Next Generation Hardware Speculation with a Technical Spin [pre E3 2019]

[goonergaz]

6 months makes more sense - maybe 10 months after initial manufacture where RAM prices likely dropped a lot.

Agreed, but either way the point is Sony won't be losing $100 per console - it's not black and white like that, if we make a very rough guesstimate at $50 per unit (assuming a similar scenario to PS4 where a game/PS+ purchase brings the loss down) and this was for the first 6 mths (around 7m units) that's a total of $350m loss and then after than they are breaking even.

That's a far cry from the $500m - $2bn posted earlier.

 

[Shifty Geezer]

(assuming a similar scenario to PS4 where a game/PS+ purchase brings the loss down)

It's lost profits. The first 5 million units sold at $100 loss is $1 billion less that Sony makes. If Sony sell a $499 console at $100 loss, plus one game and one year's subscription, and sell 5 million units, they'll be $1 billion poorer than if they sell the same package but with a $399 console at zero loss. The money men will be weighing up that amount, and not how much the balance sheet per unit lies in the red.

 

[AlNom]

Indeed. Even if the brand is strong with consumers, It's a substantial dollar risk, and at the end of the day it has to weigh upon what the benefit is to the company vs winning some pissing contest online, so to speak.

The trick for their bean counters is to only spend enough to not be an utter market failure and hit closer to profit instead of burning cash for internet loyalists.

Microsoft has a much bigger uphill battle to climb, and so they have been and will be accordingly aggressive in their spending sprees. Scorpio & buying up studios -> leading to the supposed dual SKU in order to win the pissing contest and also hit the price conscientious consumer.

That Sony may have been targeting/attempting to launch in 2019 is certainly an aggressive maneuver, but it would have to be at a somewhat conservative performance bump & price point strategically with decisions made well in advance while considering that both GDDR6 and 7nm are new and hella expensive, and who knows what they have planned on the storage side of things. Afterall, there would only be so much they could pack in at a reasonable price point, and developer feedback can additionally inform them whether they should delay with a bevy of reasons (not just a single one).

"Rushing" to launch 2019 when they're just finishing up their current gen release schedules may not be the best either, which may or may not imply an extended and lacklustre cross-gen drought of software (if that makes sense) for a hypothetical launch.

That said, architectural decisions would have to be made well in advance when considering any plans for BC, which may have consequences for an earlier targeted launch (less to time to work the kinks).

...... I digress.

I know nothing. :V

edit: 9001.

/Al B. Rant

 

[goonergaz]

It's lost profits. The first 5 million units sold at $100 loss is $1 billion less that Sony makes. If Sony sell a $499 console at $100 loss, plus one game and one year's subscription, and sell 5 million units, they'll be $1 billion poorer than if they sell the same package but with a $399 console at zero loss. The money men will be weighing up that amount, and not how much the balance sheet per unit lies in the red.

Surely you mean $500m?

Those same bean counters took a loss last time and it's paid dividends.

If the previous loss was $60 and it's worked out really well then there's every chance they will do something similar, with a potential increase.

And it's more than just a pissing contest...it will be to do with what they can achieve and if it's worth it...that extra loss might mean ray tracing for example. They might save some of the advertising budget towards it, after all this gen they are in a strong position so won't need as much... especially if they launch ahead of MS.

Anyway, time will tell.

 

[BRiT]

If the previous loss was $60

I doubt it was. The ranges could be as low as $18. This depends on what scenario would pull them to break-even -- 1 Game, 1 Game + PSN Sub, 2 Games + PSN Sub.

