Merrill Lynch's Next-Gen console prediction

xbdestroya said:
I mean for crying out loud - $25 for the 360's DVD-ROM drive? $10 after three years? It should be $10 *now.* Hell, you and me the consumer can buy these things for less than MS if this report is to be believed: DVD Drive
Is it possible that shipping and manufacturing (estimate) costs are spread across each group? It might make the numbers more reasonable.
 
expletive said:
The ML report was based on them seeing the BOM for each console...

Well I have to correct you here and say the ML report was based on them speculating on the likely BOM for each console, rather than actually knowing what the BOM for each console is. There is a difference. ;)

It would just be absolutely impossible for them to know the BOM for almost every one of the important PS3 components - and again to restate I think they have the 360 wrong also.
 
Sis said:
It's not the same technology. The differences are:

1) The Cell chip is bigger, maybe more complex
2) The Blu-ray drive
3) The RAM
4) the number of ports/complexity of ports

.Sis

I find these analyses to be highly incomplete. Microsoft's CPU (168mm2) is not that significantly smaller than Sony's Cell DD2 (235mm2) when it comes to yeildable volume. Granted there maybe cost bebefits in fabbing the XCPU due to mask size limtations initially, but to counteract that, Cell has over 50% of it's structure devoted to redundant structures and since manufacturing flaws likely follow a power-law distibution, they will yeild extremely well with the 8th SPE turned-off. A single flaw in the Microsoft CPU will be fatal and the die is garbaged. Sony, who unlike Microsoft is a semiconductor copany, pays per SOI wafer and can use every Cell processor that doesn't make the 7SPE limit, but doesn't have a fatal error in another device. This will help minimize or sum-over the costs of Cell via the entire Sony Group. Microsoft has no such ability.

Microsoft also went for an amazingly awkward GPU configuration which requires a MCM to be built and manufactured. Not only is it more prone to errors and adds in the costs (as you need a multi-step manufacturing process), but it's cost scaling in time sucks. Sony's GPU scaling is bounded by Moore's law and can be aggressively reduced in size as Sony has done with the EE+GS. Microsoft cannot do the same, outside of Toshiba & Sony, the use of embeddedDRAM and logic sucks and has a tremendous lag time compared with bulk silicon. Microsoft is going to be stuck producing these awkward 2 ASICs + MCM + manufacturing cost for each for some time to come. Any attempt at assimilation will be in the distant future if ever, meanwhile Sony will be mass producing RSXs on their bulk CMOS5 (65nm) and above (45nm, 32nm) processes at cost.

The ports are all commodity items which are totally negligable. Costs for WiFi, Bluetooth, et al will all collapse in time as Sony has the ability, as they've done with PS2's Dragon IC and recently as I think they replaced the LSI ASIC, to create and fabricate inhouse solutions. How is Microsoft going to do the same?

Why do people never contemplate the cost of the MCM/GPU on XBox? It's a significant fixed cost that doesn't scale, much to the opposite of Cell and the RSX.
 
Inane_Dork said:
Is it possible that shipping and manufacturing (estimate) costs are spread across each group? It might make the numbers more reasonable.

Well, the report certainly lacks assembly and shipping prices, but who knows? I would really view each component in isolation though. Having a factory in China assemble and package these things isn't going to cost that much per console. Certainly when reading this report though - going back to the DVD snce it's tangible - it just certainly seems like an egregious lapse of effort, because truly that $25 figure is unbelievable.
 
xbdestroya said:
Well I have to correct you here and say the ML report was based on them speculating on the likely BOM for each console, rather than actually knowing what the BOM for each console is. There is a difference. ;)

It would just be absolutely impossible for them to know the BOM for almost every one of the important PS3 components - and again to restate I think they have the 360 wrong also.

Merrill Lynch said:
Our analysis of the bills of material for the Xbox 360 and the PS3 indicates that the PS3 will not only be significantly more costly than Xbox 360 at launch.

So based on that quote from the report youre saying they dont have access to the BOM?

If they dont then yes ia gree with you. I dont see another way to interpret it though.
 
