NVidia management

[caywen]

It's going to be really interesting to see how NVidia management gets out of this stinker. On one hand, they made many of the same mistakes that 3dfx's bumbling management made during the Voodoo5 fiasco. On the other hand, Huang is a realistic and smart guy and is likely to take dramatic actions to correct this path (if the Board of Directors lets him, that is).

The best thing for NVidia to do at this point is to recognize that they have indeed lost this battle and focus their attention on the market that actually pays for all this amazing chip design: the budget category.

The truth is, NVidia doesn't really need to be the performance king right now. They need to produce a budget card that beats the 9500 on price/performance while leveraging its continuing good relations with OEMs and and computer makers. If they can keep their cash flow positive, then they can capture the performance sector again within 2 years. But until then, having to maintain the performance crown costs more than it's worth.

This is especially true given that the performance range on systems today has widened immensely and game developers are struggling to move beyond the capabilities of the GeForce 3/8500. It's still the "MX" budget market that is generating the game sales.

People have noted that their NV30 fiasco has many similarities to 3dfx's VSA100 fiasco. This is true, but it also has many similarities to their GeForce 1 fiasco. GeForce 1 was a company launching a new chip whose theoretical capabilities couldn't be fully realized by its real-world performance. It was not a big success and it did not take 3dfx long to trump it. However, it represented a great platform on which NVidia could build 3 generations of successors that would generate huge revenue for them. I fully expect GeForce FX to be the same. I fully expect the technology ramp FX->FX2->FX3->FX4 to stay extremely competitive with ATI both on performance and profitability.

My 2 Taiwan NT dollars.

 

[Fuz]

If they can keep their cash flow positive, then they can capture the performance sector again within 2 years.

Actually, the NV35 should be able to do that, convincingly. Of course all based from rumours at the moment, though the rumours do come from credible sources.

 

[John Reynolds]

It's going to be really interesting to see how NVidia management gets out of this stinker.

What stinker? You mean the stinker of actually growing your market share in spite of the fact that a competitor has a faster product than you do in basically every segment? The stinker of being in the black while your competitor is operating in the red?

 

[duncan36]

Like the Rampage was 'just' around the corner to save 3dfx too. Nvidia has some severe internal engineering/managerial problems that are responsible for their recent debacles. I think the people who are saying 'oh well in just a few more months they'll do it' aren't recognizing the severity of this.
An entire product line that was 6 months overdue is basically ruined, Nvidia engineers were soundly beaten by ATi engineers.
Nvidia has always relied on process and memory improvements for their products, this is the first time their creativitiy has been put to the test, needless to say they failed miserably. I just dont see why everyone assumes that a shaken engineering team thats failed miserably will suddenly come to life and produce a world beater.
Look at 3dfx once they were in a tailspin, they went splat.
Thats the more likely scenario when you're in a tailspin you go splat, and Nvidia definitely is in one.

 

[duncan36]

Slamming Nvidia to the mat isnt like kicking over a can, Nvidia is a weighty, sweaty, SOB. But with this news and the R350 coming out ATi's got got the momentum and the heavy Nvidia will be slammed to the mat.
Theres nothing magical about Nvidia, just like ATi lost OEM contracts to Nvidia's superior product in the Rage days. Nvidia will lose to ATi's superior product.

What stinker? You mean the stinker of actually growing your market share in spite of the fact that a competitor has a faster product than you do in basically every segment? The stinker of being in the black while your competitor is operating in the red?

 

[WaltC]

It's going to be really interesting to see how NVidia management gets out of this stinker. On one hand, they made many of the same mistakes that 3dfx's bumbling management made during the Voodoo5 fiasco. On the other hand, Huang is a realistic and smart guy and is likely to take dramatic actions to correct this path (if the Board of Directors lets him, that is).

Good sentiments--just realize it was Huang who got them to GF FX Ultra, though. I don't think he was realistic *at all* with the GF FX Ultra concept.

The best thing for NVidia to do at this point is to recognize that they have indeed lost this battle...

