I believe it will be alot more than 10TF. I have my money on at least 14TF. When sony went from 28nm to 16nm, not only did their CU count double the GPU clock went up by around 15% from the base console. The XB1x (getting rid of their esram on SOC had more room to work with and of course usng a better cooling solution, they practically quadripled the amount of CUs in their GPU and ran it at an even higher clock. 

As cool as that sounds, the XB1x GPU has only 4 CUs more than the PS4pro but is just clocked much higher. So if you use the XB1X as the baseline, simply going from 14nm to 7nm and keeping the clocks the same will at the very least mean you are by default going from 6TF to 12TF and this is not considerring any other architectural improvements made. 

No they don't. If 4nm doesn't come along soonish its not thje end of the world. And there are other ways or things that contribute to price reductions than just using a smaller chip.

That is unnecesary. First thing here is the chosen interface. SATA or M.2? 

SATA will mean the best they could ever get will be around 400MB/s. And thats if they put in a SATA SSD in the console. SATA 3 is raed for 600MB/s but thats not areal world number per say. Anif they are gonn put in a SATA SSD from launch then they might as well just go with M.2.

M.2 wll mean the best they could get is around 2GB/S+ speeds. But it also mean they could use a SATA based M.2 ssd in there allowing those that want to go faster (going from aropund 400MB/s sata to around 1.8GB/s nvme) upgrade their drive.

They honestly don't hve to put in a HDD bigger than 1TB in the console. All they have to do is spport external HDDs from day one. Its just all round better for the platform as far as future proofing goes if they go with an M.2 drive. Honestly and this may sound crazy, I can even see them soldering their storage directly onto the PCB and just support external storage on day one. 

1TB pcie 3 based nand flash storage soldered onto the PCB. Will make it eve cheaer for them to do it.

I guess PS5 will look like the original xbox one but with black color instead of white :). This is obviously fake, like 6-core cpu seems unlikely as we now know zen2 is 8-core ccx and this is what next-gen consoles will use. And DLSS is nvidia tech not AMD.

As what spec ps5 will have my new latest prediction is that Sony will aim for native 4k resolution and not some fake 4k with CB or temporal injection as I believed before. This will require a lot of Teraflops and memory bandwidth.

So the GPU will have 72-80 compute units clocked about 1500 mhz, this will give 15 TF. 72-80 compute units is twice the amount ps4 pro and xbox one X have so it's not unreasonable. The memory will be 12 Gb of ram Gddr6 in 384-bit bus and will give a total of 672 GB/s bandwidth. The console will have 4 GB of ddr4 ram to run the OS on.

2 TB mechanical hard drive with 128 GB super fast flash storage(3000 mb/s read and write speed) The flash drive will make up for the lack of Ram increased as it can be used to store textures, geometry, ai and other stuff someone else probably can explain better. 

I think both ps5 and next xbox will use something close to these specs and both will be very similiar, Sony and Microsoft will launch in 2020 for 399$.

Quantum Break runs poorly on PCs?

You probably refer to benchmarks in Ultra setting or above in native resolution without the "temporal reconstruction" performance cheats.

If you choose similar settings to the Xbox One version or the Xbox One X version, the PC version delivers similar or even better performance.

My first playthrough was the 30fps-locked Xbox version. After playing it through again on PC I can't go back to the 30 fps.

I thought that DLSS is nVidia exclusive tech.

I am talking about the problems with the PC version upon launch. 

And it is as I pointed out earlier in the thread, but no one seems to pay attention to the many arguments against why this leak is fake.

... which had nothing to do with the Xbox One optimization. And don't we forget theproblems of the Xbox One X version:

View on YouTube

Were all these graphical glitches fixed?

Except all the optimization in the world hasn't done wonders for the Xbox One or Playstation 4. - Those consoles still cannot guarantee 1080P, 60fps with anything more than medium quality settings.

The Playstation 4 SoC is doing absolutely nothing that I wouldn't expect out of a Radeon 7870. So... It's not punching above it's weight.

