X2 is only 25% faster than X1 and hasn't been used in any mobile platform yet, only car so far. Not for general public mean  it will be more expensive. Till 2020 they will use X3 instead. And your 7nm exist in your theory no way is just $250 with 4 -5 hours when play on the go. No battery tech with cheap price like that exist yet 

So you agree the X2 is already faster.

The X2 is built on a smaller fab, so it is cheaper to produce per chip. If Nintendo wanted to use it in a revision the cost would be slightly lower per chip at those rates. There is absolutely no issues with 16nm FinFET.

I don't  doubt Nintendo will skip the X2, but I also am not 100% sure they will use Xavier either. They may end up using X4 in 2019 as it should be more mobile focused than X3 (Xavier).

7nm is absolutely not something that only exsists in my theory. It is 100% on AMDs road map, and signs point to it coming as soon as late next year.

Chip costs are determined by the size and the yield rates. A mobile chip wiith PS4 level power will be smaller than the current PS4 Slim 14nm design. That only leaves yield rates, and that depends on how smoothly and quickly 7nm matures.

You do realize that smartphones will be more powerful than PS4 by 2020 right? Smartphones only cost about $200 - $300 to manufacture, the Smartphone companies are just taking massive advantage of everyone that buys into the hype. Non of the console manufactures have ever marked up their products, in fact it has generally been the exact opposite. Don't be mis-lead, Switch is not cutting edge, and phones don't actually cost $600 - $1,200. 

Cutting edge or not, can you show me a portable system that is stronger than Switch which cost $250 and play HD game like Doom which last at least 3 hours? You say smartphone cost only $200-$300 but how about buttons and analogs, they are free? Also,none smartphone play game at 720-1080p last 3 hours yet, 1,5-2 hours at best, let alone 4-5 hours or even more

I do agree with you overall about how large of performance upgrades we will see in the mobile space in the next few years (die shrinks being largely responsible for that), but I do feel you are missing a few major points and there are some inaccuracies as well.

"Smartphones only cost about $200 - $300 to manufacture, the Smartphone companies are just taking massive advantage of everyone that buys into the hype".

This part is quite true. Although it is doubtful that any of these manufacturers will sell their flagships (ie the ones that have the more powerful processors and GPUs) close to the price of cost. This is because unlike console makers, most smartphone manufacturers do not get high revenues (if any revenues at all) from software sales. Hence their goal is to sell their devices for as high of a price as possible and make it so that you upgrade frequently. So I do not suspect we will see a PS4-speced (or even Switch-speced) smartphone from these manufacturers that cost $200-$300 any time soon (most the devices in those price ranges ship with Snapdragon 2XX, 4XX, or 6XX SOCs which are far below the Switch in terms of power, particularly in the GPU department). Sony though can do something like this themselves with their existing console model.

"A mobile chip wiith PS4 level power will be smaller than the current PS4 Slim 14nm design".

I feel this well happen but it will not be as soon as you think. The current Ryzen 5s do not have GPU performance on par with PS4 Slim and they are targeting ultrabooks and convertible tablets (it consumes too much power to be placed in a mobile tablet console like Switch or much less a smartphone). I do not even believe 7 nm will get us there. The X1 SOC in the Switch (released in 2015) is about 2 times powerful (when docked) and slightly less powerful or about the same (when undocked) as the Xbox 360, which was released in 2005; that is a 10 year gap. Moreover, as the die shrinks are becoming increasingly more difficult, it could take even longer to get a PS4-speced mobile chip that fits in a device with the form factor of the Switch or even smaller. And this brings me to your next statement...

"You do realize that smartphones will be more powerful than PS4 by 2020 right?"

As I mentioned above it is highly unlikely that we will get mobile SOCs that fit into smartphones that are as powerful or even more powerful than PS4 by 2020. The other issue with that statement is that even if an SOC was as powerful as a PS4, there is very little reason to believe such a chip would perform like a PS4 when placed in a smartphone. The big issue comes down partially to thermals and partially to smartphone software. Smartphones do not have fans and the software is designed so that they do activities as quickly as possible (so they run the CPU and/or GPU depending the task as fast as the thermal design allows) in order to finish a certain task as quickly as possible so the processors could return to an idle state and conserve powers. This works great when you are doing quick tasks (and ones where flactuations in performance do not greatly matter) such as web browsing, document editing/viewing, etc. However, it does not work well in games because when the CPU and GPU run that fast for a prolonged period of time at a continuous basis (which is what needs to happen when playing games) they start to throttle because the thermal envelopes in a smartphone do not allows the SOC to run at those speeds for a long period of time. Hence, you may start with a game running at 60 FPS but in less then a minute to a few minutes you will see the frame rate reduced because the SOC is throttling. Now this issue can be addressed in software and in hardware. For example, Nintendo added a fan and two heat pipes to Switch in order to desiccate the heat; PSV and 3DS were underclocked so that their thermals could keep up. I suspect most smartphone manufacturers will go with the fan-less underclocked route (in the unlike event that they actually took gaming seriously, which is another issue I will get into) which would mean that the hypothetical PS4-speced mobile SOC would not perform as fast as it needs to in order to provide PS4 like performance.