Agreed. nVidia has already done this, as Huang as publicly called the R300 a "wonderful" chip and admitted that ATi had taken the performance crown from them. (I posted a link to the Mercury News story in the "canned" thread.) I do think that it was necessary for nVidia to admit it was now second fiddle before there would be any possibility of the company coming up with competitive 3D chips. You have to be able to see what you need to do before there's any chance of doing it. Now that the blinders are off, they at least have a chance.

The truth is, NVidia doesn't really need to be the performance king right now. They need to produce a budget card that beats the 9500 on price/performance while leveraging its continuing good relations with OEMs and and computer makers. If they can keep their cash flow positive, then they can capture the performance sector again within 2 years. But until then, having to maintain the performance crown costs more than it's worth.

Not true in the slightest. I have seen this particular apology for nVidia posted several times and it is definitely not true. Companies do not lose money on the high end, they make money hand over fist on the high end, which is why they try so hard to compete on that level--even much smaller companies like Matrox shoot for it, too. Not only is the high end 3x-5x more profitable *per unit sold* than the low-middle end, the high end captures the mindshare of the market which results in many additional low-value segment sales. At a recent conference an ATI executive was asked to describe the income split in his company between the high-end and the low-middle end, and he respond that the split was "about 50-50." People tend to think that because the volume is so much higher at the low-middle end, so is the profit. Nope, they have to sell 3-5x as many low-middle end products to equal the profit they make on their high end sales. Holding the high end performance crown is literally worth several tens of millions of dollars a year to the company holding it (provided they can meet demand for their products, etc.)

This is especially true given that the performance range on systems today has widened immensely and game developers are struggling to move beyond the capabilities of the GeForce 3/8500. It's still the "MX" budget market that is generating the game sales.

Nope...this is nothing new. The general developer market for 3D games has always written to the lowest common denominator, in most cases. That's been the norm for the last 15 years in computer gaming--even before 3D games eclipsed 2D games in sales. Even companies like ID Software which do their best to hit the "cutting edge" of 3D hardware when it ships inevitably lower their sights a couple of notches in some ways so that they can grab more sales. Game developers aren't "struggling" *chuckle* they deliberately target a lower common denominator many times simply for the sake of greater sales of their products.

However, that doesn't mean that titles like NWN don't look grand on a 9700 Pro using more modern code--they do. And that won't change the fact that software like DOOM III will actually serve to drive high-end hardware sales--Carmack's software has a history of doing just that. The $400 high-end 3D card is not dead--not by a long shot. As long as companies can come up with compelling reasons for people to spend that kind of money on 3D cards--in other words, as long as companies can make terrific 3D chips with a lot of power and longevity in them, this market will continue to flourish.

What I'm saying is that comforting yourself with the idea that nVidia "doesn't need" to be number 1 in performance to hold its ground is a delusion--to the victor go the spoils--just watch and see. Watch how hard nVidia tries to retake the crown from ATI as a measure of how important they deem that crown to be. OTOH, watch how hard ATI works to keep that crown, for the same reason. Being considered number 1 in 3D performance is considered *vital* by both of these companies, yet only one at a time can occupy this position. May the best team win! As consumers we benefit regardless of who fills the slot, and we benefit a lot more from the competition than we would from a single company holding that position with no competition in the works.

People have noted that their NV30 fiasco has many similarities to 3dfx's VSA100 fiasco. This is true, but it also has many similarities to their GeForce 1 fiasco. GeForce 1 was a company launching a new chip whose theoretical capabilities couldn't be fully realized by its real-world performance. It was not a big success and it did not take 3dfx long to trump it. However, it represented a great platform on which NVidia could build 3 generations of successors that would generate huge revenue for them. I fully expect GeForce FX to be the same. I fully expect the technology ramp FX->FX2->FX3->FX4 to stay extremely competitive with ATI both on performance and profitability.

My 2 Taiwan NT dollars.

