For a 5 to 20 person NZ firm, Dataverse and Azure SQL cost the same per user, because either way the people using the app need Power Apps Premium at NZ$32.40 a month. What differs is the database bill (Dataverse's is included to 10 GB; Azure SQL is a separate USD line from about USD 5 a month upward) and who runs it (Microsoft runs Dataverse end to end; Azure SQL still needs someone to own schema, backups and access). The row that decides it: where does the logic live? In T-SQL, Azure SQL. In the tables and forms, Dataverse.
What is each option and what does it do?
| Azure SQL Database | Dataverse | |
|---|---|---|
| What it is | SQL Server, run by Microsoft in Azure. Tables, T-SQL, stored procedures, views. A database and nothing else | A managed data platform for Power Apps: tables plus security roles, auditing, business rules, formula columns, forms, and the connector Power Apps is built around |
| Licence for app users | Power Apps Premium, NZ$32.40 per user per month (the SQL connector is premium) | Power Apps Premium, NZ$32.40 per user per month |
| Database cost | Separate Azure bill: Basic from about USD 5 a month; general-purpose serverless typically tens to low hundreds of USD a month | Included: 10 GB database with the first licence, 250 MB per licence, NZ$64.70 per GB beyond |
| Who administers it | You or your IT partner: schema, logins, firewall rules, backup policy, performance tier | Microsoft. You manage tables and roles in the maker portal |
| Logic | Stored procedures, triggers, views: keep working from a migration | Business rules, formula columns, flows, plug-ins for developers |
| Security | Database roles, row-level security if you write it | Row and column security roles, business units, auditing, by configuration |
| Reporting | Anything that speaks SQL; Power BI direct | Power BI, Dataverse views, Excel; no ad-hoc T-SQL |
| Canvas app gotcha | Delegation stops at 500 or 2,000 rows for formulas SQL cannot delegate; same as Dataverse | Same |
Licence prices from Microsoft's NZ list, September 2026; Azure SQL from the Azure pricing calculator, which bills in USD or NZD by region. Australia East is the nearest region and most NZ firms use it. Dataverse's platform layer keeps paying off past this table too: an AI agent can read a written business skill stored next to the tables, which a bare Azure SQL database has no equivalent for.
Which factor decides the right database?
Logic. A SQL Server system that is mostly tables and a front end migrates to Dataverse and gets the security, auditing and app layer as part of the deal. A SQL Server system whose value is in forty stored procedures that price jobs, allocate stock and produce the month-end report should go to Azure SQL, keep every one of them, and get a Power App on top through the premium connector. Rebuilding forty procedures as flows is the expensive way to end up with the same system. The inventory step in the migration checklist exists to count them before anyone quotes.
Which option suits each of the three example systems?
- Job register, quotes, clients, twelve tables, three procedures: Dataverse. The procedures become business rules in an afternoon and nobody administers anything again.
- Pricing engine with fifty procedures, an accountant who writes views, an ODBC link from Excel: Azure SQL. Keep the logic, lift the database, build the app on it. The accountant keeps their views.
- A vendor product that requires SQL Server: Azure SQL or a supported on-premises SQL Server, never Dataverse; migrating breaks the product.
What costs does the comparison usually leave out?
With Azure SQL, someone still owns it. Backups exist by default, but restore is a skill; firewall rules and logins are yours; the performance tier is a decision that costs money in one direction and speed in the other. For a firm with an IT partner on retainer that is fine. For a firm whose IT is the owner's nephew, Dataverse's absence of a database to own is the feature, and it is worth more than the price difference. The other forgotten cost is the developer: Azure SQL work is done by people who write SQL, Dataverse work by people who build Power Apps, and in a small market those are often different people at different rates. Whichever you choose, the app layer and its licence are the same, and moving a SQL Server database into Microsoft 365 covers the decision that comes before this one: whether to leave SQL Server at all. The keep-SQL-behind-a-gateway option on that page and Azure SQL here are cousins: same licence, same logic kept, one on your server and one in Azure. One difference worth knowing before committing: Dataverse throttles API traffic per user and per licence, a ceiling Azure SQL doesn't impose the same way, covered in the Dataverse API request limits post.
Common questions
What is the difference between Dataverse and Azure SQL?
Azure SQL Database is Microsoft's cloud SQL Server: a real relational database you write T-SQL against, with stored procedures, views and any reporting tool. Dataverse is a managed data platform built for Power Apps: tables with security roles, auditing, business rules and forms included, no T-SQL, no server to manage. Both live in Azure; only Dataverse comes with the app-building layer attached.
Does a Power App on Azure SQL need a premium licence?
Yes. The SQL Server connector is premium, so every user of a Power App on Azure SQL needs Power Apps Premium at NZ$32.40 per user per month on Microsoft's NZ list, the same licence Dataverse needs. The licence line is identical; the difference is the database bill and who runs it.
How much does Azure SQL Database cost a small business?
The Basic tier starts around USD 5 a month for a small database, and a general-purpose serverless database that pauses when idle typically lands in the tens to low hundreds of US dollars a month depending on size and hours. Azure bills in USD or NZD by region through the pricing calculator; Australia East is the nearest region to New Zealand. That is on top of the Power Apps licences, where Dataverse's storage is included up to 10 GB.
When should a small firm choose Azure SQL over Dataverse?
When the value of the system is in its T-SQL: dozens of stored procedures, views the accountant runs, an integration that speaks SQL, or a developer who will keep building in SQL. Then keep the logic and put the app on it. When the value is in the tables and the forms, Dataverse gives you security, auditing and the app layer for the same per-user licence and no database to administer.
Where to start
Count the stored procedures and tell me who wrote them. One free 30-minute call and you get Dataverse or Azure SQL with the monthly numbers. If it is Azure SQL and the work is all T-SQL, I will say whether it is mine to do or a SQL developer's.