What if PC ..... never existed?

Early MS-DOS / PC-DOS was only little more than CP/M (unofficially) ported to x86. Very, very similar under the hood.

Quite possible. At the time, there was much more money involved in the office "business machine" market.

Maybe MSX had survived longer. And Apple would be even bigger than it is!

If you ever have the time to do a little archaeology by digging into the magazines and book archived in the various archival sites you will find out that the MSX standard was never meant to be a dominant standard. It pains me to say that there was never the intention, much less a chance, for the MSX standard to be something more than what it achieved.

The reasons stated by the very inventors of the MSX standard were various and I will try to enumerate them

1) The MSX was underpowered in comparison to the other computers of the time. When it appeared the Japanese market was already served on the high end by the likes of NEC PC88-98 and other offerings by Sharp and Fujitsu. Everyone was also getting ready for the 16 bit era to arrive, as it did 2 to 3 years after the launch of the MSX.

2) The MSX was not a gaming powerhouse as it lacked obvious gaming features like scrolling that the NES offered at about the same time. The SEGA SG1000 with almost exact MSX specifications was no match to the NES either, cementing the fact that the TMS 9918 chip (a 1978 product) was well past its expiration date.

3) The MSX standard was envisioned with many applications, few of which could carry it into mainstream status:
3a) Cheap computing for the low end that leverages compatibility and interoperability - This goal was achieved. MSX was popular as a home computer for the low end, successful in the education industry and the software library maximized the potential that standardization offered.
3b) The MSX wanted to be the Arduino of its time, offering a cheap computing platform with easy programming and an open expansion bus standard. In fact, upon release of the MSX, home automation was one of the first application demoed in various tech shows. As you could guess, it was 35 years too early for the IOT industry to take off.
3c) The MSX also wanted to be the Java like standard that would be baked into every appliance imaginable, similar to the current SmartTV abilities and other Smart features. There were a few appliances that ware based on the MSX for their computing abilities but, again, there was never a good case to use the MSX as the brains for many devices and manufacturers turn to cheaper and better solutions that were soon available.

None of these original goals of the MSX would carry it to the top of the home computer food chain and targeting to the low end of the market doesn't allow you to do true innovation.
By the time the MSX turbo R came, there was the Sharp X68000 in Japan, the Amiga, Atari ST, Archimides and Apple IIGS as well as the PC and the MAC with the Megadrive and Super NES soon to follow in the gaming category.

No matter how far we would like our beloved MSXes to go and how much we upgrade them the truth remains that MSX designers knew the could use a 16 bit CPU and the MSX2 VDP for the 1st iteration of the Standard and didn't. They couldn't because they would miss the price point.

This does not mean that they failed. On the contrary, they achieved far more than they set out to do and this is why we are here. They could go for the High end but that would be replicating the already existing PC clone industry.

So there, these are my 2 cents and I would love to read your thoughts!

Let's see where the Z80 and the VDP ended up:

The Z80 went nowhere. The R800 was painfully unoptimized and by that time any other modern CPU architecture had better emulate the Z80 rather than keep compatibility with it in silicon. So that lineage is dead even though Z80 still exists as a microcontroller chip.

The Texas Instruments TI9918 evolution is an interesting story with 3 main branches:

1) The MSX branch sees the 9918 superseded by the MSX2 VDP that adds screen modes 4-8, increases palette to 512 and resolution to 512X424 (interpolated)
Then comes the MSX2+ VDP that adds proper scrolling and the compressed 19k color modes (9-12)
The MSX TurboR adds nothing as it misses the V9990 which eventually surfaces and ends the branch. It ads parallax and multicolor sprites.

2) The Sega branch uses the 9918 for the SG1000, and evolves it with the introduction of the MK2/Master System adding 64 color palette, 1 multicolor mode ala Screen 5 (most games) and a 2 layer tile mode (Golden Axe, Space Harrier) that uses tiles instead of sprites and only has choppy tile scrolling. There is also 1 game (F-16) that uses the Screen 2 mode and is a direct port of the MSX1 game.
Then comes the Megadrive VDP that adds 320X224 modes (and 320 X 448 interpolated, only ever used in Sonic 2). It drops the original 9918 modes 0-3 but retains the Master system modes. This is really the end of the line as subsequent addons and consoles by Sega used other architectures.

3) Texas instruments continued evolving the original 9918 separate of Sega and MSX and much improved models where used mostly in various Arcade games like Mortal Kombat and NBA Jam that used the TI 34010 which is a combination of CPU and VDP. The 34010 was also used in Graphic Accelerator cards in the Windows 3.0 era. It was also present in addon cards for Amiga and the final model was the 34020

Have fun learning about the TMS 34010!
https://www.youtube.com/watch?v=h5w3a5K6hbU

I do not agree. MSXs were relatively expensive compared to other equivalent machines (but the material was of good quality). The MSX suffered from a lot of things like for example the TVs which were not enough evolved for the computers, etc, and MSX prices have really dropped only towards the end. ASCII felt that the monitors were not necessary despite the increasing demand. These are the two main points that have made the success of Amstrad (low price and monitor included).

MSX's VDPs can not support the resolutions beyond those of TVs. Only the V9990 started to do it.