GeForce 1 mainly sucked at first release because it was an SDR card and it took nVidia about a year to develop decent drivers for it, and it could barely keep pace with a TNT2 at first (BTW, the architecture of a GF1 is basically two TNT2s merged--that's almost exactly what it is, so the archtecture that has served nVidia so well all of these years has literally been the TNT, which nVidia has greatly enlarged and expanded and built on.) The nv30 is a *new architecture*--it's not a "new architecture based on the TNT/GF core"--it's an entirely new architecture, period. As such, it can either be a jewel or a dog, and that's true for any company, regardless of type of chip, when they move to a new architecture.

The problem with nv30 is not that it is inherently a bad chip, the problem is that ATI's new architecture chip is much, much better--so much better, in fact, that the R300 clocked 175MHz slower and at .15 microns, outperformas it in many cases. nv30 is not an R300-killer, not even close. It probably would have been an OK chip had ATI never produced R300, but that's as far as you can take it. The architecture you should be watching with great interest is R300--I think nVidia's going to have to take nv30 back to the drawing board for some major surgery before it will be competitive with it, IMO.

 

[mcshaner1]

I think the best thing that nVidia can do at this point is to focus their attention on the professional market.

http://www.xbitlabs.com/video/3dsmax5-quadrofx/

The QuadroFX seems to do pretty well with 3ds max.

 

[KimB]

No, the professional market, while lucrative, is very small compared to the consumer sector. nVidia makes most of their money in the low-end, selling TNT2 M64's and various GeForce MX chips to OEM's for putting in PC's people buy. Very, very few people actually like to upgrade their PC. Most just purchase a new one.

 

[demalion]

Hmm...

I think nVidia engineers recognized the comparison of nv30 versus R300 long ago. It is simply that marketing comments didn't.
Huang's comments and marketing decisions reflect...marketing, but unless he acted the fool, his other decisions would have reflected the evaluation of the engineers.
I think that admission concerning performance leadership deviates from reflecting marketing only as much as necessary to convey to investors that he hasn't acted the fool...

I maintain it is a fallacy to view the nv35 design decisions as something that just occurred, or that admitting the loss of performance leadership in public now has anything to do with being able to address the nv30's deficiencies in the nv35. Note where that comment begins and ends, and the issues with nv35 launch it does not seek to address.

Also, I think the nv30 is BOTH based on the GeForce 4 design (fixed function and some of the "shader" functionality) AND a new architecture (the advanced shaders capabilities). No period emphasized....
I also persist in my opinion that dismissing the magnitude of the impact of the 128-bit bus is a folly, and basing expected architecture performance on such a dismissal leads to a very flawed analysis.

At the same time, I think there are all sorts of positive comments about the R300 in comparison to the nv30 possible even ignoring the 256-bit bus advantage, and all sorts of opportunities for the R350 to warrant even more such comments.

 

[demalion]

No, the professional market, while lucrative, is very small compared to the consumer sector. nVidia makes most of their money in the low-end, selling TNT2 M64's and various GeForce MX chips to OEM's for putting in PC's people buy. Very, very few people actually like to upgrade their PC. Most just purchase a new one.

I don't think the professional market can be ignored. 3dlabs seems to fund some decent engineering work by catering to that market...but perhaps they get income from patents as well? <- I have no idea how many patents they hold and how much income could be garnered from holding them, just theorizing.

Hmm...why no 's for other directions? This site is displaying an obvious bias for the "right"!

 

[duncan36]

nVidia makes most of their money in the low-end, selling TNT2 M64's and various GeForce MX chips to OEM's for putting in PC's people buy. Very, very few people actually like to upgrade their PC. Most just purchase a new one.

While true you're omitting/ignoring that the only reason that Nvidia has all these big contracts to sell cheaper 3d devices is because they ruled the high end.
It works out like this, Nvidia comes to big OEM and says look we can get you a million Gf4Mx's for $10 each with $2 profit on each for us. Dell says great ATi can only sell the 9000 for $15 each what a bargain we'll go with Nvidia.
Now with their high end ownage ATi can walk in and say well we can give you the 9700pro for $150 which should make you $150 per 9700 sold. Nvidia cant match this product the 9700pro is your only choice.
Oh btw if we offer you these terms we'll expect you to start buying the 9000 for $15.
Thats how it works, Nvidia is going to start losing big contracts its inevitable.

