View Full Version : NV40 failure, Nvidia out of graphics?
I was wondering if the NV40 is anything like the NV30 was, will Nvidia drop out of the graphics market? Maybe sell their IP to another company, or will they push on for another generation?
My guess is if the NV40 is a flop, they will exit the graphics market. If the NV40 does just "average" or better, they will stay for another generation.
AlphaWolf
15-Oct-2003, 20:20
Define the scope of the failure. It would have to be a massive failure (ie their manufacturing partners refuse to produce it) in order for them to drop out.
cthellis42
15-Oct-2003, 21:32
Hardly. All things considered, few really note the "failure" of this generation either. And considering the companies have a lot more cash and other methods of support, I don't see nVidia or ATi or analling themselves unless they have two full generations with ABSOLUTELY horrible performance which can't even be compensated for by other means. Erosions will likely be slow, and one REALLY pull one out if the company decides fighting over X particular market just isn't worth spending the resources for the low margins any more.
If nvidia fails at the nv40 (like someone else said, depends on how bad), you'll probably just see more dirty tricks more than anything. I doubt they'll go out of business anytime soon. Probably the most that would happen is you'll see them relegated to the backseat of the graphics business, with ATI as the driver.
Besides, now that nforce has taken off, you'll see them milking that for all it's worth. They even named one of their new mediaq chips "nforce3 go" :roll:
No is the answer you're looking for. And, as someone else noted - The NV3x can't really be called a disaster or a flop.
Now, the NV40 and NV50 being a flop? That's a different story, and I suspect nVidia would pull back considerably if that was the case. I doubt they'd sell their IP though, even in that case.
And, as someone else noted - The NV3x can't really be called a disaster or a flop.
:shock: :shock: :shock:
Actually, you're right... it's an understatement... rather should be called for 'nightmare' or something similar.
3dilettante
16-Oct-2003, 03:45
While the nv30 core was short lived, the value versions seem to have sold rather well, despite the shortcomings of their base architecture.
Does anyone have some money figures that show if the nv30 made any money?
digitalwanderer
16-Oct-2003, 14:25
Does anyone have some money figures that show if the nv30 made any money?
I really don't think it did and I think everyone who is claiming the nV3x is not a failure is very much wrong and in denial. :) (BTW-Hi BM! Sorry ya got banned at EB, they're protective of me. ;) )
NV30 was a flop, not the entire NV3x family. NV31 and especially NV34 are rather successful.
The NV3x architecture is pretty good with DX8 titles, and with the "optimisations" it can muddle through on DX9 apps. It's hardly as if it's crawling along at 1fps in every single application ever, and is blowing up people's computers.
Saying it's a "disaster" is flat out lying. Is it as good as the ATi offerings? Clearly not, but that doesn't mean it's failure.
digitalwanderer
16-Oct-2003, 16:14
Saying it's a "disaster" is flat out lying. Is it as good as the ATi offerings? Clearly not, but that doesn't mean it's failure.
I humbly beg to differ. I feel the damage that the nV3x did to nVidia, or rather the damage that nVidia did to themselves in their questionable promotion of the nV3x series, have had and will have a huge negative impact on nVidia that they will be a long time recovering from. :(
nonamer
16-Oct-2003, 17:23
Saying it's a "disaster" is flat out lying. Is it as good as the ATi offerings? Clearly not, but that doesn't mean it's failure.
I humbly beg to differ. I feel the damage that the nV3x did to nVidia, or rather the damage that nVidia did to themselves in their questionable promotion of the nV3x series, have had and will have a huge negative impact on nVidia that they will be a long time recovering from. :(
In case you haven't figured, the NV3x series is outselling the R3x0 series unit-wise, last time I checked.
Rodéric
16-Oct-2003, 17:58
I wouldn't mind nVidia leaving the 3D market, provided PowerVR takes its place :)
Nite_Hawk
16-Oct-2003, 19:12
Saying it's a "disaster" is flat out lying. Is it as good as the ATi offerings? Clearly not, but that doesn't mean it's failure.
I humbly beg to differ. I feel the damage that the nV3x did to nVidia, or rather the damage that nVidia did to themselves in their questionable promotion of the nV3x series, have had and will have a huge negative impact on nVidia that they will be a long time recovering from. :(
In case you haven't figured, the NV3x series is outselling the R3x0 series unit-wise, last time I checked.
