NVidia Boldly Lies at its on Conference

NewsFactory

Warp2Search have passed us a link to a short but interesting story from Nvidia's latest press/industry briefing on their business...

GameSpot

Al_Vampyre

Smoke & Mirrors....nuff said!

Vengeance

True......

wertret

WTF! Nvidia stills thinking people is idiot...

MSX

I think the worst part of this whole thing is their Prez. He knew before the FX line came out that it was nothing spectacular and he was excited for the NV40. I'm guessing the NV3X came out and they decided to sick their PR hounds on it.

Time will tell. Yearly updates are not what enthusiasts want... Its why people are so pissy about the current situation with nforce.

WaltC

I want to kind of apologize for the length of this post as it wasn't intended. I found the Gamespot report of the nVidia Editor's Conference to be an exceptionally interesting read on a number of levels. As such, this post is divided into two parts. Part One consists of my initial, knee-jerk reaction to the statements by nVidia as Gamespot reported they occurred at the Conference. When I made these comments it never crossed my mind that there might be an alternative background interpretation in which the Conference remarks as related by Gamespot might be framed that would actually allow them to sound sensible, rational, and logical from nVidia's point of view. But after reading my initial remarks over (Part One) it just seemed to me that it all seemed so weird and bizarre because I was reading into the Gamespot report nVidia's intention to remain a strong competitor in the 3d-gaming chip & reference-design markets. Part Two emerged when I realized there was an interesting, alternate view of the context of the Conference in which everything nVidia said there would make perfect sense.


Part One


That was a disappointing read. That "content management team," stuff is really bad news--what "content" is nVidia going to try and "manage"? Of course, the "content" they are targeting is nothing less than the games produced by the companies they mentioned. It's clear that nVidia is still dead set on forcing the industry to capitulate to its standards and goals, as opposed to acknowledging that the 3d gaming industry itself is much larger than nVidia, and that nVidia is only a *part* of that industry. These guys have a severe case of delusionitus.

I suppose that this kind of thing is very good news for ATi, whom I am sure will be more than content to stand back and watch nVidia drive itself into the ground in the 3d chip marketplace. But I am nonetheless disappointed in nVidia--profoundly. Why is it so *hard* for them to admit that ATi has indeed one-two-three-upped them for the last year? Truthfully, of the many people I know few are using nVidia products these days and of those who are most are disappointed with nVidia's current products, drivers, and practices in relation to the 3d-gaming industry. I would much prefer this was not the case and that nVidia would take a page from ATi's book and emulate what ATi did after being perceptually and factually behind nVidia for years in terms of 3d hardware performance and driver support. ATi remade itself from top to bottom and decided to compete with nVidia and beat them at their own game, if possible.

I never recall at any time listening to ATi support its second-string position by stating over and over again that basically ATi was right and the rest of the world was wrong. Nope, ATi turned the corner because unlike nVidia it had few delusions about itself in the 3d-chip markets, or its influence, and realized that the way to win was simply to build a better mousetrap. In order to change dramatically, the first thing a company must understand is that its present course isn't getting it anywhere. Before ATi could improve such as it has done, it was first necessary to be brutally honest with itself and identify all of the areas in which it was failing relative to its competition. Without that kind of acute self-understanding and flaw diagnostics the major changes we've all seen this year at ATi would never have occurred. Companies are run by people and don't run themselves. So it's not surprising that the approaches of ATi and nVidia to the same markets for the last year have been dramatically different--the people running these companies are dramatically different, after all. I am still very disappointed in nVidia, however.

So why is that too much to ask of nVidia--that they simply re-enter the real world and start competing once again? Reading this article made me almost feel sorry for the top management at nVidia, as they seem totally lost and without a clue as to what they need to do to compete. If ATi had been that way, and ATi's top management had professed the general attitudes I've seen quoted from the nVidia CEO in the last year--there never would have been an R3x0--and the last year would never have happened for ATi.

Over and over again, ceaselessly and repetitively, all year long, the message from nVidia has been that only nVidia knows what is best, and that Microsoft, game developers, benchmark developers, and even their successful competitors like ATi--are all wrong. In short, the view from the top of nVidia is that among everyone else only nVidia is right. That's just so sad it's almost pathetic. I really, really never expected to see nVidia sink this low and display this kind of irrationality. I thought--and hoped--that nVidia would "catch on" to what it needed to do months and months ago. But after reading this report at Gamespot I no longer think the corporation management has the needed judgement to be able to do that.