 

[Tahir2]

It works out like this, Nvidia comes to big OEM and says look we can get you a million Gf4Mx's for $10 each with $2 profit on each for us. Dell says great ATi can only sell the 9000 for $15 each what a bargain we'll go with Nvidia.
Now with their high end ownage ATi can walk in and say well we can give you the 9700pro for $150 which should make you $150 per 9700 sold. Nvidia cant match this product the 9700pro is your only choice.
Oh btw if we offer you these terms we'll expect you to start buying the 9000 for $15.

Not from what I have seen.. the bottomline is the bottomline and a million GF4MX's will sell a million systems which is better than 10'000 systems with a highend card for definite. (NB: I only have knowledge of the UK market).

 

[duncan36]

Bottomline is the most important thing in a top-down product line. When there are holes in the product line thats when theres problems.
Classic case is ATi around 2000, their product gaps left the door wide open for Nvidia who promptly clobbered them.

 

[caywen]

It's going to be really interesting to see how NVidia management gets out of this stinker.

What stinker? You mean the stinker of actually growing your market share in spite of the fact that a competitor has a faster product than you do in basically every segment? The stinker of being in the black while your competitor is operating in the red?

The stinker of being perceived as the market loser (as opposed to leader). It's a position they'd rather not be in. Before, they had money and the performance crown. Now they've lost the latter (if the rumors of the 5800's demise are, in fact, true).

As far as how they are financially, that's a completely different matter, as I've pointed out already. Being the performance king and technology leader doesn't equate to being a good business.

 

[John Reynolds]

As far as how they are financially, that's a completely different matter, as I've pointed out already. Being the performance king and technology leader doesn't equate to being a good business.

In the end, that's all that matters to a company: "how they are financially".

 

[caywen]

Not true in the slightest. I have seen this particular apology for nVidia posted several times and it is definitely not true. Companies do not lose money on the high end, they make money hand over fist on the high end, which is why they try so hard to compete on that level--even much smaller companies like Matrox shoot for it, too. Not only is the high end 3x-5x more profitable *per unit sold* than the low-middle end, the high end captures the mindshare of the market which results in many additional low-value segment sales. At a recent conference an ATI executive was asked to describe the income split in his company between the high-end and the low-middle end, and he respond that the split was "about 50-50." People tend to think that because the volume is so much higher at the low-middle end, so is the profit. Nope, they have to sell 3-5x as many low-middle end products to equal the profit they make on their high end sales. Holding the high end performance crown is literally worth several tens of millions of dollars a year to the company holding it (provided they can meet demand for their products, etc.)

Good points, overall, except that I'm not apologizing for them. Normally, I'd agree with you except that NVidia has spent something in the order of half a billion on developing NV30. If they were leading the performance pack by a considerable margin then they would be able to count on healthy sales to recoup the costs. Given the rumors (and that's all they are at the moment) that only 100,000 NV30 chips will be produced, they would have to recoup up to $4000 per chip to break even. Realistically, only a fraction of that money was actually intended for the 5800 marketing, but the recoup costs are still very high (I'd estimate between $500-1000 per chip sold). So, obviously trying to keep the performance crown for the 5800/5800 Ultra is a deeply red situation for NVidia. Fortunately, I think they'll be able to make it all back from subsequent products that capitalize on the NV30's IP (ie NV31, mobile versions, etc).

Nope...this is nothing new. The general developer market for 3D games has always written to the lowest common denominator, in most cases. That's been the norm for the last 15 years in computer gaming--even before 3D games eclipsed 2D games in sales. Even companies like ID Software which do their best to hit the "cutting edge" of 3D hardware when it ships inevitably lower their sights a couple of notches in some ways so that they can grab more sales. Game developers aren't "struggling" *chuckle* they deliberately target a lower common denominator many times simply for the sake of greater sales of their products.

