Honestly, I am surprised that they did not included this into the firmware from the start. The only way to truly go SEM system is if you can have independent control over each voice. It seems like an easy concept to conjure since it the voices are under total software control. The only thing missing is variability between HP/LP filter states.
The voices are under software control but the hardware generates the CVs and the hardware is not there for 100% independent CV per voice like the SEM polysynths. The OB-X8 is not designed to emulate the SEM polysynths, it is designed to emulate the OBX/OBXa/OB8. Like the original OBXa/OB8, the hardware supports duo timbrality in that the total voices can be split or layered, but only by two. That means four voices produce one sound while the other four produce another.
Component wise, duo timbrality requires two sets of CMOS switches and S&H buffers to product the CVs to two sets of voices. Hardware costs money.
To accomplish eight voice timbrality like the SEM polysynths, you need four times the hardware components. More hardware does not come free, it costs more money.
This is not a feature that firmware can solve.