Maybe upskilling is a worry for some, but frankly a developer who doesn't understand how to leverage SQL Server to their advantage will not even see the need for DevOps. And if they don't see the need, they simply will not do it, no matter what incentives or approach is taken.
Frankly, I don't understand that. Databases are exceptionally good at doing data-related things. Duh. Although, to be fair I've been immersed in database-essential applications development since the FoxPro days. The fact that developers treat databases as dumb persistent storage is face-palm territory.
Train the developers in relational theory. Show them the performance difference between using indexes correctly and not. Show them the advantages of constraints. Warn them about the dangers of triggers, about the advantages of stored procedures. The security model using SPs is simply light-years easier than direct table access. Sure, there's a ton of SPs needed, but so what?
Show them how SQL Server will reduce their pain. That will do more for DevOps adoption than anything else. Of course they'll have to learn T/SQL so they can write SPs but (shrug) once they stop laughing at how primitive T/SQL is and how many warts it has the other advantages of SQL Server should placate them.
I will say one advantage of T/SQL's primitiveness is to encourage KISS on a massive scale. Annoyingly massive, but massive non-the-less.
You want to turbo-charge DevOps? Show developers what's in it for them!