It's nothing new, but it's true. The lowest common denominator hasn't moved ahead as fast as the high end. Furthermore, fewer people are upgrading their old PC's as consumer spending on technology has slowed down dramatically in recent years. Furthermore, there's been a glut of low-priced PC's with integrated mobo chips (very slow). Thus, you've got a handful of blisteringly fast high end cards (R300, NV30) and lots of slugs (Integrated mobo 3D). The performance range is getting wider from a perspective of both fill rate, triangle rate, and features.

What I'm saying is that comforting yourself with the idea that nVidia "doesn't need" to be number 1 in performance to hold its ground is a delusion--to the victor go the spoils--just watch and see.

With all due respect, I sense you think that I'm an NVidia apologist and that I have some kind of agenda. I don't.

I neither own their equipment (i've got a cheap Rad 7500) nor their stock. Nor do I work for them in any way. They are in a serious pickle and if they make the same mistakes that 3dfx made, they are in huge trouble. But again, Huang's a tenacious and smart CEO and I don't think he's going to let this happen.

 

[Sxotty]

It's going to be really interesting to see how NVidia management gets out of this stinker.

What stinker? You mean the stinker of actually growing your market share in spite of the fact that a competitor has a faster product than you do in basically every segment? The stinker of being in the black while your competitor is operating in the red?

I think it is going to show up this summer. The market lags behind, I know that I waited to the GeforceFX, and now I decided it is not worth the price tag. Other people waited as well, the longer ATI rules the surer Nvidia will start to loose market share. Even in a much more unbalanced battle between Intel and AMD, when AMD started to walk all over Intel after a bit they started to gain market share as well.

 

[Hellbinder]

What stinker? You mean the stinker of actually growing your market share in spite of the fact that a competitor has a faster product than you do in basically every segment? The stinker of being in the black while your competitor is operating in the red?

I would like to remind you that up until now the entire world thought that the GFFX Ultra was going to be far superior to the 9700. Coupled with the fact that the 9500's were not out for an entire sales quarter. Its really not that suppriseing that Nvidia sold more low end GF4's Etc toward the end of last year.

You guys seem to think that These things change over night. They dont. The bottom line is that Nvidia has now been delt a major PR blow, and now has to wait another 5 months before they can counter. Meanwhile Ati is about to lauch another new product, further driving the prices down on there current DX9 line.

Nvidia has no chance whatseover to right the ship until sometime in 2004. Neither their Nv31 nor Nv34 are going to be performance, or price performacne competative with the Spring ATi line up. Nvidia now is stuck in the unpleasant position to design their first 256bit MC, tweak the Nv35 design up as close as they think they can based on what they are *guessing* about the R400, and somehow Rush it out the door 5 months ahead of schedule.

I seriously doubt that they mantain the same marketshare they have now over the next 12 months. Im not saying they fall apart. But I would look for a change towards ATi.

 

[Nagorak]

The best thing for NVidia to do at this point is to recognize that they have indeed lost this battle and focus their attention on the market that actually pays for all this amazing chip design: the budget category.

So...basically they'll be like ATI has been for the last 3 years?

What stinker? You mean the stinker of actually growing your market share in spite of the fact that a competitor has a faster product than you do in basically every segment? The stinker of being in the black while your competitor is operating in the red?

Do you have to bring this up every time someone criticizes Nvidia?

Seriously, I thought record players went out back in the 70s...but this keeps going round, and round...

 

[Mulciber]

]

What stinker? You mean the stinker of actually growing your market share in spite of the fact that a competitor has a faster product than you do in basically every segment? The stinker of being in the black while your competitor is operating in the red?

I seriously doubt that they mantain the same marketshare they have now over the next 12 months. Im not saying they fall apart. But I would look for a change towards ATi.

I'll actually be looking for a change in market share towards ATI and Intel. But I still expect growth for nVidia this year in the mobile and chipset segments. ATI should deffinately be gaining marketshare in the retail section.