Vince said:
I find these analyses to be highly incomplete. Microsoft's CPU (168mm2) is not that significantly smaller than Sony's Cell DD2 (235mm2) when it comes to yeildable volume. Granted there maybe cost bebefits in fabbing the XCPU due to mask size limtations initially, but to counteract that, Cell has over 50% of it's structure devoted to redundant structures.
I stopped reading by this point. 168mm^2 isn't significantly smaller than 235 mm^2? :oops: 50% of Cell's structres are redundant? :oops:

Other someone's wording themselves very badly, or they need to go check where their getting their info from.
 
expletive said:
So based on that quote from the report youre saying they dont have access to the BOM?

If they dont then yes ia gree with you. I dont see another way to interpret it though.

That is exactly what I am saying though - that they do not have access to any such documents, that they are hypothetical in nature. That the image pakpassion providedis arbitrary and 'best guess.' How could there be? There are no blu-ray drives, there is no RSX (so to speak). The Cell figure jives with someone who thinks it's a very complex chip, but not with someone who knows about the fabbing process - because really it should cost less than RSX.

So, though I can't prove it, their having no 'official' BOM documents from either of these companies is indeed what I feel is the case.
 
xbdestroya said:
That is exactly what I am saying though - that they do not have access to any such documents, that they are hypothetical in nature. That the image pakpassion providedis arbitrary and 'best guess.' How could there be? There are no blu-ray drives, there is no RSX (so to speak). The Cell figure jives with someone who thinks it's a very complex chip, but not with someone who knows about the fabbing process - because really it should cost less than RSX.

So, though I can't prove it, their having no 'official' BOM documents from either of these companies is indeed what I feel is the case.

Given the reputation that these financial companies need to uphold and protect (i work for a large one myself), imo, there really isnt a way they would release such a statement without data to back it up. Whether or not its the whole BOM, i dont know but they have access to enough to be able to make that assumption. There really is too much at stake for them to base it on less, especially considering they indict other associated companies along with this analysis.

What they consumer will or wont want is debateable but when it comes to the financial side of things and them publishing this report, you (plural) have to give them the benfit of the doubt.
 
Shifty Geezer said:
I stopped reading by this point. 168mm^2 isn't significantly smaller than 235 mm^2? :oops: 50% of Cell's structres are redundant? :oops: \

My statement was buttressed on several conditions, none of which you considered. Namely, as we've seen recently by both ATI, nVidia and Intel the fabrication of large ASICs on 300mm wafers has become viable; the problem which increases rapidly when you increase die size is the defect rate, which is counteracted by the fact that, yes, almost half of Cell is redundant or repititious structures. The XCPU's only known redundancy presumably is in it's L2 cache. In addition, Sony, unlike Microsoft, pays per SOI wafer; any IC which is yeilded without a fatal flaw can be sold in some capacity, in some device.

Just making blanket statements with those dumb :shock; icons doesn't accomplish anything. Here's a pathological case for you, if an IC is 225mm2 but composed of 200 identical structures (with 100 needed operation) and minimal control, communication and bus logic is that going to yeild worse than a 150mm2 IC that's composed of 3 major structures all of which need to be functional?
 
Last edited by a moderator:
Sis said:
What numbers are you referring to?

.Sis

These numbers from Gamespot. http://www.gamespot.com/news/6136682.html
For the second quarter ending September 30, sales and operating revenue at Sony Computer Entertainment (SCE) was 214.2 billion yen ($1.86 billion), up 79.1 percent from a year earlier. The idiom "you have to spend money to make money" holds true for SCE, as the company spent millions on marketing for the PSP and on research and development for the PlayStation 3, yet it still made a profit of 8.2 billion yen ($71 million). Last quarter it took a loss of 5.9 billion yen ($51 million).
Sony explained that shipments in both hardware and software were up for the quarter, stating "in addition to significant contribution of sales from PSP, an increase of PlayStation 2 unit sales in Europe and the US compared to the same quarter of the previous fiscal year resulted in a significant increase in sales in all geographic areas."
Sony sold a little more than 5 million PS2s this quarter, a huge leap over the 3.02 million units from the same period last year. The company also sold 3.75 million PSPs, which debuted in Japan last December during the quarter. 10 million PSP handhelds have been shipped worldwide as of October 21.
 
