Just because something is planned doesn't mean it will happen exactly that way. As SvennoJ points out above.....

The tech was not created yet, so when a die shrink failed, the contract was not applicable to the situation.

I feel it's because current ps4 can't do VR very well.

Either way a ps4 neo and a ps4 slim are both good ideas.

What do you mean they don't manufacture?

I always thought Nvidia and AMD designed and manufactured their GPU's.

Neither Nvidia or AMD own their own fabs. They use manufacturers like TSMC, Globalfoundries or Samsung to produce the chips. Intel is the only one with its own fabs.

What does that have to do with anything? The 20nm die shrink simply doesn't work, no matter who makes it.

http://www.pcper.com/reviews/Editorial/28-nm-GPUs-Extended-Through-2015-20-nm-Planar-Bust

The essential difficulty of the 20 nm planar node appears to be a lack of power scaling to match the increased transistor density.  TSMC and others have successfully packed in more transistors into every square mm as compared to 28 nm, but the electrical characteristics did not scale proportionally well.  Yes, there are improvements there per transistor, but when designers pack in all those transistors into a large design, TDP and voltage issues start to arise.  As TDP increases, it takes more power to drive the processor, which then leads to more heat.  The GPU guys probably looked at this and figured out that while they can achieve a higher transistor density and a wider design, they will have to downclock the entire GPU to hit reasonable TDP levels.  When adding these concerns to yields and bins for the new process, the advantages of going to 20 nm would be slim to none at the end of the day.

The planned die shrink turned out to be a bust. Switching to finfet is not trivial and requires new designs for the entire tool chain. Sony might have contracts ensuring the price won't go up for production of the current 28nm planar apu, yet they simply can't make the originally planned slim version. The current ps4 design is a dead end.

I don't know if Sony could have made a 14nm slim version with exactly the same specs, however it would probably still cost roughly the same to make as the planned Neo and a lot harder to sell back at the starting price. Might as well test the waters with a ps4.5, easier to sell, and can be produced for less in the future.

Any citation for this?

This doesn't make sense. There is always a choice. And there is no business sense to the premise. After all, if Sony makes the rumored console, at the rumored price, they would always be making less money then cutting the current model to 299. It is hard to imagine a scenario where a 399 neo is more profitable than a 299 PS4.

The only thing I can think of is if either Neo is at least 450 or if they can gett the cost of production down substantially (using magic).

In fact now that I think about it, the premise of yhis article makes me doubt that the Neo is even real.

Also the idea that AMD would abandon production of their fastest selling console semiconductor is asinine. 

If AMD doesn't wan't to make the chip anymore and do not have a contract holding them to it, they can certainly do that. The chip may have sold very well but it is outdated and unsuitable for a fast moving market. 

The way the cost reduction works is due to economies of  scale, this is the economic theory that the larger your production capacity for one item, the cheaper it becomes to produce per item. If they kept just one production line open for the old chips, they would need different equipment, different equipment and different techniques as well as people trained for that process.... that adds up to huge costs. It is better to have a streamlined process, all the same. 

So will original ps4 be phased out?

Smells like bullshit.