How much money has OS4 made so far?
I doubt any numbers have been published on that.
How do MorphOS, AROS and OS4 compare in terms of development practice and development speed?
I don't know either. All four do not run on classic hardware (or, rather poorly, as AROS is concerned). I personally consider it quite obvious that they haven't found their customers, lacking both affordable hardware and a sufficiently large software library. Development pace is at glacial speed for all of them.
I doubt you can really compare AOS/Morphos with AROS, and all of the three with the classical AmigaOs. AOS/Morphos address approximately the same market (PPC Amiga "wannabees"), or the lack of a market, so to say. Morphos made at least not the error of depending on custom-made overpriced hardware, which is an advantage. But none of them supply a sufficient library of interesting applications, so both are a rather pointless exercises, if you ask me.
AROS I do not quite understand either. Development is slow, it is neither an alternative to AmigaOs, does not run well on the existing systems, so there is no usable software library either. Otherwise, if I already have a PC, then there is Linux and many applications for them, so why would one pick AROS?
It's really not a miracle that none of them has found a wide user basis.
The reason why I believe that one can probably/hopefully come up with a self-sustainable AmigaOs for 68K is that there is now both: a) an interesting hardware alternative (the FPGA implementations) and b) a large legacy software library.
If you come up with an alternative business model that helps to make this possible, go ahead. Up to now, I received nothing, and until then, I believe I'd rather stick with the original idea: Sell it for (gosh!) money.
