However, the lack of open hardware for the Rpi means that it is not possible to replicate one at all. You basically can't make a Rpi-compatible board to suit your own needs.
It isn't the fault of the Rpi fundation, it's the sorry lack of openness in electronics hardware that is to blame. Chip manufacturers license lots of IP from each other and I doubt we'll ever see a truly open microprocessor/controller that would have all its features accessible without restriction.
We can hope that one day a RISC V implementation could be open, but then the GPU, controllers, etc would also need to be unencumbered.
Of course, it would be possible for chip manufacturers to allow this licensing -maybe to the detriment of the cost of the chip- but it would need to really be profitable for them to go through the 'hassle'.
After all, the die doesn't need to be open, we only need open access to all the functionalities of the chip.
I read this complaint about the RPi firmware so often on HN but I have never read how it actually is being used to stop RPi owners from doing what they want with their RPi. It is theoretically a problem but in practice what has actually happened. There has been an enormous amount of tinkering with RPis, people are using them for all sorts of things, and are able to build on the work of others. DRM is important but viewing copyrighted works is not the only use for a computer.
There is no perfect computer but the RPi for its OS-versatilty is arguably better than many others who fail miserably on that point.
There are certainly better dev boards from a spec standpoint, but if it cannot boot off external media and is only intended to run a single OS, I think it is missing some "freedoms" that the RPi provides.