It’s due to the maths behind. Special algebra is used for video compression, and a discovery has been made something like 15 years ago that allows a better video compression. It fueled technical progresses of the last years.
For images, we basically hit the wall quite some time ago. The new technologies are more about engineering improvement than math improvement.
Then there is the technical environment. It doesn’t matter if your technology is a bit better than the old one because the cost to change the whole technical environment is insane. That’s why ipv4 is still there for example. Changing everything for a new technology to be used is a long, costly and painful progress. But this is something only developpers can’t cope with, because the development culture is painfully ignorant of industry constraints and time lines.