That's where the "content management team," comes in. To me, even the nomenclature used here denotes an unsavory and ill-advised use of nVidia's resources. They are looking at their customer base through the particularly distorted lens of game developers, and imagining that all the people who buy their 3d-chips as end-users through nVidia's board OEMs are essentially idiots not worth the time and trouble to directly talk to without condescending. Top nVidia brass obviously feels that "all" it needs to do is co-opt game developers into doing certain things with their game engines, and the stupid customer base will fall into line and accept whatever status-quo nVidia decides to dish out. So what this "team" is seeking to do is nothing short of the usurpation of software game publishers and developers simply to make sure that their "game content" is very friendly to whatever hardware nVidia decides to make. I don't think this approach will work for nVidia.

Rather, they should again take a cue from ATi and seek to entice developer support of their hardware by making darn good hardware in the first place! If your hardware and driver picture are where they should be then the interest of game developers will automatically follow. nVidia seems congentially unable to understand that its goals are not the same as the goals of software developers. I think they've become very confused about that in the years since 3dfx and pre-R300, and have mistaken developer interest in their products during that time as stemming from some source other than the hardware and drivers nVidia was putting out then in relation to what the other 3d IHVs were offering. I find it very sad that nVidia is retreating from engineering into the delusional fixation that they can accomplish all of their goals through the marketing efforts of "content management teams" (when the "content" to be "managed" does not even belong to nVidia in the first place) regardless of how their products compare to those made by their competitors in reality.

The comment they made about their products currently losing the benchmark wars only because nVidia wasn't cheating anymore is so incredibly bizarre that it has to stand on its own. Nothing anyone could say about such a strange delusion could top nVidia's direct comments here as reported by Gamespot.

Next up was the equally strange statement that nVidia wanted everyone to know that if it was losing market share it certainly wasn't to ATi, but it was to Intel. Just one utterly self-evident problem with that proposition: the only graphics chips Intel makes these days are 2d value chips, usually motherboard integrated, targeted at the very low end of the corporate markets and including only partial DX7/8-ish (not really sure if DX8 support is even involved, unless it's done in software through the drivers.) In fact, it's nVidia's 440MX el-cheapo chips that nVidia and intel are competing on, as Intel doesn't even make a DX9-supporting chip, as Ati makes in R3x0 and nVidia claims to make in nV3x.

This leads into the most bizarre statement in the whole Gamespot account--and possibly the most telling statement of all--that nVidia is giving people heads up that even FOUR quarterly driver releases for their products annually "might be" unattainable from now on--and that the company might seriously reduce even that low number to ONE DRIVER PER YEAR.

Everyone should realize that for mid-range to high-end 3d graphics reference designs made primarily for 3d gaming, such as those currently made by ATi and nVidia--ONE DRIVER RELEASE A YEAR is simply not feasible. It's just not possible, for the very practical reason that in order to stimulate sales of your products on an ongoing basis you need to address BUG FIXES in your drivers (if nothing else) so that the 3d games people buy those refrence design products *for* will actually be able to run those games with a high degree of reliability and stability--not even talking about performance at all here. Regular driver updates is part & parcel of the business and there's just no getting around that. So, what could nVidia's top management be talking about here?

(Continued in Part Two)

WaltC

Part Two



If we take all of these pronouncements as a whole we *might* be able to infer a future implied course of action by nVidia that few people, including me, might ever have suspected or guessed--at least at this particular time.

(1) What nVidia's top brass is actually saying, I think, relative to its "content management team," is actually not that "We're right and everybody else is wrong," as I interpreted above. What they are actually saying (without literally saying it) is that, "We don't care if we're wrong relative to what the 3d-gaming industry is moving toward in reference to API standards--we don't care if we are wrong, because shortly that market will no longer be of any major concern to our company."

If the remarks are viewed in this context, then they lose all of their seeming irrationality, and become instead a simple statement by nVidia as to what the company cares about and what it doesn't care about moving ahead. This interpretation would also explain why the company, at this time, seems much more obsessed with marketing than in engineering its next generation of 3d architectures. If nVidia is contemplating getting out of the 3d-gaming reference design marketplace completely, getting out of the 3d-chip markets, then far from being strange, odd, and irrational, everything they said in the Conference with respect to these issues actually makes a great deal of sense.

Of course, they cannot announce such an intention all at once at this time, since they have to finish up whatever financial commitments they have contracted for with various sectors in the 3d-gaming industry--such as board OEMs, software publishers, developers, etc. But at the same time they would want to begin laying the ground work for the eventual announcement so that it doesn't completely hit like a bolt out of the blue. In this context the actual purpose of the "content management team" could in fact be much different than what I assumed at first at the beginning of this post--it could be to begin winding down nVidia's current levels of involvement in the 3d-gaming marketplace.

(2)Next, the remarks about losing benchmarks because they stopped cheating also becomes much less bizarre and strange, if we consider that nVidia simply doesn't care any more about winning them in this market segment--that the whole issues of 3d-gaming benchmarks and all of the associated aspects will soon not be of importance to the company because it will have left the 3d gaming reference-design market.

