nVidia wins this round - Charlie D.

[blastingcap]

Again, they spend weeks studying this stuff. AMD is not stupid. They determine what amount of profit is earned if boards are priced at 400$. They study what amount of profit is earned at 450$. They study the profit level at 500$.
Please understand that pricing decisions are not taken lightly, anyone that has been inside these companies would be well aware of this. The 550$ pricing model gives them the most profit and most of the public is willing to pay the premium for it because it does outperform the GTX 580.

If kepler is released 6 months from now with better performance, obviously the current pricing model will either change, or NV will charge that much more.

You guys act like pricing decisions are taken lightly. The spend MONTHS studying what gives them the most sales and profit. I don't like premium prices anymore than you do -- but its easy to understand why they do it. Simple economics

Dude. Quit it. You speak condescendingly based on, what, 2 semesters of college econ, as you put it? (By the way, companies are not guided by the invisible hand; MARKETS are. Thank me later if that saves you a point on your midterm or something.)

Real life doesn't work the way the stylized models in your textbook say. Okay? It's messy and chances are, nobody in marketing remembers what a derivative was. Okay? There are more incompetent people in positions of power than you can imagine. And even the competent ones are not perfect.

It is trivially easy to obliterate your contention that marketers know best. Look at the price-gouging prices charged for 8800 GTX, or HD 58xx at release. Marketers are not gods. They make best guesses based on market data, and guess what, sometimes they are wrong. Things change, variables change. If they set the price while yields were X, then yields go to Y, what happens? If the price of transporting a box of cards doubles in two months, what happens? If a general strike springs up at the assembly plant, what happens? Etc.

In this case, I suspect 28nm yield suck and AMD is more than happy to start off at a higher price point than in the past, since there is not much in the way of supply anyway, plus NVDA has no competing part anyway. Similar to why Nikon did NOT even try to match Canon's sales this past holiday season, because Thailand destroyed much of Nikon's DX lens and camera production. As yields improve or as NVDA introduces competitors, I expect prices to soften accordingly, all else equal.

Please don't drink the neoclassical economics Kool Aid. It got us crap like utter faith in the market leading to Bear Stearns and Lehman. NO IT IS NOT ALWAYS PRICED INTO THE MARKET ALREADY. The assumptions built into neoclassical economic models are laughable. Perfect information does not really happen, there are vast information asymmetries leading to well-documented and growing problems like insider trading. How serious is the problem? People have won Nobels for asymmetric information work. The root causes of the 2008 crisis stemmed from an utter lack of due diligence and common sense, and bad incentives coupled with moral hazard of governments bailing out undeserving private sector companies. You had idiot counterparties not doing their due diligence on absolutely insane mortgage-backed securities products.

If you are an econ or business major, change your major before it's too late. I suggest statistics, mathematics, computer science, physics, or any engineering.

P.S. As a practical matter it is easier to overprice something and then give rebates or outright price reductions later, than the other way around. People get especially angry at you for RAISING prices over the product launch number.

 

[blastingcap]

lol

And this my friends is what you call a successful FUD campaign. Nvidia has no product, no benches, no credible leaks, no nothing. But hey, Nvidia is going to crush AMD, because Charlie said so. You know, that same Charlie that was reviled and ridiculed over and over again. But now his word is suddenly the gospel! How can one not laugh at this?

FUD campaign? Are you suggesting Charlie is now on NVDA's payroll or something? Shyeah right.

These true/false/in-between leaks and rumors happen all the time, and if NVDA really wanted to FUD it up, they would have done so at around the launch date, not weeks afterwards.

Edit to add: To be clear, "launch date" meant either the paper launch or hard launch of 7970, take your pick.

 

[gramboh]

Product pricing is not done by the marketing department alone. It is usually done by Finance folks with input from marketing and operations. The Finance people (in large consumer organizations) do extensive modelling to optimize returns/profit. Companies like AMD/NVidia have huge datasets (from past products) that business intelligence folks would be analysing in addition to consumer sentiment, competitor offerings/timing etc.