Do you have a link for this? I'm rather curious to see how the various cards have been stacking up in terms of sales...
As Dig mentioned though, even if the NV3X cards are selling well, it's atleast partially due to the massive scucess of the GF4s. IF the NV3X cards don't perform well with DX9 titles (they are argueably decent with DX8) then a lot of people will not be interested in buying the next generation, even if they are good cards.
Nite_Hawk
cthellis42
16-Oct-2003, 19:35
In case you haven't figured, the NV3x series is outselling the R3x0 series unit-wise, last time I checked.
Strip out the 5200's and you'll see who's more embraced by the game/enthusiast community, of course. (And last I knew there were few sales numbers, but that the 5200 had 70% of the sales.)
I'm questioning the level of "failure" of the nv3x mainly from the general public's perspective. Other than those of us who REALLY read the review sites and are in general video and gaming nerds--with whom nVidia has lost much ground--just about EVERYTHING that's been going on the past year has flown right off the general public's radar. Most still carry lingering "nVidia is the best" attitudes--even among gamers who just don't care to read too many reviews--and the bulk of cards sold and profit made from any line are the mainstream ones anyway, and it takes a LOT for people who buy those cards to notice, and for OEM's to shift gears.
I certainly think that nv30's overall lifespan, and the manner in which nVidia has been handling themselves during it equates to a big failure, but I certainly don't account for the bulk of their sales. :P I even see that they've lost a lot of ground and respectability from certain sectors, but from enough to make a MAJORITY of people think nv30 was a failure, public and businesses alike? No way. And even if nv40 follows up with equivolent problems compared to R420, there's still plenty of wool left to pull over the general public's and companies' eyes. (Especially since they usually shut their eyes voluntarily anyway.)
Bouncing Zabaglione Bros.
16-Oct-2003, 19:55
In case you haven't figured, the NV3x series is outselling the R3x0 series unit-wise, last time I checked.
Be interesting to see what the profit margins are. Nvidia have had poor yeilds of NV3x and are cutting margins in order to keep market share. IIRC, dollar for dollar, ATI is making more money for each card they sell than Nvidia.
NV3x is not a failure in general terms, but if you look at what Nvidia *intended* for NV3x to do, how they expected to dominate the market, control DX9, and sell loads, I think you'd have to agree NV3x has not achieved it's aims. When you take into account losing the high end technology race, cancelling NV30, and tarnishing Nvidia's name with the digerati, I don't think you could call NV3x a success. At best it's a failed attempt to do anything more than tread water in the DX8 marketplace which only came unstuck because of ATI's stellar DX9 part.
digitalwanderer
16-Oct-2003, 23:45
I think all those 5200s are going to come back and hurt nVidia harshly. I'm already seeing "I have an FX but my framerates are kind of low" messages crop up on some sites, I expect it to be raining 'em before long. :( (Frowny because I am NOT happy that so many people are going to be getting a lousy gaming experience who's only mistake has been to fall for some PR hype. :( )
RussSchultz
17-Oct-2003, 00:16
The GF4MX didn't hurt them, I don't see why the 5200 will.
digitalwanderer
17-Oct-2003, 00:21
The GF4MX didn't hurt them, I don't see why the 5200 will.
But the GF4MX didn't bill itself as "dx9 for $79". ;)
ricercar
17-Oct-2003, 00:25
It's absurd to think NVIDIA's going to fold with only failure in consumer graphics. The NV3x Quadro FX workstation GPUs simply rape the competition, nForce is the clear performance king for the AMD platform, and the purchase of MediaQ and deal with Transmeta gives NVIDIA a good shot at dominating handhelds in a few years.
NVIDIA is the world's largest fabless semiconductor firm, fastest firm in history to reach $1 Billion USD in annual revenues, and just last month was still the fourth fastest-growing technology company.
Saying NVIDIA will fold with a failure of NV40 is like saying Apple will forld with $6B cash in the bank. Not going to happen. NVIDIA will be around for a long time to come.
thegrommit
17-Oct-2003, 00:33
You forget that ATI was in Nvidias and 3dfx's shadow for many years. Yet look where they are now.
ricercar
17-Oct-2003, 00:51
I forget nothing. :D ATI used to be king in the mid 90's, the competitor everyone wanted to beat. Yet ATI survived as #2 from the graphics shakeout until the Radeon 9700.