The final issue is that consoles and portables like PS4, Switch, 3DS, PSV, etc. provide low level APIs (that are consistent across models and revisions), they have unified standards for physical inputs, and, as mentioned above, the software clocks the hardware so that it can sustain performance over long periods of time, smartphones are so fragmented that none of this can be guaranteed. Hence you may have PS4-speced mobile SOC inside a piece of hardware where the manufacturer does not really want to provide low-level API support (this is an issue I currently have with my Moto Z which has an SOC that supports Vulkan but Motorola refuses to support anything higher than OpenGL ES 3.x). There is also lots of fragmentation with regards to game controllers (and since most developers rightly assume that people on smartphones are playing games using touch, many neglect proper controller support all together). Finally, we have the app stores which are filed with cheap and free to play low-budget, small-scale content which makes it difficult for developers risk putting higher priced but higher-budgeted and larger-scale games on these platforms.

All in all, if anyone is going to do a portable gaming device or a portable smartphone gaming device, it is going to be Sony, Nintendo, MS or anyone else interested in actually making a device for gaming. Apple and Google (and Google's hardware partners) do not care. The hardware partners do not make money out of software so they do not care what type of games their devices play so long as it is generally the same as their other smartphone competitors. Apple and Google make so much revenue from free-to-play and cheap games that they will not really care about investing in order to make their digital platforms more appealing to higher cost games.

For me Sony nearly got it right with the Xperia Play. The big problem with that phone was it had outdated hardware and it was running Android. If Sony put the Vita's OS on there (with greater focus on apps and such, and attempting to make it into a full fledge mobile platform), plus based the phone on Vita's hardware, and made it so that it could play Vita and all of the PSP library then that would have been a very attractive proposition (particularly for 2011/2012).

After selling over 80 mill PSPs, Sony blew it with the Vita. Expensive memory cards, rear touch instead of L2 + R2, the name "Vita" instead of PSP 2, and last but not least, a weak game library. Both Sony's home and handheld consoles have always been sold primarily on the basis of 3rd party support (Gran Turismo is the exception, but ironically that series never released on Vita).

Sony also thought 3rd parties would (help) carry the Vita but 3rd parties were more interested in the mobile platform where they could earn at least as much money with much less effort, investment and risk. Even Nintendo, whose handheld consoles sell gangbusters, have learned that you cannot rely on 3rd parties to sell your console. At best, they'll arrive later when there's a guarantee for profits. Therefore, they made Switch to be a hybrid so they could combine the home and handheld libraries and support the Switch themselves.

With 3DS and Switch, Nintendo provides all of the best-selling games themselves with only Monster Hunter as the exception. Sony would also have to be able to provide the content themselves and the only way I see them capable of that is if they make a hybrid or a handheld that can run all PS4 games in downscaled form.

However, a "PSP 3" would still need an identity and exclusive games, and for this I think Sony would be wise to buy a smaller developer such as Sanzaru Games (Sly Cooper: Thieves in Time, Secret Agent Clank) or Level- 5 (Dark Cloud, Rogue Galaxy, Yo-Kai Watch, Ni no Kuni, Professor Layton, Inzuma Eleven). These games are not AAA games, but they fit the handheld format so well. It would also need an exclusive Gran Turismo, God of War and Daxter, and Sony funded third party games such as a smaller GTA, Fallout, Minecraft, Metal Gear Solid and Final Fantasy.

I've enjoyed my Vita a lot. Especially on long distance travels. Therefore, I might be interested in a "PSP 3" if it supported SD cards, had L2 + R2 and ran the entire PS4 library in downscaled form. Though I doubt they'll release a new handheld. I think they're more likely to introduce a new kind of controller for phones as an alternative to touch and tilt. Soon, phones are capable of running downscaled PS4 games and there's a lot of potential for Sony to grow in the mobile market and increase the number of subscribers of services such as PS Now and PS Vue.

It's right next to your imaginary custom X3. How did you miss it? I thought you said you would be able to tell the difference.

I think the point was that if a smartphone can house a PS4 level APU/SOC, then that same hardware in a tablet like format, or handheld/hybrid, should be able to get near to, if not the same level of performance as the PS4 by then. If Nin can do it, PS can do it. Who would do it better, and who would sell more, is the bigger question.

Double post.

X3 is Xavier and X1 chip in Switch is custom. There is no reason to believe X3 won't be custom. Also, Sony can do like Nin, right? I wonder what Sony game still sell in Japan after being released 3-4 years like Animal Crossing or Mario Kart. What Sony franchise can make impact comparable with Pokemon?

PS can do what Nin can, in terms of hardware, based on my reply solely about hardware. They could also try and create similar games as well, but trying to compete with Nin's specific franchises is a challenge in itself. I don't really see PS targeting that specific audience though, since they would simply be looking for customers who like PS and third party games, but want or require a mobile platform to play them on. Having some mobile based games for it would also be a good idea though. I don't think X3 would happen in the next Switch and even if it did, the battery life would be the same as now at best, if not worse, plus the price would be more expensive. Who's going to buy an upgrade like that? Nin would have to wait until like 2025 and by that time the PS Portable will have taken the entire handheld/hybrid market as well. Pokemon will have no choice but to release on PS devices going forward after that.

Nvidia can give Nintendo something that matches/betters anything Sony can get.

That's the problem for Sony, once you remove the "oooh, better hardware" appeal from them, they don't really have much on Nintendo. Third parties can't ignore the Nintendo platform either because Nintendo has a proven track record of dominance in the portable field.

If I was Nintendo and Sony tried this, I'd just wait 6-8 months and then launch a Switch branded upmarket competitor with even better specs, and I'd keep the current Switch at a lower cost, redesigned with a more kid friendly design.