Damn. This thread astounds me. A world renown financial analysis firm predicts something remotely negative about the PS3 and all of a sudden they are a bunch of chimps with calculators.

It seems that it's Internet forum groupthink that anyone that doesn't predict PS3 to be a rip-roaring, utopia-ushering, new world order-like success is an uneducated hater with his head in the clouds.
 
jvd said:
They are making money on the system because in many countrys its been out for a year or so now . Thus the owners of the original systems launched are buying umd movies and games and memory sticks and less people in those areas are buying new units meaning sony is making money back.

However it took them over a year to launch (or just about a year) to launch in the 3 major teritorys

Ding Ding Ding!!! You understand what I'm trying to say now. Yes. Yet it didn't take them over a year to launch the PSP it took them 9 months to launch in 3 territories.
 
Alpha_Spartan said:
Damn. This thread astounds me. A world renown financial analysis firm predicts something remotely negative about the PS3 and all of a sudden they are a bunch of chimps with calculators.

It seems that it's Internet forum groupthink that anyone that doesn't predict PS3 to be a rip-roaring, utopia-ushering, new world order-like success is an uneducated hater with his head in the clouds.

Ahh you just don't understand. Read the forum some more and learn a little something. Just because ML states something doesn't mean it's gospel. They more than likely will be wrong. We need to see what they thought the price of the PS2 and PSP were going to be. I swear some analysts thought the PS2 was going to be $400 and the PSP was going to be $500.
 
Vince said:
I find these analyses to be highly incomplete. Microsoft's CPU (168mm2) is not that significantly smaller than Sony's Cell DD2 (235mm2) when it comes to yeildable volume. Granted there maybe cost bebefits in fabbing the XCPU due to mask size limtations initially, but to counteract that, Cell has over 50% of it's structure devoted to redundant structures and since manufacturing flaws likely follow a power-law distibution, they will yeild extremely well with the 8th SPE turned-off.

Hunh? The difference between 168mm2 and 235mm2 is significant and fairly large.

vince said:
A single flaw in the Microsoft CPU will be fatal and the die is garbaged. Sony, who unlike Microsoft is a semiconductor copany, pays per SOI wafer and can use every Cell processor that doesn't make the 7SPE limit, but doesn't have a fatal error in another device. This will help minimize or sum-over the costs of Cell via the entire Sony Group. Microsoft has no such ability.

True but its not because Sony is a semiconductor company and MS isnt. IBM designed both chips for their customers based on customer needs. Other than 360 what other use does MS have for the chip? Why compete with Intel, AMD or Texas Instruments or Sony in the semicon arena whenthey make more not competing with them?

vince said:
Microsoft also went for an amazingly awkward GPU configuration which requires a MCM to be built and manufactured. Not only is it more prone to errors and adds in the costs (as you need a multi-step manufacturing process), but it's cost scaling in time sucks.

Sony had edram as did GC. I would agree with you when it comes to assembly costs and yield timing across separate assembly lines but each manufacturer/fabber (NEC and TSMC) is highly capable of producing required chipsets on-time with high yields.

vince said:
Sony's GPU scaling is bounded by Moore's law and can be aggressively reduced in size as Sony has done with the EE+GS. Microsoft cannot do the same, outside of Toshiba & Sony, the use of embeddedDRAM and logic sucks and has a tremendous lag time compared with bulk silicon.

I disagree strongly and although I cannot find the quote, MS surely has plans to consolidate the edram with the mainengine as soon as the manufacturing process allows it (probably @65nm and surely at 45). I dont know where you get your suppositions from but they are awkward in how unrealistic they given your fairly advanced knowledge of fab techniques and timing.

vince said:
Microsoft is going to be stuck producing these awkward 2 ASICs + MCM + manufacturing cost for each for some time to come. Any attempt at assimilation will be in the distant future if ever, meanwhile Sony will be mass producing RSXs on their bulk CMOS5 (65nm) and above (45nm, 32nm) processes at cost.

see previous para.

vince said:
The ports are all commodity items which are totally negligable. Costs for WiFi, Bluetooth, et al will all collapse in time as Sony has the ability, as they've done with PS2's Dragon IC and recently as I think they replaced the LSI ASIC, to create and fabricate inhouse solutions. How is Microsoft going to do the same?

