You have been designing an SoC/MCU for a while, you would know well that development of a CPU from scratch is a daunting task requires long term commitment, huge investment and lots and lots of trial and error, I don't have the temerity to convince propaganda bot otherwise than their official stance.
It is tough to develop an OS from ground up, the only true OS I know of is Unix, everything else is either a cheap RIP off (microsh*t) or descendant linux, BSD, etc, etc. Same challenges as above and huge investment, debugging with a known instruction set.
But when we add a new instruction set and a new OS, well we are looking at at-least a decade and change timeline while the other challenges remain.
I do remember about two years ago Chinese domestically developed CPU was at par with a pentium II and if I am not wrong it used the same instruction set, I may be wrong about the instruction set.
About this project its most likely based upon a licensed ARM core, or a purchased older model with local development and most likely a linux offshoot. There is room for a newer CPU, instruction set and OS but the challenges are monumental there were two outstanding OSes I saw OS2 warp and Be-OS both failed OS2 due to typical IBM arrogance and Be-OS was deliberately destroyed by the larger a**holes.
If any single country can do it it would be Chinese they have the resources, all that is required is the long term commitment and huge investment.
I have seen screenshots of BeOS before and it seems clean. And from reading about it from that time the OS supposedly had a reliable file system and the OS was supposedly good for multimedia.
But of all the OSes I know about I believe QNX Neutrino to be the best. It is microkernel-based with message passing as IPC and has the Adaptive Partitioning feature to ensure that a certain set of programs ( like networking - TCP/IP and Ethernet drivers ) do not eat away the execution cycles at the cost of others. This comes handy when say there is a network attack on a machine but the Partitioning does not hang the machine because the non-network programs have been alloted some cycles too.
About Microsoft, I think XP was the best in terms of UI and general usability. I say this of course with the background that I haven't used MS for many years and have been using pen-drive-bootable Linux ( like Slax now ) to use the machine. So I don't know about Windows 7 and so on.
About developing an OS from scratch, inspired by QNX some years ago I had written a basic OS. It didn't have memory protection and timeslicing but had message passing, UNIX-like signals ( async notifications / program control ), timers and a few keyboard commands. I and a financing partner started a company to commercially develop this OS but the company closed down for a non-technological reason. Nevertheless, developing another OS is not an impossible task for me for the new SoC / MPU.
But you are right. Designing a processor is a long-term commitment. It has taken me years to think of a simplified, RISC instruction set, and then there was the question of what hardware interfaces to have ( the SoC part ) which I have narrowed down to USB ( to implement 3.x or 4.0 ), display, speaker ( which can double up as DAC ) and mic ( which can double up as ADC ). And then there is the GPU part which I am still studying and which will take months.
Some days ago I had mailed two American companies to inquire about certain battery tech and memory tech. I received encouraging news from the battery company. Yet to hear from the memory company.
I want to start a company whose product will not only be the SoC but also a portable ( or wearable ) computer in which the SoC is used.
And yes, you are right about SoC / MPU implementation taking time and trial and error. The basic element I will need is a FPGA board and to obtain that I will have to approach a venture capital company.
About the Chinese, they have the resources but their processor implementations have been SPARC or MIPS or ARM or now RISC-V.
In India there is an ongoing competition called "Swadeshi Microprocessor Challenge" organized by the Ministry of IT ( MeitY ) which called for individuals or groups to use the Shakti processor ( RISC-V ) or the Vega processor ( some other foreign design I suppose ) and implement them on Xilinx FPGAs provided by the ministry for various applications ( either from the provided list or some other ). I had inquired with the organizers and they said that since my is a separate design I will have to use my own resources.