It doesn't seem like any of us know the profit margins on Network Pass. However we know it's not 100%. It's likely more than 25%. I doubt it's even 75%. So should we assume its a very robust 50%? Sony increased the suggested retail price of PSN+ in 2017 in the US to $60 from $50, but there are typical sales at $40 even to this day. (Amazon Price History - https://camelcamelcamel.com/PlaySta...hip-Digital/product/B004RMK5QG?context=search )

Assumption: 1 Game Purchase
The profit on 1 game purchase would be $18 Publisher License (based on $60 game).
Result: $18 Profit

Assumption: 2 Game Purchase
The profit on 2 game purchase would be $36 Publisher License (based on $60 game).
Result: $36 profit

Assumption: 1 Game Purchase + Network Subscription
The profit on 1 game purchase would be $18 Publisher License (based on $60 game).
The profit on 1 Network Subscription would be $20 (based on $40 price typical sale).
Result: $38 profit

Assumption: 1 Game Purchase + Network Subscription
The profit on 1 game purchase would be $18 Publisher License (based on $60 game).
The profit on 1 Network Subscription would be $25 (based on Retail $50 price at launch in 2013).
Result: $43 profit

Assumption: 2 Game Purchase + Network Subscription
The profit on 2 game purchase would be $36 Publisher License (based on $60 game).
The profit on 1 Network Subscription would be $20 (based on $40 price typical sale).
Result: $56 profit

Assumption: 2 Game Purchase + Network Subscription
The profit on 2 game purchase would be $36 Publisher License (based on $60 game).
The profit on 1 Network Subscription would be $25 (based on $50 price at launch in 2013).
Result: $61 profit

​

 

[Deleted member 13524]

$100 loss during the launch window isn't out of the realm of possibilities considering it was done before with the PS3, and during a time where Sony didn't have >40 million people paying them a $60 annual fee.

Bankrolling the console to increase the chances of a sizeable userbase advantage during the launch window isn't anything new.

Even more if they're actually planning for a more expensive "premium" service.

If they put the PS4 plus titles available for the PS5 as BC only for people who pay for the premium plus, many won't resist the urge.

 

[Shifty Geezer]

Surely you mean $500m?

Yes. Originally I had first 10 million being sold at a lost, but refactored that downwards to a safer number - PS4 took some 8 months to get to 10 million.

Those same bean counters took a loss last time and it's paid dividends.

Last time they knew that loss wouldn't last long. If you can't be sure of a quick price reduction, you're facing significant drops in profitability.

If the previous loss was $60 and it's worked out really well then there's every chance they will do something similar, with a potential increase.

If the previous $60 loss was expected to last two years, it may well have not happened.

$100 loss during the launch window isn't out of the realm of possibilities considering it was done before with the PS3...

PS3 was an absolute money pit because of its loss leading! Sony were willing to take a significant hit for a long period, and it wiped PS out. By being far more conservative this time, and only taking a short-term loss in order to squeeze an unexpected but highly valuable 8GBs at launch, PS4 has made many monies. Had Sony put in more hardware and taken a large loss on PS4, they likely wouldn't be better off in terms of market share or revenue but would be down on profits. PS4 was well engineered at the sweet spot between power (market desirability) and cost.

The issue isn't whether a loss leader is possible or not, but how long a loss leader will be around. If a $50 loss can be eliminated within 6 months of launch, it's not so bad and high probability. If the hardware can't be cost reduced by $50 for 18+ months after launch, it's a helluva lot more expensive. Rather than just saying, "it could happen," speculation should revolve around identifying which components are likely to significantly cost reduce in the short term to enable a loss-leader. If no components can be identified, probability is slim.

 

[Xbat]

PS3 was an absolute money pit because of its loss leading! Sony were willing to take a significant hit for a long period, and it wiped PS out.

Now with PS+ though it's a different time. I know I've said this many times but I believe in the present and future that the network is all that matters and anything that makes that network grow is the be all and end all.

The money they make on PSN dwarfs everything else and so that must play a massive party in the decision making. Go cheap and sell lots of consoles but possibly lose the big spenders who buy 10+ games a year and definitely sub to the network?
Would be fascinating to know what decisions they making and why.

 

[Tkumpathenurpahl]

Digital has also seen a massive rise this generation, which will further increase the average loss per unit they'll be able to sustain, even if they follow the same "1 game + PS+" model.