(3)Let's also examine the strange remark about "not losing market share to ATi but only to Intel," even though it has been more than obvious to everyone that in the 3d-gaming market segments, in the middle and high ends especially, that nVidia has taken a frightful pounding from ATi over the past year in terms of market share. And of course the reasons go far beyond anything to do with cheating in benchmarks and a poor public image and poor drivers--the company has had yield problems *all year long* with respect to nV30/nV35. In fact, nV30 was cancelled and officially described a "failure" by JHH himself and never made it to market. And according to statements released by nVidia in August 2003, yield problems relative to nV35 production were not corrected until September '03--or just last month (if they in fact have been corrected as of yet.) So even if you discount all of the negative publicity nVidia has brought upon itself with all of its activities this year as being a major sales damper for this market segment, the fact that nVidia's not been able to make enough high-profit gpus to sell in quantity all year long is quite enough by itself.

So maybe the proper interpretation of these otherwise bizzare, seemingly RDF pronouncements the company has made about market share losses to Ati this year is that what nVidia meant to say was, "We aren't in the least concerned with whatever marketshare we lost to ATi this year, because going forward those markets will no longer concern us. Rather, nVidia views its foremost market-share challenge to be Intel, in the the high-volume, value-oriented gpu market segments that really have little or nothing to do with the 3d-gaming chip market segments, and as this is where we will be competing from now on it is market-share losses in this market we must address. In this market we do not compete with ATi, but with Intel."

Looking at it that way the RDF no longer applies and nVidia's statement is transformed from a bizarre denial of reality into a sane and rational assessment.

(4)If Gamespot accurately reported the remarks nVidia made about the possibility of only releasing ONE driver per year in the future, then I think this remark pretty much clinches it that nVidia has either already decided to leave the 3d-gaming markets, or else is seriously considering doing so at this time. In the corporate, low-end market where Intel lives graphically, and where nVidia says its competition actually is, it's not uncommon to see Intel generate 1-2 driver releases per low-end corporate-targeted graphics chip per year. If nVidia dumps the 3d-gaming graphics market completely then suddenly all those bugs (not to mention "optimizations" and additional feature support) relating to 3d-gaming support will no longer be a consideration for nVidia--just as they aren't presently for Intel--because its chips aren't targeted for the 3d-gaming market and are not meant to support such games in the first place. DX9 and Intel graphics such as they are definitely do not mix.

To me, the "possibly only one driver per year from now on" remark is simply too pregnant with meaning to ignore. You just cannot remain competitive in the 3d-gaming graphics-chip marketplace and even THINK about "one driver a year" if it is your desire to compete in that market segment with companies like ATi.

So basically, if we view nVidia's comments made at this conference as reported by Gamespot to be an outline of nVidia's "game plan" with respect to its future competition with IHVs like ATi, then the remarks as reported show such a striking disdain for reality that they cannot be considered rational on the whole. But, if we view them from the context that what nVidia *didn't directly say* is missing, and the intention of nVidia is to get out of these market segments in the coming months and restrict itself to competition with Intel on the low-profit, mainly 2d, high-volume end of the gpu markets, then I can honestly find nothing irrational in the Conference remarks at all, and an almost complete consistency.



Summary Factoids


Looking at some other peripheral issues which have occurred in the past year which may lend to the context...

(1) I recall nVidia stating publicly last year prior to shipping nv30 in '03 even as it actually was able to do, that the company had invested some $400M in the development of nv30. I believe AFAIK that this was another of JHH's own remarks. Don't know if it was PR boast or fact, but if it was true, then when we add in the nv30's failure in the market and the subsequent further development work on nV35--complicated all year by yield and production issues--recurring driver issues, etc.--and we add in associated marketing, then it's certainly possible nVidia's spent some $500M-$600M in the past 18-24 months relative to nV30/35, with very little to show for the money spent in the end. After all of that time and money and work, high-profit nV3x is still shy of being competitive with ATi's R3x0 and is generally perceived in its markets as inferior to R3x0. You could argue about lower-end .13 nV3x--but indeed there have been yield and production issues with it as well (though not to the same extent as the high-profit chips), and their DX9 performance is quite dismal in relation to the the competition's--even the low-end, .15 micron nV3x chips nVidia could actually ship performed poorly relative to the competition. OK, that's just gotta' hurt, no two ways about it.

(2) This is slightly parenthetical, but there is an historical precedent for a major chip manufacturing company (though not fabless and certainly far wealthier than nVidia) getting into the 3d-gaming chip & reference design markets, getting soundly whipped by the competition, and taking its marbles home with its tail between its legs. That company was Intel, of course, after abandoning its i74/5xx line when it simply could not compete with 3dfx, nVidia, ATi and even Matrox reference designs at the time. Conditions were different, then, of course--the 3d-gaming chip market was nowhere near as large as it is now, but on the other hand Intel never had anywhere near the yield and production problems that nVidia's had all year with .13 nV3x, and Intel had its own fabs and a whole lot more money to throw away. You know things are indeed hot in a market when Intel can't take the heat.

