Slim Revision Speculation

mpg1

Veteran
Since slim revisions wouldn't depend on the entirely on the same technologies as next gen consoles I think it deserves it's own thread.

To me, slim revisions should be possible as soon 7nm is available...which would be next year.

I think at the very least Sony will release another revision. Personally I'd like to see a revision of the Pro as it's pretty big and noisy now.

Also think 7nm could cut costs of X1X. Probably wouldn't require the same cooling components.
 
Since slim revisions wouldn't depend on the entirely on the same technologies as next gen consoles I think it deserves it's own thread.

To me, slim revisions should be possible as soon 7nm is available...which would be next year.

I think at the very least Sony will release another revision. Personally I'd like to see a revision of the Pro as it's pretty big and noisy now.

Also think 7nm could cut costs of X1X. Probably wouldn't require the same cooling components.

I'm not expecting either Sony or Microsoft to use 7nm for PS4 or X1X.

PS4 because Sony themselves are predicting 2-3 million less shipments next financial year and cost of 7nm chip redesign and Microsoft because of the same redesign cost plus
I believe they will remain stuck with 12 GDDR5 chips limiting motherboard redesign? and the 384bit bus doesn't shrink very well?

Can't rule either out just yet, though.
 
2-3million less per year would still be well over 10m units over the lifespan of the shrink. There's still a 4 yr tail of PS4 sales from today even if PS5 launches in a couple of years.
 
Some are projecting PS5 to hit in 2019, if it has PS4 BC I dont see slim revisions or any PS4 models continuing to sell. Same even if it doesnt hit until 2020.
 
The supposed density improvement of 7nm is ~3.2x following TSMC's website comparisons between 16nm & 10nm (2x), and 10nm & 7nm (1.6x). Due to the required I/O, a jump to 7nm would probably not be feasible.

The layout arrangement of the DDR3 memory channels for Durango is certainly different from that required for GDDR5, but the cost : benefit may not be there for that particular design. I'd start to wonder if it would be more DDR3 chan than transistor area. :p

There is 10nmFF, but that cost : benefit is somewhat suspect.
 
Last edited:
A slim and cheap PS4 seems highly desirable to me. Getting it down to $200 will increase longevity of the platform. Historically there's been a market for <$200 consoles that hasn't been served since PS2 was discontinued in 2013. This is what price managed for PS2 -

ps2-wii-hw-trends.png


It sold ~70 million in its first four years, and ~80 million after that at <$150.

Getting PS4 cheap enough could double its lifetime sales. Giving up on a platform shrink and not having a full micro-slimification that their other PS's had, just seems like giving up on easy money. Sony could even go one further and put in UHD playback and sell it as a cheap accompanyment to 4K TVs, or go the other route and leave out the optical drive and sell a download only console at nearer $150 and test the waters for how desirable such a device would be now that people are used to download only Apple TV and Amazon FireTV and GoogleTV widgets.
 
I also don't think there will be the need to make a slimmer PS4 or Xbone S.
Besides, are they going to launch modern consoles with ~90mm^2 SoCs at 7nm?
For how much longer can Microsoft keep purchasing 16 DDR3 chips for a supposedly low-cost console? Speaking of 16 DDR3 chips, can they even make the PCB that much smaller?

If the 9th-gens are coming later than we think, I could see 7nm bringing slim revisions to the mid-gens during 2019, with the current Slim and S being slowly phased out.

At least for the PS4 Pro, because we're looking at practically the same hardware except for the larger SoC and 1700MHz GDDR5 (which is about the slowest being manufactured today). And IMO it's not the die size difference between 150 and 80mm^2 for the SoC that will prevent the console from being sold for $200.
 
A slim and cheap PS4 seems highly desirable to me. Getting it down to $200 will increase longevity of the platform. Historically there's been a market for <$200 consoles that hasn't been served since PS2 was discontinued in 2013. This is what price managed for PS2 -

ps2-wii-hw-trends.png


It sold ~70 million in its first four years, and ~80 million after that at <$150.