Competition is not going to destroy NVIDIA. Ephemeral market dominance is not going to destroy NVIDIA. As long as NVIDIA turns out a competitive product in all their markets, they will be a player in all their markets.
For NVIDIA to be destroyed, the company has to maintain last year's corporate incompetence for a significant duration. Shareholders and the board simply will not permit such an ongoing failure without a management overhaul. Even founder/CEOs can be dethroned if they're demonstrably non-profitable.
nonamer
17-Oct-2003, 03:35
For NVIDIA to be destroyed, the company has to maintain last year's corporate incompetence for a significant duration. Shareholders and the board simply will not permit such an ongoing failure without a management overhaul. Even founder/CEOs can be dethroned if they're demonstrably non-profitable.
No offense but last year was when the GF4 came out. They would last quiet a while with last year's "imcompetence."
ricercar
17-Oct-2003, 04:57
No offense but last year was when the GF4 came out. They would last quiet a while with last year's "imcompetence."
No offense taken. However, that turns out not to be the case. GeForce4 NV25/NV17 "came out" in February 2002, considerably more than 1 year ago.
Over one year ago, September 2002, was the release of GeForce4 Ti AGP8x (NV28). NV28 has never been a performance peer for the ATI Radeon 9700 Pro, released August 2002. NVIDIA has been playing catch up since August 2002, more than one year.
In the last year, since October 2002, NVIDIA design incompetence and TSMC fabrication incompetence kept the NV30 from its well-announced and repetitively missed release dates. NVIDIA marketing incompetence kept NVIDIA from admitting to the delay, and instead pretend that nothing was wrong. Qualitative credibility and quantitative sales were lost due to NVIDIA design and marketing incompetence demonstrated in the year since October 2002.
I stand by my statement. For NVIDIA to be destroyed, the company has to maintain the last year's corporate incompetence for a significant duration.
Holy cow, off topic we go (not that I am complaining). I could care less what everyones' opinion is. I just would like to see some hard dollar facts on if the NV30 lost money and how much. I realize that the R&D money gets spread out over all the nv3x series. I read it was 400 million or so. And they shipped about 100,000 FX5800's. So the question is how much of the 400 million went into the nv30. I can't see the nv30 ever making money.
ricercar
17-Oct-2003, 06:37
how much of the 400 million went into the nv30.
If you discount any NV30 R&D applying to the NV31/34/35/36/38/4x, then NV30 GPUs would have had to sell in the neighborhood of an average of $4000 each to break even.
digitalwanderer
17-Oct-2003, 14:32
I stand by my statement. For NVIDIA to be destroyed, the company has to maintain the last year's corporate incompetence for a significant duration.
What would you consider a "signifigant duration"? I'm not disagreeing with you, I'm really curious. I'm feeling that with a year+ of bad behaviour under their belt they've already got their work cut-out for themselves to re-establish their credibility.
Holy cow, off topic we go (not that I am complaining). I could care less what everyones' opinion is. I just would like to see some hard dollar facts on if the NV30 lost money and how much. I realize that the R&D money gets spread out over all the nv3x series. I read it was 400 million or so. And they shipped about 100,000 FX5800's. So the question is how much of the 400 million went into the nv30. I can't see the nv30 ever making money.
Problem with that is the 400 mil that went into research for the NV30 is also paying a dividend on every NV3x board they make, the research applies to all of 'em so nVidia can spread out it's loss.
It's damned hard to nail down R&D money with precision on a good day, with nVidia intentionally trying to muddy the issue it's pretty hopeless. :(
EDITED BITS: "for the" instead of "fort he"... :roll:
RussSchultz
17-Oct-2003, 15:05
It's damned hard to nail down R&D money with precision on a good day, with nVidia intentionally trying to muddy the issue it's pretty hopeless. :(
WTF? They share with you the cost in R&D for their new architecture and you're accusing them of intentionally muddying the waters because you can't calculate with ease how profitable a particular product line within that new architecture is?
Never mind that we have almost 0 data on these topics from any other IHV with regard to their own architectures or products.
digitalwanderer
17-Oct-2003, 19:02
It's damned hard to nail down R&D money with precision on a good day, with nVidia intentionally trying to muddy the issue it's pretty hopeless. :(
WTF? They share with you the cost in R&D for their new architecture and you're accusing them of intentionally muddying the waters because you can't calculate with ease how profitable a particular product line within that new architecture is?