(3) In what had to be a major blow to nVidia this year in terms of future income prospects, Microsoft abandoned nVidia as the primary xBox chip supplier and tapped nVidia's competition, ATi, instead. For fiscal '03 I read one analyst's report that nVidia received $445M in xBox revenue, or, about 1/3 of the company's gross annual income. That's not going to hit them for a couple of years, maybe, although it will surely hit them to some degree for the remainder of the contract with M$, but nVidia knows the xBox cash-flow train will be derailing without a doubt. And so it has to figure out how it will compensate moving ahead.

(4) nVidia's stock capitalization has suffered this year as the result of the events listed here and all the others we've seen, and investor and analyst confidence in the company has deteriorated markedly.

(5) Largely due to its own short-sightedness and intractability nVidia has suffered major damage this year to its PR image inside of its target markets for 3d-gaming refrence design products. This of course has further eroded an already poor business climate for nVidia and been grossly amplified by the fact that it is almost universally perceived that nVidia has been the author of its own PR misery for most of this year. It is much more difficult, and far more expensive, and pays far fewer dividends, when a company embarks on major marketing campaigns which are perceived by its markets as negative if not false or at the least highly misleading and incomplete. Yet nVidia has shown a marked disconnect all year in communicating with its target end-user markets so that it might have understood that such tactics were not welcomed. Indeed, the company seems to have cared very little for the opinions of the consumers of nVidia-based products, or *potential* consumers of those products, and rather has chosen to overtly attempt to manipulate those opinions to some advantage it could see for itself that few others could.

(6) In the years between the demise of 3dfx in late 2000/early '01 and the advent of R300 in late 2002, nVidia enjoyed a superlative position in the 3d-gaming reference design & chip markets. It had many exclusive chip manufacturers and board OEMs as manufacturing and marketing partners, and the general perception of the 3d-gaming target markets was that nVidia delivered the best of both performance and driver support among all the relative IHVs, including both ATi and Matrox. But with the erosion of its capability over the last year in the face of ATi's R3x0, combined with the failure of nv30, nVidia not only lost the xBox, but it has also lost virtually all of its formerly exclusive board OEMs who have now picked up ATi, as well. Pre R300, nVidia was widely reported to have forced its board partners into exclusive deals simply because there were no other competitive chips for them to choose from at the time and nVidia had them over a barrel. Such advantages as these formely exclusive arrangements provided nVidia, especially in terms of large bulk orders from huge OEMs like Asus, are now in the past for nVidia, as even Asus is now selling ATi-based Asus 3d cards.


______________________________________

Whew! *chuckle* I knew I shouldn't have put on a pot of coffee an hour ago! Sheesh.

Anyway--if we look at the picture as a whole--and view those conference comments as reported by Gamespot in the context that they *might mean* that nVidia is at least considering just pulling out of the general 3d-gaming chip markets and going into multi-function ASICs, motherboard chipsets, and restricting itself to competing with Intel in the value-end 2d corporate graphics-chip market where high-end 3d chip design and advanced API hardware feature support and constant driver support is not involved, their remarks make some sense to me. Otherwise, I'm afraid they don't. It was really the thing about "we compete with Intel and not ATi," and "We might be unable to provide any more than one driver a year in the future for our graphics chips," that really put me on this tangent. Floating the once-a-year-driver concept is just something I can't imagine any 3d-gaming gpu IHV to ever say. BUt it matches the Intel 2d-chip paradigm (with very limited 3d support comparatively) to a T, seems to me.

MSX

It just makes what I said above sound worse. The NF3 is is not a superior chipset to VIA's and it'll likely have the same issues with (lack of)drivers as the nf2. Though perhaps they're aiming lower...

Plus the fact that they were considering dropping the soundstorm, because of its inability to mimic EAX correctly, where they must realize if they put a little time into it they could take a bite out of creative. The 3D soundcard market is so dismal right now as we are practically forced to buy creative... There's lots of great professional cards(loves his M-audio), but they were not designed for games and are in most cases out of price range for most home users that instead opt for an old sb or an integrated solution.

So there's lots for nV to do outside the top 3D market. If (and how) they try to take it is up to them. Or their just changing strategy to have a turnaround with the nV40... time will tell.

blade5545

WOW!!! Thats great thesis on the current state of Nvidia. You should email that to one of the popular Nvidia sites and see what they have to say.

MrThc

That was an excellent summary there WaltC, it really put it all into perspective. Thanks.