Can't be, for 2 reasons:

1. Heat. A Vega, which comes close to the 1080 in power, consumes more power than an entire console. And downclocked to it's sweet spot the performance gain over the X is pretty weak.

2. Price. High-end GPUs cost way over 500$ (the Vega VII that got released today costs 700$ for instance), so you can't expect these to be built into a console unless production costs would at least match, if not exceed PS3 initial costs

Much more than you think, I fear.

Have a look at the NPD prediction tool, where I try to find exactly that out - and over half of the PS4/XBO consoles sold nowadays in the US are Pro or X. In fact, the X has a whooping 80% ratio over the S right now (which is probably also why the S is so often on sale - it just doesn't sell anymore otherwise), while Pro and slim are more or less at equal terms.

If we expand that to global sales, and reduce the ratio a bit in favor of the slim/S models, then out of 280k weekly sales between PS4 and XBO probably still mean 120k Pro/X models sold per week outside of holidays by now. And while the sales of PS4 and XBO are slowly dropping, it's almost exclusively the base models that are dropping right now, giving the Pro and X an increasingly large market share.

In fact, since the Pro is already 2 full years on the market and the PS4 as a whole sold over 38M WW, I'm very sure at least one quarter of those sales are Pro sales, which by itself would already almost match your figures. And with 11.5M in the US alone, it's quite possible the Pro is closing in on 5M sold there by now. As for the X, look the first paragraph.

In other words, those 10-13 million you mention can very well be just globally in the US right now instead of worldwide. I'm pretty sure that Pro+X, taken together, already sold 15M+ and are by now close to half of all their consoles' share in weekly sales. Which means that until the next gen's release, there will be at least some additional 15M Pro and X consoles being sold, bringing their share to at least 30M consoles if the ratio stays the same, but 35-40M seem pretty plausible with their rising market share for then by now. And 30-40M who potentially won't upgrade because there's not too much of an upgrade can be pretty hefty, especially early on.

Also, don't forget that those enthusiasts who will upgrade mostly also sell their old console. A new new console for 499$ or an used Pro/X with almost the same visuals for 249$, what would you choose?

Can't be, for 2 reasons:

1. Heat. A Vega, which comes close to the 1080 in power, consumes more power than an entire console. And downclocked to it's sweet spot the performance gain over the X is pretty weak.

2. Price. High-end GPUs cost way over 500$ (the Vega VII that got released today costs 700$ for instance), so you can't expect these to be built into a console unless production costs would at least match, if not exceed PS3 initial costs

Much more than you think, I fear.

Have a look at the NPD prediction tool, where I try to find exactly that out - and over half of the PS4/XBO consoles sold nowadays in the US are Pro or X. In fact, the X has a whooping 80% ratio over the S right now (which is probably also why the S is so often on sale - it just doesn't sell anymore otherwise), while Pro and slim are more or less at equal terms.

If we expand that to global sales, and reduce the ratio a bit in favor of the slim/S models, then out of 280k weekly sales between PS4 and XBO probably still mean 120k Pro/X models sold per week outside of holidays by now. And while the sales of PS4 and XBO are slowly dropping, it's almost exclusively the base models that are dropping right now, giving the Pro and X an increasingly large market share.

In fact, since the Pro is already 2 full years on the market and the PS4 as a whole sold over 38M WW, I'm very sure at least one quarter of those sales are Pro sales, which by itself would already almost match your figures. And with 11.5M in the US alone, it's quite possible the Pro is closing in on 5M sold there by now. As for the X, look the first paragraph.

In other words, those 10-13 million you mention can very well be just globally in the US right now instead of worldwide. I'm pretty sure that Pro+X, taken together, already sold 15M+ and are by now close to half of all their consoles' share in weekly sales. Which means that until the next gen's release, there will be at least some additional 15M Pro and X consoles being sold, bringing their share to at least 30M consoles if the ratio stays the same, but 35-40M seem pretty plausible with their rising market share for then by now. And 30-40M who potentially won't upgrade because there's not too much of an upgrade can be pretty hefty, especially early on.

Also, don't forget that those enthusiasts who will upgrade mostly also sell their old console. A new new console for 499$ or an used Pro/X with almost the same visuals for 249$, what would you choose?

Sony is not spending what normal consumers spend for the GPU, and all other components(https://www.gamespot.com/articles/we-built-a-pc-using-ps4-pro-specs-how-does-it-perf/1100-6443491/), so the price point you made is a bit off;  I agree with you that we can't have the state of the art inside a console for all the reasons we know, but by the end of 2020, at a price of $449 you might have an interesting piece of hardware, sure much much more powerful than PS4 PRO.  We have just to wait and see.

It's not by the end of 2020. Sony isn't going to choose the GPU at the end of 2020. They've already got the GPU at this point if they are releasing at the end of 2020.

Of course.   Hardware must be finalized and tested, no way to choose it in Nov 2020.  I was talking about hardware ready to be shipped by end of 2020, and how much it would cost to Sony to produce it.