I was referring more to how nVidia is trying to keep exact sales figures of the NV30 under wraps. :roll:
RussSchultz
17-Oct-2003, 19:11
So, what are the exact sales figures of the 9500?
Where's the intentional obsfucation accusations there?
I was referring more to how nVidia is trying to keep exact sales figures of the NV30 under wraps. :roll:
So is everyone else in this business.
I stand by my statement. For NVIDIA to be destroyed, the company has to maintain the last year's corporate incompetence for a significant duration.
I firmly agree with you. Although I'd see a few more factors to it.
Plus, I disagree on the Quadro being the clear professional leader: it badly lacks behind in shading performance against ATI, but that doesn't matter a lot for *most* of that market, and NV got the flexibility lead.
So you're right, but I think that small detail is rather important, because it shows just how fast nV could lose the market if they lose the flexibility lead, and the market orients itself more towards shaders.
---
In my eyes, NVIDIA got one BIG project int he GPU market: The NV50. It's the card they're betting a lot on, probably expecting it and its refreshes to last a good 24 months. It's a LOT of R&D going in it, and the design being planned upon a lot longer than all other nV designs.
The NV40 can't kill NVIDIA. Anyone pretending that simply doesn't have a clue about business. But should the NV40 not be a good success or a failure, and should the NV50 be a failure, their core market is, IMO, toast. Their reputation would have gone the way of the dodo by then, so only their secondary markets could bring them any profit.
So you've got nForce and Transmeta. In nthe next 5 years, the mobile GPU market as done by Transmeta will require more complex engineering, so if they lose their core GPU market, they'll evantually lose a big part of that one too. But that's not in a sufficently short term for it to matter much, IMO, so they do have an important win in that direction.
Looking at the nForce3 though, their engineers better realize their mistakes, because right now, I know I'm not too tempted by one. And if the Soundstorm rumors are true, they lost a key niche market too. Not that it matters much, but at this rate, their solutions might focus on value instead of quality. Not a bad business solution, it all depends on how well it is executed, of course.
---
So, for NVIDIA to go bye-bye, they'd have to:
1. Not have a big success with the NV40 *and* have a failure or not a particularly good success with the NV50.
2. Their corporate techniques have to remain the same ( although if everything else happened, and not this, they might still be in trouble - the other way around is true too, but to a lesser extent IMO ).
3. They've got to slowly but surely lose their nForce market - losing it fully is not required, but becoming a smaller player would be sufficent to be in trouble overall if the rest is true.
NVIDIA becoming a smaller player, should their corporate incompetence as you say persist, is however a much more likely scenario IMO: Never becoming the new Intel they've always dreamed of, just doing alright at $2B market capitalization.
---
Regarding the NV30: $3670 per CHIP not to lose money on it alone, to be exact ;) And if the board sold at $400, it's likely NV sold the chip at 50-75 bucks. Maybe 250-500 for Quadros, which are half their production.
Fact is, spread out the costs whatever way you want, some of them you just can't honestly say they contributed to making the derivatives.
Someone please explain me how all the failed tape-outs and the dozens of millions of dollars lost on failed accelerated risk production runs would contribute to future chips? :lol:
---
Once again, I agree with you ricercar, and I hope you'll also agree with me. I'm sure I'll know soon enough :) ;)
Uttar
EDIT: Changed "sure" by "soon", last sentence made absolutely no sense.
digitalwanderer
17-Oct-2003, 21:27
So, what are the exact sales figures of the 9500?
I think it was somewhere between "lots" and "god-awful lots!", depending on if you include the Pros or not. ;)
Where's the intentional obsfucation accusations there?
No one asked about ATi. :roll:
I'm not accusing nVidia of anything unsavory here, I'm just pointing out that it's very hard to get accurate information about these things. :)
RussSchultz
17-Oct-2003, 21:45
Of course nobody asked about ATI. That would be unwarranted.
Of course nobody asked about ATI. That would be unwarranted.
Indeed.
I'd like some numbers for XGI, Matrox and S3 though ;) j/k
Digitalwanderer, check your PM - trust it or not, the reliability on that number is very high.