 

[Deleted member 11852]

It's lost profits. The first 5 million units sold at $100 loss is $1 billion less that Sony makes. If Sony sell a $499 console at $100 loss, plus one game and one year's subscription, and sell 5 million units, they'll be $1 billion poorer than if they sell the same package but with a $399 console at zero loss. The money men will be weighing up that amount, and not how much the balance sheet per unit lies in the red.

The economics of consoles is also rather opaque. When Sony announced they were making a profit on every PS4 sold I took that to mean that the total cost of BOM, packaging and shipping was now less than the baseline RRP in each region, not that Sony has recovered all their R&D costs for the project.

I do wonder how Sony and Microsoft account R&D of consoles to determine, de facto, whether any given console/generation is profitable. I.e, did PS4 R&D just suck from PS3 lifecycle profits and begin life on a project cost basis basis at launch at $0 profit/loss or did PS4 start at $0 profit/loss back in 2007 - when R&D began - and was racking up costs until launch in 2013.

 

[chris1515]

Yes. Originally I had first 10 million being sold at a lost, but refactored that downwards to a safer number - PS4 took some 8 months to get to 10 million.
Last time they knew that loss wouldn't last long. If you can't be sure of a quick price reduction, you're facing significant drops in profitability.
If the previous $60 loss was expected to last two years, it may well have not happened.
PS3 was an absolute money pit because of its loss leading! Sony were willing to take a significant hit for a long period, and it wiped PS out. By being far more conservative this time, and only taking a short-term loss in order to squeeze an unexpected but highly valuable 8GBs at launch, PS4 has made many monies. Had Sony put in more hardware and taken a large loss on PS4, they likely wouldn't be better off in terms of market share or revenue but would be down on profits. PS4 was well engineered at the sweet spot between power (market desirability) and cost.

The issue isn't whether a loss leader is possible or not, but how long a loss leader will be around. If a $50 loss can be eliminated within 6 months of launch, it's not so bad and high probability. If the hardware can't be cost reduced by $50 for 18+ months after launch, it's a helluva lot more expensive. Rather than just saying, "it could happen," speculation should revolve around identifying which components are likely to significantly cost reduce in the short term to enable a loss-leader. If no components can be identified, probability is slim.

PS3 was losing 200 to 250 dollars per console sold at 599 dollars, it was insane. An analyst seeing the 2021 forecast of Playstation division said PS5 will be a bit more subsidized than PS4, he never said at PS3 level... After I am sure they know what they do and they have a solution to lose less money soon after the launch. There is a big margin between lose 60 dollars per console like PS4 at launch and 200/250 dollars per console like PS3 at launch.

 

[MrFox]

They didn't plan to lose over $200 per console at launch. The blue laser was over $100 because of production issues which didn't last long. The cell had very bad yield at launch, multiplying the cost of working chips. None of those were part of the expected BOM cost at launch. The head of sony was supposedly very angry at the situation.

The cost reduction was quite dramatic with ps3 because it combined multiple aspects: the low cost shrink of both cell and gpu, the cell yield issues being solved, the blue laser optical block went from over $100 to $8 because they improved yield, volume, and competition of bluray drives.

The ps4, comparatively, had no yield issues, the ODD is already near the lowest possible cost, and there are no more easy shrink, since each new node is extremely expensive.

I wouldn't be surprised the ps4 "plan" included and overhead (hoping for break even at launch), and with everything going so well they had the margin left to get that 8GB approved with a small loss.

No guarantee they won't have a big loss per unit at the ps5 launch, but they wouldn't do it on purpose.

 

[mrcorbo]

The cost reduction was quite dramatic with ps3 because it combined multiple aspects: the low cost shrink of both cell and gpu, the cell yield issues being solved, the blue laser optical block went from over $100 to $8 because they improved yield, volume, and competition of bluray drives.

