The X16 should take hints from the X8. Honestly, feature creep plagues the X16. The Yamaha chip, the Nintendo ports: it's for a niche audience, and that's what makes the X8 appealing: the X8 looks like feature creep never happened, it looks simple, it's fresh air. I can't see people willing to learn, code and assemble (the original stated goals of the X16) needing an FM chip or Nintendo ports. With all the sympathy I have for the project, the day the FM chip was discussed was the day I saw the ugly face of feature creep entering the project.
If it were my project, I'd build an enhanced X8 with 256KB to 1MB of memory and USB or DB9 ports as an FPGA and sell it as soon as possible (or sell the cores for the MiST and MiSTer, which would reduce the manufacturing and shipping costs to 0 and allow real and fast profit). During the next months after the release, I'd get the system rock stable (the beauty of FPGA is that not only the firmware but also the specs can evolve if some roadblock occurs) and I'd get a good software library for it thanks to the community. As a bonus, it would make the current chip shortage irrelevant.
Then, when the rock stable FPGA is done (18 to 24 months later), I'd build the X8 as real hardware (in that scenario, phase 1 occurs at the end of the development). I'd name the "chips version" Commander X16 and the only difference would be that the X8 is FPGA and X16 is a kit or assembled computer made of as many real parts as possible (the idea of having certified assemblers around the world for people wanting it done for them looks perfect – and it would avoid customs nightmares, especially in the EU).