Over the years I've seen so many solutions to development but not one is a panacea. Example: I had a 28 day embedded project that would have blown its 28 day limit if there were any attempt to force DevOps onto the plan. This was 28 days from start to finish including a custom printed circuit board, circuit design, software coding and final delivery. We made thousands of these boards as well as outsourcing the remote control for the system.
The speed of development was a bit hard on management but we got it done and moved onto the next project. Even the DevOps people in the company were caught flatfooted by the short cycle from start to finish. There was no need to update the software over time so it didn't fall into their DevOps way of how things work.
The second problem I've run into DevOps is lack of financial commitment. All the DevOps I've seen was costly and management I've had is very focused on costs. So the dance continues to be to automate what you can and watch that budget.