Getting PS4 cheap enough could double its lifetime sales. Giving up on a platform shrink and not having a full micro-slimification that their other PS's had, just seems like giving up on easy money. Sony could even go one further and put in UHD playback and sell it as a cheap accompanyment to 4K TVs, or go the other route and leave out the optical drive and sell a download only console at nearer $150 and test the waters for how desirable such a device would be now that people are used to download only Apple TV and Amazon FireTV and GoogleTV widgets.

I think it is a very different market than the PS2 era where it had the market to itself and multiple node shrinks to take advantage of that were no doubt much cheaper than what 7nm is purported to be. With nicely selling Xbox and Switch, I just can't see a clear ROI.

As AINets asks above what is the BOM and what are the possibilities of reducing it outside of the SoC shrink? Are there 2GB GDDR5 chips to go to 4 chips and simpler/cheaper motherboard for example?
 
Some are projecting PS5 to hit in 2019, if it has PS4 BC I dont see slim revisions or any PS4 models continuing to sell. Same even if it doesnt hit until 2020.

PS3 Super Slim and Xbox 360 Model E were announced/released 4-5 months before launch of Xbox One and PS4 IIRC...
 
Are there 2GB GDDR5 chips to go to 4 chips and simpler/cheaper motherboard for example?
There are 2GB chips but they can't use 4 chips because each GDDR5 chip only has a 32bit bandwidth, so 4 chips would result in half the bandwidth.

A new SoC could adopt 14Gbps GDDR6 at 128bit, though I don't know how much time and money that would require.
 
Whose moving to 7nm to save cost on chips? Design costs for 7nm run about a quarter of a billion dollars. 14/16 requires double patterning while 7nm requires quad which affects design and production costs. It’s so expensive at lower nodes now GF, TMSC, Samsung and others introduced new processes at 22 nm to accommodate customers still stuck at 28nm but who can’t afford the costs at 16 nm and below.

Right now, lower power or higher performance is necessitating the move to 7nm, not lowering costs.
 
Last edited:
If lowering of costs isn't possible, then it's not going to happen. That'd be the only reason not to though; if it's possible economically, it should be done!
 
PS3 Super Slim and Xbox 360 Model E were announced/released 4-5 months before launch of Xbox One and PS4 IIRC...

Because neither of which next-gen had BC.
 
There are 2GB chips but they can't use 4 chips because each GDDR5 chip only has a 32bit bandwidth, so 4 chips would result in half the bandwidth.

A new SoC could adopt 14Gbps GDDR6 at 128bit, though I don't know how much time and money that would require.

Thanks. I did wonder about that limitation as it came up regards PS5 RAM. I really doubt a redesign using GDDR6 as the cost would be even higher.

No I think a much more likely path Sony will take is to just reduce the retail price of what they have by $50-$100 from the beginning of next year to boost the sales of the final first-party games.
 
PS4 is selling 20M units a year. Don't see how they don't release one more hardware revision.

Was selling. Now its lower, no?

Even if PS5 is BC and released in 2019 for $399, you still expect new PS4 revision to be made and use up valuable mfg resources like ram, hdd, flash, and wafers and about $750 Million for R&D on 7nm?
 
If 70 million PS4s makes Sony $1.6 billion a year profit, another 70 million PS4's will easily be worth a $0.75 billion investment.
 
If 70 million PS4s makes Sony $1.6 billion a year profit, another 70 million PS4's will easily be worth a $0.75 billion investment.

And now consider what your profit would be if you can drive all those people to the PS5 instead.
 
You can't though. You can't convince people wanting a $200 console to buy a new $400+ console. You release the expensive product for the big spenders, and the low cost item to suck up the small spenders with different priorities. Not having a low-end option just means losing that segment. And if your rival has that option, they'll be the ones mopping up and attracting consumers to their ecosystem.
 
Back
Top