 

[AnandThenMan]

FUD campaign? Are you suggesting Charlie is now on NVDA's payroll or something? Shyeah right.

Then I take you believe Charlie is on AMD's payroll currently?

 

[96Firebird]

Then I take you believe Charlie is on AMD's payroll currently?

Was.

 

[blastingcap]

Then I take you believe Charlie is on AMD's payroll currently?

What? That was a total non-sequitur.

You are taking Charlie's recent article to be part of, as you put it, a FUD campaign. That implies some sort of positive link between Charlie and NVDA which I find hard to fathom.

And for the record, I don't know if Charlie is paid directly or indirectly by AMD for special services, so I can't comment on that.

It's apparent to anybody visiting SA that AMD and AMD affiliated companies have advertised a lot on that site.

Not saying that is a cause of bias or whatever, and in fact the causality may run the other way around... like if Charlie independently bashed NVDA and AMD saw that and liked it and then independently decided to advertise on SA.

 

[SirPauly]

Many people have told me if one meets Charlie -- he is a nice guy. For Example: Max, whom reviewed a bit for SemiAccurate told me this and keep an open mind.

I still feel Charlie is a sensationalist but does seem to have contacts but he certainly had an axe to grind with nVidia and actually used a mistranslation as a source against nVidia.

Totally surprised by the tone of the offering about nVidia's Kepler but he claimed he saw it first hand. Grain-of-salt and place it into memory and will find out how accurate, semi-accurate or way the hell off it may be.

 

[blastingcap]

Many people have told me if one meets Charlie -- he is a nice guy. For Example: Max, whom reviewed a bit for SemiAccurate told me this and keep an open mind.

I still feel Charlie is a sensationalist but does seem to have contacts but he certainly had an axe to grind with nVidia and actually used a mistranslation as a source against nVidia.

Totally surprised by the tone of the offering about nVidia's Kepler but he claimed he saw it first hand. Grain-of-salt and place it into memory and will find out how accurate, semi-accurate or way the hell off it may be.

Charlie certainly has access to sources and people that is sometimes surprising. E.g., http://semiaccurate.com/2010/06/14/tim-sweeney-and-andrew-richards-debate-future-graphics-hardware/

http://www.youtube.com/watch?v=tnogwO84O0Q

 

[AnandThenMan]

What? That was a total non-sequitur.

When you say, "Are you suggesting Charlie is now on NVDA's payroll" the now strongly suggests Charlie being paid off previously. For example if I said, I'm now working for Honda, the obvious question would be, who did you work for before?

As for Charlie, I've always found him to be fairly accurate after you filter out the sensationalism. So this time, I believe what he is saying is fairly close to the truth.

 

[blastingcap]

When you say, "Are you suggesting Charlie is now on NVDA's payroll" the now strongly suggests Charlie being paid off previously. For example if I said, I'm now working for Honda, the obvious question would be, who did you work for before?

First of all, this was my full quote:

"FUD campaign? Are you suggesting Charlie is now on NVDA's payroll or something? Shyeah right."

My quote does not suggest what you say it suggests. It says that a) I doubt that Charlie was on NVDA's payroll before his most recent article (the "now" part); b) I don't think Charlie is on NVDA's payroll or working with them on a FUD campaign (the rest of the quote).

If I had written something like this it could have been interpreted the way you interpreted it: "Are you saying that Charlie switched sides and is on NVDA's payroll now?"

But I didn't write that, because I was in fact not making any such suggestion.

As for Charlie, I've always found him to be fairly accurate after you filter out the sensationalism.

Fairly accurate after filtering meaning... semi-accurate? Sorry, I could not resist.

 

[AnandThenMan]

I'm sure people realize this by now, but semi-accurate is a play on words based on semiconductor. Actually it does seem like some don't realize this. :\

 

[Arkadrel]

Well... only time will tell...

Is charlie trolling....?
trying to fish out people that take credit from his articels?
Doing something so unexpected to generation page hits -> ad $ ?

or

Does he simply write what he thinks after haveing seen a kepler?

