My buddy who used to design these things* agrees with most of your assessment. He thinks the chip cost will be slightly higher, though:
“The reduction in SPUs from 8 to 7 helps; but this does not reduce all of the logic in MFCs and bus. The LS and L2 cache should be able to be repaired on die. They will be struggling to get above a 30% yield. Long term they will need to get over 80%.”
-he’s not using any insider information in his estimate, rather his experience with fab processes. We’re a semiconductor company after all. :-)
Especially amusing is the stories all quote the mistaken $900 total that we all know was wrong in the original! Amazing the speed with which misinformation spreads on the miraculous intarweb :-D
But nonetheless, I do indeed smell big trouble for Sony…
“There is no change in our original plan to release the console in spring 2006.”
The difference between that and “I, Kei Sakaguchi, 100% guarantee the PS3 will be available in the Spring” is amazingly huge in corporate-speak. And he would have said something closer to the latter if he could have (it certainly would have been a more investor-soothing statement).
In the spectrum of possible “denials” of the Merrill Lynch report, the one they picked falls very much on the weak end, and is about as much a confirmation as any denial could be.
The only way they can pony up and counteract this news now is announce a release date and pricing. Failure to do that is to allow their shares to slip on ‘misinformation’.
That being said, if they don’t announce, then there’s a good chance the analysts are right, since an announcement of date and price would send their stock straight back up.
In all likelyhood they’ll announce at GDC during Phil Harrison’s keynote. They’re under no obligation to alter their timeline for releasing information. And there’s no chance this analyst is “right” considering how abosolutely laughable and nonsensical the report is.
I will bet you $50 that by January 13, 2007, the PS3 will have sold less than 600,000 consoles in North America according to NPD statistics.
If NPD statistics aren’t available for that precise date, we’ll have to get legalistic, but my hunch is it won’t even come close.
That, to me, would corroborate the hell out of what you consider this laughable and nonsensical analyst report. The bottom line of this report is that the PS3 is getting pushed way, way back. I believe that bottom line. You, apparently, don’t?
Now, mind you, I do agree that it’s pretty hilarious how Lynch’s own report mis-totaled their own cost estimate. But I do think Sony is in deep trouble with the PS3, and I do think their 2006 PS3 North America launch will be worse than Microsoft’s 2005 Xbox 360 North America launch. And I’m putting my money where my mouth is.
So what do you say? In fact I’ll take up to four bets on this. Anyone? Anyone? Bueller? Bueller?
The thing is, the report basically makes two arguments:
The PS3, as it stands, will cost Sony somewhere around $800-900 (ML can’t make up their mind so I’ll just split the difference) per unit.
Because of 1, they can’t launch in the spring because the costs will be too high, and they may not even be able to launch before 2007. Initial supplies will be limited because Sony can’t afford anything else.
Nobody seems to be arguing that the PS3 won’t be delayed (to fall 2006 or some other date), but it’s been pointed out both here and elsewhere that ML’s price estimate leaves a lot to be desired. So their prediction of a delay is most likely correct (not that it’s a big risk to say “the PS3 will not make spring”), but the argument they use to back it up is most likely wrong. Likewise with the supply issue – I would be shocked if there’s enough PS3s in the first batch to meet demand, but if Sony has supply problems it won’t necessarily be because the thing costs them $800 to build. Charles’ argument for a delay – the lack of debug kits, finalized hardware, etc. – seems better and more firmly grounded than anything in the Merrill Lynch report. So whether or not the report is “right” depends on which of its claims you’re looking at, and even the “right” conclusion was probably based on bad data.
Those are all excellent points. If Brad agrees with you – if he feels that the report is tripe, but that the PS3 probably will be delayed for other reasons (no dev kits yet, etc.) – then yeah, it wouldn’t make sense for him to take that bet.
I also notice no one else is particularly interested in taking the bet yet :-) We’ll see what happens by tomorrow, but if no one here is willing to bet on Sony, then it’s pretty clear that Sony’s in deep, deep trouble. If Microsoft has all of 2006 to itself in North America, it’s going to be a very hard row for Sony to hoe.
Don’t forget kids you can’t have analyst without A-N-A-L! Besides that, I have yet to see a mainstream analyst projection about the games industry that turned out to be even close to accurate. Not saying it doesn’t happen, but I haven’t seen it.
I’ll take the bet, no problem. For the record, I always thought Spring was an optimistic goal, one Sony would like to make in at least one territory. I also think it’s even possible they would launch in the US first since it is the most competitive market. But a fall launch has always been a likely fallback and I don’t think Sony will have any trouble outselling the Xbox on your proscribed terms. To nail it down, it’s sales from now to January 16 next year, regardless of the American launch date, versus Xbox 360 from its launch to 1/16/06. Correct?
Consider this, STI have had a lot more time to work out manufacturing kinks in the Cell. The design has already seen 3 core revision. By contrast MS had far less time to work out the problems from first silicon to launch. IBM will have had, what, two years? Is going to be selling blade servers this year, Toshiba is putting Cell in their TVs… Hell, like a year ago IBM demonstrated a Cell CPU clocked above 4GHz.