They also hacked bits of hardware off at each revision. Each successive model was less capable overall than the last. Not saying this was a bad thing, since the bits removed were obviously not supporting core features (the hardware BC was probably the most impactful cut), but hopefully a potential PS5 wouldn't, from the start, come with hardware features that were so unnecessary to the most prominent usages of the device as to not be missed when they were removed.

See also: Xbox One jettisoning the Kinect 2 pack-in.

 

[MrFox]

They also hacked bits of hardware off at each revision. Each successive model was less capable overall than the last. Not saying this was a bad thing, since the bits removed were obviously not supporting core features (the hardware BC was probably the most impactful cut), but hopefully a potential PS5 wouldn't, from the start, come with hardware features that were so unnecessary to the most prominent usages of the device as to not be missed when they were removed.

See also: Xbox One jettisoning the Kinect 2 pack-in.

Oh yes I forgot about the entire PS2 hardware on the board. They even have a patent showing they considered making it as an optional module. After a few years, the market for it is probably not enough to warrant developing a product for it. It should have been an optional module right at launch. (yep, like kinect 2)

In 2010 there was a big email survey to psn accounts. They asked about BC, remasters, bluray, first parties, favorite genres, usage pattern, online gaming, etc... It's obvious the ps4 was designed, in part, based on the result of that market research. 2010 would be when design goals should be decided for a 2013 launch, and considering the demography was exclusively ps3 owners, it was the best place to ask about BC. Obviously not enough gamers cared about it, at least not enough for either a higher BOM, or an optional module. (a fast cell emulation being currently impossible)

This time we have x86 and nothing crazy (the wildest was the xb1 esram, and it was easily solved by mapping it elsewhere), so BC comes for free as long as they keep the hardware reasonably compatible. BC is the number one reason I don't see ARM being ever considered. But I wonder about new GPU architecture, how having BC as an absolute design goal can prevent new architectures from being considered. I think we will get additions which are not in the way, like simultaneous FP/INT, double rate fp16, quad rate int8, tensor, RT, new codec blocks, maybe new real time memory compression, etc... As long as previous gen games not using them are still running equal or better performance and are binary compatible (or at least translation friendly from the hypervisor or abstraction layers). I think someone here mentionned sony's gnm not being very friendly to translation?

 

[iroboto]

how having BC as an absolute design goal can prevent new architectures from being considered.

BC seems a licensing/business challenge than a technical one.
MS has managed 2 different architectures to run and to some degree enhanced.

 

[mrcorbo]

BC seems a licensing/business challenge than a technical one.
MS has managed 2 different architectures to run and to some degree enhanced.

The 2 architectures aren't that different, though. OTOH, I'm not sure there is all that radically different of a hardware architecture out there to choose anyway. But, if there were, the ease of achieving BC might stack the deck in favor of more familiar.

 

[iroboto]

The 2 architectures aren't that different, though. OTOH, I'm not sure there is all that radically different of a hardware architecture out there to choose anyway. But, if there were, the ease of achieving BC might stack the deck in favor of more familiar.

More or less yea. I mean if you get nvidia and older amd stuff working on your newer amd stuff, and the power PC stuff to work, then the only thing more radical is Cell. I don’t see BC being a technical issue for ps5

 

[AlNom]

I don’t see BC being a technical issue for ps5

Could have said the same thing about Gamecube/Wii/WiiU. :V

Comes down to paranoia, time, and investment on the software side, of course.

Not that I'm suggesting anything for Sony. I know nothing. (devil's advocate)

 

[mrcorbo]

More or less yea. I mean if you get nvidia and older amd stuff working on your newer amd stuff, and the power PC stuff to work, then the only thing more radical is Cell. I don’t see BC being a technical issue for ps5

Those more significant translations only achieved partial BC, though. And it was a tremendous amount of effort to get that. Going forward, with persistent game libraries, the expectation will be full BC. The way Xbox 360 achieved BC with selected OG Xbox games and the One family obtained BC with selected games from both of the prior gens is not going to work.

 

[BRiT]

The BC on Xbox One being only "select games" is because of nontechnical license issues.