The short story is that Nvidia (NASDAQ:NVDA) will win this round on just about every metric, some more than others.

He's probably talking about:
performance/mm^2
performance/watt
performance/$

this *doesnt* (have to) mean the card being released will be faster than the 7970.

My guess is the comming card has good performance/mm^2.

 

[blackened23]

Dude. Quit it. You speak condescendingly based on, what, 2 semesters of college econ, as you put it? (By the way, companies are not guided by the invisible hand; MARKETS are. Thank me later if that saves you a point on your midterm or something.)

Real life doesn't work the way the stylized models in your textbook say. Okay? It's messy and chances are, nobody in marketing remembers what a derivative was. Okay? There are more incompetent people in positions of power than you can imagine. And even the competent ones are not perfect.
.

I'm not saying i'm an economics expert -- In fact, I was pointing out my lack of expertise. People who say "AMD should price it at 400$ or they will lose sales" No they won't. They know what they're doing, they have tons of experts that stare at spreadsheets all day long and AMD has ridiculous amounts of statistics and market research that tells them how they should price. You have internet experts here who think they know what they're talking about in this matter (and i'm not saying i'm an expert) --they suggest that AMD should price their product lower when it outperforms the competition. That is utter stupidity. Fact of the matter is Kepler may outperform 7970, but is it here yet? No. You don't price based on something that isn't here yet. You price based on here and now, and right now 7970 is the king. If Kepler is released and all the suspect rumors are true, then they will adjust accordingly. Its just 1 of the basic principles of economics that anyone who has studied eco in college knows. Charge what the market will pay. And judging by how quickly etailers are selling out of their stock, it seems to be working.

Besides all of this, all of the Kepler rumors are extremely suspect. I feel like Charlie may be trolling nvidia based on inside information. You have rumors last week indicating that the first kepler would merely match 7970. Now all of a sudden Charlie says its better. Now if it is better, thats great, i'll be the first to buy it in June. But rumors are rumors, and can I buy a rumor? Nope.

 

[Imouto]

I'm sure people realize this by now, but semi-accurate is a play on words based on semiconductor. Actually it does seem like some don't realize this. :\

http://www.thefreedictionary.com/semi-

loled

Seriously.

 

[blackened23]

Product pricing is not done by the marketing department alone. It is usually done by Finance folks with input from marketing and operations. The Finance people (in large consumer organizations) do extensive modelling to optimize returns/profit. Companies like AMD/NVidia have huge datasets (from past products) that business intelligence folks would be analysing in addition to consumer sentiment, competitor offerings/timing etc.

Exactly this. People here who think they know more about how AMD should price than AMD themselves, are ridiculous. AMD has an entire department full of finance/research types that study this crap all day long.

 

[Keysplayr]

Well... only time will tell...

Is charlie trolling....?
trying to fish out people that take credit from his articels?
Doing something so unexpected to generation page hits -> ad $ ?

or

Does he simply write what he thinks after haveing seen a kepler?

He's probably talking about:
performance/mm^2
performance/watt
performance/$

this *doesnt* (have to) mean the card being released will be faster than the 7970.

My guess is the comming card has good performance/mm^2.

While I don't believe a word uttered out of Charlies mug, It's probably not a good idea to start selectively defining the meaning "every metric". Every metric means every metric. Which goes against whatever logic you used to determine that it *doesnt* have to mean it's faster. If it wasnt' faster, the words every metric would probably not have been used because that would account for what most people are looking for.. I mean really.

And so we might really know what Charlie is mulling over, maybe we should appoint somebody to ask him. Maybe Anandthenman could do the deed. After all, he is the one who correctly defined the name of Charlies web site when nobody else had a clue. I'd say that makes him the closest to the situation. I nominate. Seconds?

 

[Arkadrel]

While I don't believe a word uttered out of Charlies mug, It's probably not a good idea to start selectively defining the meaning "every metric". Every metric means every metric. Which goes against whatever logic you used to determine that it *doesnt* have to mean it's faster. If it wasnt' faster, the words every metric would probably not have been used because that would account for what most people are looking for.. I mean really.

