SQL ConstantCare® Population Report: Winter 2025
Every quarter, we publish adoption rate data showing how quickly people are adopting new versions of SQL Server. Today it’s time for the winter 2025 version of our SQL ConstantCare® population report.
SQL Server 2022 is finally seeing some good adoption with about 21% of the market. It’s gradually taking market share from all other versions, including 2019, but 2019 still dominates with 44% of the market:
- SQL Server 2022: 21%, up from 13% last quarter
- SQL Server 2019: 44%
- SQL Server 2017: 12%
- SQL Server 2016: 15%
- SQL Server 2014: 4%
- SQL Server 2012 & prior: 1%
- Azure SQL DB and Managed Instances: 2%
Here’s how adoption is trending over time, with the most recent data at the right:
About a decade ago, we saw rapid-fire releases with SQL Server 2014, 2016, and 2017 all coming out in quick succession. Their expiration dates are fast approaching too:
- SQL Server 2014 was unsupported as of July of last year
- SQL Server 2016 will be unsupported in July of next year (2026)
- SQL Server 2017 will be unsupported in October 2027
Those rapidly approaching dates means that the next time I publish another one of these 3 months from now, we’re likely to see a really, really big jump in the SQL Server 2022 adoption rate. While companies tend to stay away from bleeding edge stuff, 2022 isn’t seen as bleeding edge anymore, especially given that it’s had 17 Cumulative Updates so far. I bet the 2022 adoption rate will jump, even at the expense of 2019.
12 Comments. Leave new
To be fair, it did take Microsoft 17 cumulative updates to get SQL 2022 *technically* to a point without any issues. (we will ignore the things that are in permanent preview, or the ridiculous things like not working with fabric when columnstore is enabled, the most likely possible source table to get integrated with fabric) Given that it seems to be getting parabolically worse starting with SQL 2019, I assume we can be looking forward to SQL 2025 reaching a measure of stability at some point in 2034
Hey Brent, I love that you publish this data for the community. One thing that crossed my mind – I’m not sure counts of Azure SQL DB vs all the others are at the same level of granularity, as the others can house multiple databases per instance. Would it be more approriate to compare counts of Azure SQL DB to counts of databases in each version of SQL Server? And if you can do that, might you also be able to do counts of databases by compatibilty level? Really just throwing these out as some ideas for interesting add-ons to the current chart. Again, really appreciate that you’re even doing what you’re already doing.
Great question. I think you’re asking, “Can you provide an inventory of databases rather than an inventory of servers?” The short story is no, I don’t think that paints a good picture of adoption either because some of the SQL Servers involved here have thousands of databases per server. The inventory would be really, really skewed.
My ask was really for “and also” instead of “rather than.” I can see your point on how different the percentages may be between server count by version vs. database count by version. However, that’s actually what I was interested in – how do those percentages change and what story does that tell if you have 50 Azure SQL DBs, 10 SQL 2022 instances with 10 DBs each, and 2 SQL 2014 instances with 200 DBs each? The server vs DB counts will absolutely be skewed, but does that indicate a deeper story about version adoption? I’m mostly playing Devil’s advocate at this point, so I’ll be happy with whatever you decide. But I think inquiring minds might like to know, if it’s easy enough to add. Thanks again for all your work.
The short story is that Azure’s market share would look dramatically smaller. I haven’t seen an Azure SQL DB server with thousands of databases, but we have a bunch of ’em in SQL ConstantCare, and even just one of them would dwarf Azure SQL DB.
Hi Brent
The org I work for is currently on SQL Managed Instance, with plans to move towards Azure SQL. Im curious if you have thoughts as to why the adoption rate of cloud native solutions is so low vs the the on-prem offerings from MS?
Thanks
Yes, and I talk about that in my Fundamentals of SQL Server in the Cloud class.
Isn’t SQL Server 2014 unsupported as of July of LAST year?
Absolutely, good catch! Fixed in the post. Thanks!
A great snapshot of where we are with versions! Any indication of percentage running on Linux? My feeling is that number is very low… single digits.
It’s typically only a handful of servers each time I run it.
[…] a survey of which versions are still running. This is from an IT asset firm and matches Brent Ozar’s Population report. 44% of you are running SQL Server 2019, which is the largest percentage. Since there’s an […]