Uttar
cthellis42
17-Oct-2003, 23:26
Why are you only sharing those numbers with dig? Many of the rest of us like numbers, too~! <snf>
Why are you only sharing those numbers with dig? Many of the rest of us like numbers, too~! <snf>
They're in the editorial too, plus, I just give him the exact figure for the NV30 R&D costs, not a lot of figures :P
I just thought of giving that number first to Dig because he seemed to believe these are nearly impossible to get...
Most of the info in the NV30 paragraph of the editorial do not come from my sources, but rather from someone also working on such matters, and he has a source in accounting. I'll always be very impressed by his sources too, as they are really much more diverse than mine.
Uttar
bloodbob
18-Oct-2003, 12:10
Why are you only sharing those numbers with dig? Many of the rest of us like numbers, too~! <snf>
They're in the editorial too, plus, I just give him the exact figure for the NV30 R&D costs, not a lot of figures :P
I just thought of giving that number first to Dig because he seemed to believe these are nearly impossible to get...
Most of the info in the NV30 paragraph of the editorial do not come from my sources, but rather from someone also working on such matters, and he has a source in accounting. I'll always be very impressed by his sources too, as they are really much more diverse than mine.
Uttar
Then maybe you should tell him to get his ass into this forum now :)
Then maybe you should tell him to get his ass into this forum now :)
His job isn't to leak confidential information, he's a reviewer. And considering how hardcore he is about reviewing, huge shout-outs that never even Anandtech or Tomshardware would dare to publish because it'd kill their bandwidth.
Smart man, too. Although rumors ain't his speciality I'd say - he just speculates WAY too much, so even though he got amazing sources, the info coming out of his mouth is only midly reliable.
Uttar
I heard a rumor that the dig might have info on some of the new graphics cards coming out.... firsthand. I wonder if that is what gave him a heart attack. If this is true dig, don't hate me. :twisted:
digitalwanderer
19-Oct-2003, 00:07
I heard a rumor that the dig might have info on some of the new graphics cards coming out.... firsthand. I wonder if that is what gave him a heart attack. If this is true dig, don't hate me. :twisted:
I could never hate ya ByteMe, never. :)
Nope, it ain't true. Chasing my kids on their bikes is probably a much bigger contributing factor than anything computer related. (I really don't get anywhere near as worked up about computer stuff as I appear to online. ;) )
Did he just dodge the question?
digitalwanderer
19-Oct-2003, 01:28
Did he just dodge the question?
No. No I do NOT have any new info on any cards. I got an AIW 9600 Pro on Wenesday to review and I'm still waiting to hear about when I should be getting an XGI to take for a test spin.
That's all I know, sorry.
just me
19-Oct-2003, 03:21
IBM will take over the GPU side of nV, it's just a matter of time, IMO. :wink:
Now on to the important topic >
You had a heart attack Dig?! :shock: Damn man you're younger than I am. What the h3ll happened & how are you my friend? :cry: Man oh man, gotta take care of yourself for them kids & the wife! :evil:
IBM will take over the GPU side of nV, it's just a matter of time, IMO. :wink:
l:
I never thought about THAT. What makes you take this wild speculation?
bloodbob
19-Oct-2003, 04:06
Well currently IBM are doing most of amd's R&D so wouldn't surprise me.
just me
19-Oct-2003, 07:36
IBM will take over the GPU side of nV, it's just a matter of time, IMO. :wink:
l:
I never thought about THAT. What makes you take this wild speculation?
Who is making the nV cores in Fishkill, NY? IBM. If they don't sell because of 'issues', ya think nV will pay IBM? I don't & IBM will show nV what Lawyers can do. :wink:
IBM has a history of getting what they want & seldom do they enter into a venture w/o thoughts of 'take over'. It is the nature of Big Blue. :twisted:
Look for an AMD/IBM chipset to take out NF3's next year. They've helped Intel, AMD is 'in house' & so is nV. They all end up going to IBM for their expertise & IBM has 'em by the [ahem] . :shock:
IBM has their finger in a lot of pies & the companies that 'lease' IBM technology (tech IBM develops only to 'lease' to others) would astound you.
You have seen the 144mb Cache CPU IBM is making, haven't you?
http://www.theinquirer.net/images/articles/bluey.jpg
.02 & time will tell
vBulletin® v3.8.6, Copyright ©2000-2013, Jelsoft Enterprises Ltd.