It's inherently limited by its geometry kernel. Most "real" CAD suites use something like parasolid, usually with a bunch of extras slapped on top. Making a new one from scratch is a massive undertaking, but I'll remain forever hopeful that we get a new, modern, open-source kernel one of these days...
This isn't really true. The vast majority of problems are in the UI. The geometry kernel is limited, but it's good enough for an open source project. Compared to say OpenSCAD, Open CASCADE is leagues ahead.
I don't necessarily agree in this case - OCCT is more than capable for what FreeCAD is offering. Add to that the development trajectory of OCCT also seems to be really taking off recently (with the 8.0-RC, they've re-worked how all B-Spline algorithms work, with implications for all operations).
Not gonna lie I just hope the rewrite it in rust community takes a stab at it at one point,
There are already at least two geometry kernels being written in scratch in Rust (see fornjot.app for one) --- the problem is the first parts are obvious/easy, so initial progress is rapid, then one hits the difficult/intractable parts and progress stalls, usually to be abandoned.
There are a couple of doctorates available for folks who are willing to research and publish in this space --- the commercial products are all holding their solutions as trade secrets in their code --- even then though, the edge cases are increasingly difficulty to solve in such a way as to not break what is already working, hence the commercial kernels having _very_ large teams working on them, or at least that is my understanding from what Michael Gibson (former lead developer of Rhino 3D, current developer of Moment of Inspiration 3D) has written on the topic.
What's wrong with OCCT?