Just as a simple example, if my competition is using a non-dead rendering library and their UI is twice as fast as mine as a result, I'm not going to be happy.
Recent commits doesn't indicate much in terms of security, fit for use, or quality. All it indicates is recent commits, and hints at reachability of the authors if you need to ask them about their intent (if code is unclear) or for license alternatives (if the license is unsuitable).
Yeah, or it could be the opposite. The newer, shinier alternative could also be faster, less buggy alternative.
The only alternative is maintained by Google and is much harder to use.
But really, QPainter suffers from a lot of the same issues as Cairo (and then some). The main "problem" with them is that they are CPU bound. Both of them have OpenGL backends, but they are not "really" faster. There is also the fact that the Qt Company doesn't invest in it anymore since QtQuick2 was released. Because of that, it doesn't get some of the newer Skia features. So Cairo/QPainter are pretty good at server side rendering. Qt was never popular for that use case. Cairo used to own that market, but it's usage is declining due to the maintenance inactivity issues. Also, once upon a time Cairo was used in Firefox and Chrome, so it's SVGs tend to be rendered more accurately in modern browsers than Qt ones.
Skia itself is an unsustainable library to depend on. Its API and behaviour isn't stable. Google doesn't really care about non-Google use cases, so the CMake is broken 90%+ of the time. It adds a large maintenance burden on projects trying to use it. Distributions also hate it for being impossible to package reliably. In turn, it makes hard for smaller projects to migrate away from Cairo because users have an easier time getting their hands on Cairo than Skia.
(bias disclaimer: I worked with both as a KDE dev and as the AwesomeWM co-maintainer, a Cairo based project. A lot of the recent unreleased Cairo commits originates from the AwesomeWM community, but not from me)
That was the theory behind Cairo. In practice it never worked out and Cairo wasn't able to fully exploit hardware accelerated rendering. Cairo was very complex so it wasn't easy for the maintainers to keep all the various backends up to date.
If something has bugs or lacks features, then it needs maintenance. "Maintenance" for pushing out versions to fulfill expectations of "maintained" is pointless.
Calling it dead implies that projects that are currently using it shouldn't use it anymore, but projects for which cairo as-is suffices will continue to work with the features that cairo already has, so why would they need to migrate?