That is a bit of a fallacy that you assume just because one game runs poorly on one platform... That it must be highly optimized for another.

Here is the kicker... The majority of ports to Windows Store tend to be terribly bad.
Quantum Break however runs perfectly fine on my PC... And actually ran like Garbage on my Xbox One X months ago.

No. I don't think you understand.
The Xbox One is using a Monolithic OS that is derived from the Windows 10 Kernel. - It then has high level API's like Direct X 11 and 12 just like the PC.
It also has low-level API's for developers that wish to leverage those.
It's also running multiple Operating Systems in tandem... And gobbling up more Ram than a desktop OS.

It is also leveraging x86 PC hardware... And you somehow came to the conclusion it's a more lean implementation than what the PC offers? Common.

From a hardware perspective... There is not a single graphics feature that is done on the old and outdated console Graphics Core Next hardware that cannot be done on PC. None at all.
And all the talk of Unified memory, large ACE counts and all the other bells and whistles that have been championed from the last 5 years? Where is the advantage it is has provided over the PC? It just  hasn't materialized.

Why not? You didn't provide a different reason. Go on...

The first party games clearly look better than the third party games on PS4. 60fps is a CPU problem, not a GPU problem. Also, I am pretty certain that no one expected PS4 to do TLOU2 graphics back in 2012 when they released the 7870. 

It isn't a conclusion I made through any fallacy. It is an educated suggestion as to how and why it happened. If you were to tell me that Remedy devs are incompetent on the tech side, I would call you a liar. This is the only way to explain why the PC version ran so poorly. The would clearly have fixed it given enough time, as they also did with a lot of the issues afterwards. Most windows store games were planned to be Xbox one only for a long time until Microsoft changed their tune.

Yep, it has a low-level API for developers that wish to leverage that level of control, but the problem comes in when you want to port it to the PC platform because then you can't directly port it without significant performance pitfalls. Simply put, some of the hardware intricacies of the xbox and playstation aren't available on ALL PC hardware (keyword here is "intricacies") so sometimes you can't rely on async compute or some specific number of scheduled wavefronts in your PC games, simply because you never know if it actually has all those ACE units, ROPs or ALU's or whatever.

Elaborate on "gobbling up more Ram than a desktop OS"  Most desktop OS's eat up more ram depending on how much ram it has. For instance, I am running 32gb ddr4 and it eats more than 10 gigs on caching etc. It will free that ram up if i actually start using it in a game, or some other application, but there you have it.

Elaborate on what you mean when you say "more lean". I didn't imply that it is automatically better than what PC offers. What I am arguing is that utilizing different specific hardware in a specialized way is preferred over using random hardware, as random hardware would have to brute-force a ton of random stuff just to get the same result.

As an example, emulators are extremely inefficient. It is a similar principle. The emulation overhead itself can be lightweight but when it comes to executing the code that was built for specific hardware, then it becomes way slow. A laptop I had from 2013 buckled hard in the attempt to render some enemies with a special specular mapping in the later stages of Persona 4 for instance. It dropped to under 10fps and it was a laptop with two GT 650m SLI and an i7. Granted these specs aren't anything special today, but it would still amount to several times the power of a PS2.

I wasn't arguing that PC as a whole cannot achieve these features. I am just saying that developers cannot use these features on PC in any significant way, because they never know if the end user even HAS IT.

The advantages have mostly shown itself in the first party space. Media Molecules game, for instance, has completely done away with the rasterization hardware (ROPs) and they are using a cutting-edge SDF solution for their graphics. If they didn't have those ACE units they would never have been able to go with this SDF solution. The Tomorrow Children is also a game that uses cascaded voxel cone tracing. No one else was doing that back in 2014. Here is a GDC PDF from early 2015  

The take away here is that, if this game was multiplat, it would most definitely mean that they would be forced to use more traditional technology. Unless they only want to target that very narrow segment of the PC market that had GPU's with those exact features.

Really not that hard to imagine that they could since they have a new GPU microachitecture in hand and have 7nm to work with ...