As for the GPU, well, assuming it’s little modified from the G70, an architecture around for a while which nVidia has lots of experince producing, it shouldn’t be too great a challenge, especially for a fall launch. This is, again, in contrast to ATI’s brand new design which wasn’t in production until only a few months before launch.
Sony could also have an advantage if we aren’t talking about a simultaneous worldwide launch.
NORTH AMERICAN sales only, yes. OK, we’re on. I will however bail on my offer to make this bet multiple times… it’s you and me, Brad, and that’s it :-)
Feel free to remind the heck out of me as the year unfolds!
Sony could also have an advantage if we aren’t talking about a simultaneous worldwide launch.
Actually I think that one works against you – they’re obviously going to launch in Japan first, and that will mean their North America launch will get pushed back…
Honestly, I’d love to see Sony bungle it. Not because I’m against them or the PS3, but simply because I’d much rather see the main consoles be on more even footing. The xbox vs ps2 stuff sucked because sony had the numbers; this pretty much forced a lot of developers to use the PS2 as their main SKU, regardless of what they would’ve liked. Should the numbers stay more even, it’s more likely going to be conducive to platform exclusives rather than LCD games.
And I don’t know about you guys, but I always thought the fact that almost all the games were available on almost all the platforms was more than a little ridiculous. Good for the consumer on one level, bad for them on another. Games are better when devs don’t have to juggle four platforms at a time.
Wouldn’t the consoles being more even give more an incentive for developers/publishers to go multi-platform? I mean, if both 360 and the PS3 hold 40% of the market share each, it’s much more desirable to make sure all games are both machines than to keep it to one…than say the current situation, where I think the PS2 has a good 70% or so of the market share.
If anything, with rising development costs I’d imagine that almost all non-first party games to go multi-platform sooner or later.
That would be a reasonable bet if they launch in late November like MS did. If the US launch date is more like September, then I think they can get more than 600,000 of them here by the middle of January.
Which is really kind of moot. Sony isn’t launching in some sort of time warp where they get to compare themselves against what MS did the previous year. In the real market, they’ll be up against what MS can do this fall (shipments, games, marketing, “buzz”, etc). It’s academicly interesting to compare, but doesn’t amount to much, really.
That, to me, would corroborate the hell out of what you consider this laughable and nonsensical analyst report. The bottom line of this report is that the PS3 is getting pushed way, way back. I believe that bottom line. You, apparently, don’t?
The report said two things, really. One, the PS3 costs $900 for Sony to make (or $800, who knows which estimate they meant us to believe since they can’t add). Two, they’re gonna miss Spring. Number two I believe. “way way back” is an exaggeration, I think. Sony said Spring 2006, and I always took that to mean around May or June. They’re gonna be like 3-5 months late.
Not necessarily. There are some other financials to consider.
Let’s say Microsoft and Sony are both vying to have your hot 3rd party title as exclusive to their platform. They’re going to pay you for it, or give you a break on royalties, or co-market it, or all of the above.
So if you go dual-platform, you expand your potential market. That’s a plus. But you lose out on that exclusivity money. And you increase your development costs, because it’s not like you just hit the “compile for PS3” button and you’re done - there’s a pretty significant engineering investment in each platform. Are you going to have to do all your own networking and matchmaking stuff on PS3, or are you going to hook into this “Hub” thing, or will “Hub” even work like that? If it does, is doing both that and Live on the 360 a significant increase in development costs?
You have to balance the exclusivity deal you’ll get against the increased cost of developing for two platforms. Then you have to ask yourself, “if GTA sold X million units on the PS2 alone, and Halo 2 sold 5 million on the Xbox alone, am I really missing out on that many sales by being platform exclusive?”
It’s a calculated risk either way, I’d say. I think you’ll see a whole lot of cross-platform games, and you sort of already are (Biohazard/Resident Evil for both PS3 and 360). But I don’t think the exclusivity game is going to settle down any.
Sony Corp’s launch of its next-generation PlayStation 3 video game console could be delayed if industry specifications for some of its technology are not finalized soon, although it is still aiming for a spring rollout, it said on Monday.
. . .
“We’re waiting for them until the last possible minute, but the launch could be pushed back if they’re not decided soon,” the spokeswoman said. If the PS3 is not ready in time, the company will choose the next best timing for the launch, she said.
Here’s an interesting statement by Nintendo’s Reggie Fils-Aime, in an interview with Engadget:
We have been sharing the controller mechanics with developers across the world. We have shipped over a thousand controller dev kits to developers so that they can begin getting experience with the controller mechanics. The response to the controller has been fantastic. Developers are truly embracing the innovation. They’re embracing the approach and quite frankly, they’re embracing our vision; our vision of creating games that are as sophisticated as the core gamer wants it to be but could also be as straightforward and as accessible as brand new gamers.
He’s saying “controller dev kits” which probably means controller add-ons to existing Gamecube dev kits. However, since the GC is supposedly very similar to the Revolution this might actually serve to create an impressive launch library for the Revolution. If those 1,000 controller dev kits translate to even a few dozen games in the launch lineup that would be more than either Sony or even Microsoft have to offer…
Almost all non-first party games are already multi-platform, and have been for years. Not necessarily at launch, though; trying to do it at launch requires targeting a mix of lowest-common-denominator features, which is what Charles is complaining about (and is 155% correct about).
