Is SQL Server really “too legacy to love”?

Listen on:

In a recent article, The Register suggested that SQL Server is “too lucrative to ditch, but too legacy to love” – but is that actually true?  

With many years’ SQL Server, PostgreSQL and Oracle experience between them, Steve Jones, Kellyn Gorman, Grant Fritchey, and Pat Wright offer their thoughts on the article, and the current state of SQL Server as a whole – including how it fares against other databases in 2026. 

Read the full The Register article here.  

DB-Engines: https://db-engines.com/en/ 

Listen or watch to every episode of the Redgate Simple Talk podcast at 
https://www.red-gate.com/simple-talk/podcasts/ 

Did you enjoy this episode and want to support us? Leave us a review or rating on Apple Podcasts or Spotify!

To review us on Apple Podcasts, simply head to our main show page (The Simple Talk Podcast) and scroll down until you see ‘Ratings & Reviews.’ From here, you can simply tap the 5 stars to rate us, or click ‘Write a Review’ to leave more detailed feedback!

To rate us on Spotify, simply head to our main show page, hit the ‘About’ tab (to the right of ‘Episodes’), and underneath the description you’ll see our current rating in the left-hand corner. Tap that to leave a rating of your own! 

Is SQL Server really “too legacy to love”? The key takeaways

If you don’t have time to watch or listen to the full episode, here’s what you need to know.

“Legacy” doesn’t mean irrelevant

The Advocates – Steve, Kellyn, Grant, and Pat – pushed back hard on the framing that SQL Server and Oracle are simply “legacy” platforms. Many features developers wish existed in newer platforms already exist in these older ones, after all. The “legacy” label really undersells how much capability is still baked into the platform.

SQL Server is still growing — just not as fast as PostgreSQL

Despite the narrative of decline, SQL Server usage is reportedly still growing at around 8%. The real story isn’t collapse, it’s that PostgreSQL is growing faster – and is on track to overtake SQL Server in the DB-Engines rankings by the end of the year.

History has seen this hype cycle before

There are parallels to be drawn to previous eras. The MySQL/LAMP stack boom in around 2000, for example, and predictions 15 years ago that Oracle was “finished”. Neither of them played out as expected. Hype around a “new” platform rarely translates into the wholesale abandonment of established ones.

Different databases still serve different jobs

The idea that just one database platform can replace everything else is a misunderstanding common outside the technical community. Transactional systems, data warehouses, and temporary/throwaway databases all have different requirements, and picking the “right tool for the job” still matters.

PostgreSQL has real enterprise gaps — for now

PostgreSQL still has shortcomings at true enterprise scale, especially in banking and insurance context – failover behavior being just one example where Oracle’s rack clustering is more mature. That said, expect the gap to close within a year or two as PostgreSQL continues to improve rapidly.

Oracle is stickier than SQL Server

A notable prediction from the Advocates: Oracle’s install base will likely hold on for another 5–10 years, largely because of deeply embedded PL/SQL packages and wrapped code that’s difficult (even for AI) to unpack and rewrite. SQL Server, by contrast, is seen as comparatively easier to migrate away from, toward PostgreSQL.

Migrations are far harder than they look on paper

Moving the data is described as “the easiest part.” The much bigger challenges are migrating application code, and — critically — rebuilding disaster recovery, backup, and restore processes to match prior RPO/RTO commitments. Teams often don’t realize how different recovery procedures are on a new platform until it’s too late.

Cloud “migrations” often become permanent lift-and-shifts

Full cloud migrations rarely happen as cleanly as planned. The panel shared examples of migrations that were expected to take a month stretching to 12–18 months, with many organizations settling into a long-term (sometimes permanent) lift-and-shift rather than a full platform refactor. This is explored more in The Cloud Migration Divide.

Business incentives drive platform decisions – not just technical merit

Sales and marketing dynamics play a big role. It’s easier to sell an existing customer base a “new and improved” product than to win new customers. This even creates competition internally within Microsoft, where the teams behind Fabric, Databricks, SQL Server, and PostgreSQL effectively “compete” against each other for investment and attention.

Microsoft’s recent PostgreSQL hiring signals a real shift

Microsoft has been comparatively slow to invest in PostgreSQL compared to AWS and Google Cloud — but recent hires (including a well-known PostgreSQL community figure moving into a PM/advocate role) suggest that may be starting to change, even if the long-term commitment from leadership is still uncertain.

The bottom line

SQL Server isn’t necessarily “dying”, but Microsoft’s investment priorities are visibly shifting toward PostgreSQL and Fabric. Whether that translates into actual customer migration depends far less on technical superiority and far more on the practical difficulty of moving mission-critical systems, code, and disaster recovery processes off a platform that’s been running for decades.