

Tags and inventory

Tag resources with their data classification and owner, then use Azure Resource Graph to report where everything actually sits.

Customer Lockbox

For supported services, Customer Lockbox requires your approval before Microsoft engineers can access your data during a support case.

Customer managed keys

Keeping encryption keys in Azure Key Vault under your control, for services that support it, helps when you need to show who can decrypt data.

Logs in the same place

Send activity and diagnostic logs to a Log Analytics workspace in an Australian region, so your audit trail follows the same residency rules as the data it describes.

Frequently asked questions

Which Azure region should most Australian organisations use?

For most workloads, Australia East is the usual starting point because it offers availability zones and a wide range of services, with Australia Southeast as its paired recovery region. The Canberra regions suit specific government and critical infrastructure needs, and access to them is restricted.

Does keeping data in Australia stop foreign governments accessing it?

Not necessarily. Residency controls where data is stored, but a provider headquartered overseas may still be subject to the laws of its home country, such as the US CLOUD Act. If this matters to you, get legal advice and look at contract terms, encryption and key control, not just the region.

Is all of Azure IRAP assessed?

No. Microsoft’s IRAP assessment covers a defined list of in scope services in its Australian regions. Check the current list for every service in your design.

Can we stop staff creating resources outside Australia?

Yes. Assign the built in allowed locations policy in Azure Policy at a management group, and review the exceptions it needs, such as global services that do not belong to a single region.

Do Azure OpenAI models keep data in Australia?

It depends on the deployment type. Regional deployments process data in the region you choose, while global deployment types can process prompts in other regions. Check both the deployment type and model availability before you commit.

What is the difference between Australia East and Australia Southeast?

Australia East, in New South Wales, offers availability zones and generally the wider range of services. Australia Southeast, in Victoria, is its paired region and a common choice for disaster recovery or for users closer to Melbourne. Check that every service you need is available in both before you decide.

Can we use the Canberra regions?

Australia Central and Australia Central 2 were designed with government and critical infrastructure in mind, and access has to be requested rather than being open to every subscription. Most commercial workloads run in Australia East or Australia Southeast instead.

A simple test: for each dataset, can you state where it is stored, where it is backed up, which services process it and who can access it? If not, the residency design is not finished.