Again do you think that IC knowledge and consolidation is the realm of Sony only? Epox, Soyo, Abit and much more less mainstream motherboard manufactures (not to mention SIS who is helping with Mainboard design) could not only do this but predict to you how long before they could afford to consolidate... and MS cant? I am again perplexed by your contentions.

vince said:
Why do people never contemplate the cost of the MCM/GPU on XBox? It's a significant fixed cost that doesn't scale, much to the opposite of Cell and the RSX.

cell is way more complex than xenon but you expect it scale more quickly? the main engine of xenos is less complex and large than RSX... the Edram is the only monkeywrench but ...

MS designed this entire system for easy consolidation and future economies of scale... It is a primary goal of theirs, that has been well known for as long as we have know about 360/xenon development. All of your contentions fly in the face of MS' planned consolidation and cost benefit plans, and frankly I cant agree with you this time vince.
 
mckmas8808 said:
Ahh you just don't understand. Read the forum some more and learn a little something. Just because ML states something doesn't mean it's gospel. They more than likely will be wrong.
My ass. In the real world I'd put my money on the word of Merrill Lynch before I'd take stock in what a bunch Internet forum nobodies talk about. But then again, maybe Merril Lynch got in the fourtune 500 by pure luck...

When it comes to technical/hardware discussions I'd lean a bit more towards what I know, but when it comes to financial matters I'm more inclined to trust the experts.
 
blakjedi said:
MS designed this entire system for easy consolidation and future economies of scale... It is a primary goal of theirs, that has been well known for as long as we have know about 360/xenon development. All of your contentions fly in the face of MS' planned consolidation and cost benefit plans, and frankly I cant agree with you this time vince.

Oh and the PS3 is not? Of course Sony can implement 65nm tech in the PS3 in 2007 right?
 
Alpha_Spartan said:
My ass. In the real world I'd put my money on the word of Merrill Lynch before I'd take stock in what a bunch Internet forum nobodies talk about. But then again, maybe Merril Lynch got in the fourtune 500 by pure luck...

When it comes to technical/hardware discussions I'd lean a bit more towards what I know, but when it comes to financial matters I'm more inclined to trust the experts.

So why not explain the errors that were explained earlier in this thread then? A $500 PS3 compared to a $250 Xbox 360?:???:
 
expletive said:
Given the reputation that these financial companies need to uphold and protect (i work for a large one myself), imo, there really isnt a way they would release such a statement without data to back it up. Whether or not its the whole BOM, i dont know but they have access to enough to be able to make that assumption. There really is too much at stake for them to base it on less, especially considering they indict other associated companies along with this analysis.

What they consumer will or wont want is debateable but when it comes to the financial side of things and them publishing this report, you (plural) have to give them the benfit of the doubt.

Expletive well I don't know - I mean is it the norm for an analysis to have access to the BOM for products that haven't even lauched yet? And for that BOM to contain price quotes for components that in some cases don't exist yet? I mean I am *certain* that their numbers are based on something, I have no doubt. And that there was a logical process followed in reaching them.

Certainly I know it always takes financial firms a while to get the BOM with regard to Apple's iPods, something they're always after, and to even predict what new iPod models might be in the pipe - their main source often being fan sites. So I don't think it's outrageous to accuse ML of creating numbers based on facts, but that the numbers themselves do not necesarilly reflect a hard and fast reality.

I mean some of these numbers just don't make sense, that's all I know.

I'm hardly trying to bring an indictment though; I wouldn't expect any other firm's numbers to be any more accurate at this point.
 
mckmas8808 said:
So why not explain the errors that were explained earlier in this thread then? A $500 PS3 compared to a $250 Xbox 360?:???:
You mean "percieved errors."

I think they have better access to financial details than we will ever have. Thus, they get a HUGE benefit of the doubt.
 
If cheaper cost is the dominant factor for success, as it's seems to be obessed over on this forum, how come Nintendo is not winning the console race?

PS3 is more cost, for more functionality, and more power. It worth paying more for, and a PS3 priced $100 more than the Xbox 360 will still beat it in number of units sold by Christmas 2006.
 
Back
Top