The short story is that Nvidia (NASDAQ:NVDA) will win this round on just about every metric, some more than others.

1) every metric = more than 1 thing.
2) just about = not everything, whos to say performance for this card isnt that?

also if every metric isnt:

performance, cost, performance/cost, performance/watt, performance/mm^2
tempatures? noise?

what the hell is it then?

Im just saying, it doesnt have to be performance.

a 760 could be slower than a 7950, but still have better performance/cost, lower power use ect.
And that would still fit what charlie is saying.

And it would still be impressive for charlie, he looks at chips from a engineering point of view.
To him stuff like performance/mm^2 is impressive, and performance/watt.

while to most other "normal" people/consumers they just care about performance, and how loud it is.

 

[blastingcap]

I'm not saying i'm an economics expert -- In fact, I was pointing out my lack of expertise. People who say "AMD should price it at 400$ or they will lose sales" No they won't. They know what they're doing, they have tons of experts that stare at spreadsheets all day long and AMD has ridiculous amounts of statistics and market research that tells them how they should price. You have internet experts here who think they know what they're talking about in this matter (and i'm not saying i'm an expert) --they suggest that AMD should price their product lower when it outperforms the competition. That is utter stupidity. Fact of the matter is Kepler may outperform 7970, but is it here yet? No. You don't price based on something that isn't here yet. You price based on here and now, and right now 7970 is the king. If Kepler is released and all the suspect rumors are true, then they will adjust accordingly. Its just 1 of the basic principles of economics that anyone who has studied eco in college knows. Charge what the market will pay. And judging by how quickly etailers are selling out of their stock, it seems to be working.

Besides all of this, all of the Kepler rumors are extremely suspect. I feel like Charlie may be trolling nvidia based on inside information. You have rumors last week indicating that the first kepler would merely match 7970. Now all of a sudden Charlie says its better. Now if it is better, thats great, i'll be the first to buy it in June. But rumors are rumors, and can I buy a rumor? Nope.

I don't necessarily disagree with your conclusion, just how you got there.

 

[Keysplayr]

Exactly this. People here who think they know more about how AMD should price than AMD themselves, are ridiculous. AMD like has an entire department full of finance/research types that study this crap all day long.

And still, they can get it wrong. All companies are guilty of this. All companies make mistakes. And yes, this does indeed mean that they are not perfect, contrary to belief.

 

[Keysplayr]

1) whos to say performance for this card isnt that?

I suppose you did. Stick to your plan. Wait and see.

 

[blastingcap]

1) every metric = more than 1 thing.
2) just about = not everything, whos to say performance for this card isnt that?

also if every metric isnt:

performance, cost, performance/cost, performance/watt, performance/mm^2
tempatures? noise?

what the hell is it then?

I interpreted it to mean "every metric that matters." Examples: perf/watt, quieter, cooler, more cost effective to produce, etc. Charlie also mentioned GK104's modest size, so it'd be ironic if NVDA won on perf/mm^2 as well.

Besides, SA admin Copper already stated in SA forums that at least one of the things Charlie meant was better or equal performance, yet also better thermals at the same time, which strongly implies better perf/watt.

 

[Guovssohas]

To the ones saying AMD is overpricing the 7970; the gtx580 3gb also costs 550.... So that is ok? A weaker product in every aspect and it's price is fine? ehh,, lol? Please explain.

 

[Arkadrel]

SA admin Copper already stated in SA forums that at least one of the things Charlie meant was better performance at better thermals, which strongly implies better perf/watt.

Nvidia did say they where aiming for perf/watt with compute... so that wouldnt be surpriseing.

 

[blastingcap]

To the ones saying AMD is overpricing the 7970; the gtx580 3gb also costs 550.... So that is ok? A weaker product in every aspect and it's price is fine? ehh,, lol? Please explain.

preexisting GTX580 to SLI with
functional 3D prior to the HD7xxx launch
CUDA
PHYSX

For most people though it is an inferior buy for the dollar, I agree.

 

[blackened23]

what this thread needs now is tweakboy

No callouts.
-ViRGE