HD 5000 series has sold 25 million units, AMD has 90% of DX11 marketshare

[Lonyo]

Would be true, if it wasn't for NVIDIA sales picking up faster than AMD's...and they(NVIDIA) have increased their DX11 market share compared ever since.

So much for a "fermi" failure...despite being 6 months late and the irrelvant "uses more power!!!1111!one" it's ganiing ground on AMD every month.

Even if AMD is selling DX11 cards, NVIDIA is selling more DX11 cards at the moment.

Do you have any actual proof of either of these statements?
Oh, and Steam numbers aren't proof, no matter how much you wish they were. I haven't seen any links stating these "facts", so I was hoping you could link me to your sources.

 

[GaiaHunter]

"AMD on Thursday announced revenue for the third quarter of 2010 of $1.62 billion, a net loss of $118 million, or $0.17 per share, and operating income of $128 million"

"AMD reported third quarter revenue of $1.62 billion, up 16 percent from one year ago. Excluding one-time charges such as those related to AMD’s sale of GloabFoundries, the company reported non-GAAP earnings of $108 million or 15 cents a share, which blew away Wall Street’s consensus estimate of 6 cents a share. Overall, under GAAP rules the chip maker took a net loss of $118 million or 17 cents per share."

 

[Lonbjerg]

Do you have any actual proof of either of these statements?
Oh, and Steam numbers aren't proof, no matter how much you wish they were. I haven't seen any links stating these "facts", so I was hoping you could link me to your sources.

I use Steam as it's vendor neutral...unlike PR form either side.
Infact the only cards that gained shared last month was NVIDIA's...I am sorry (not really) the world dosn't conform to "Fermi is a failure" as you had hoped *shrugs*

 

[Dribble]

Ati operating income of $1 million!!

Didn't expect that - as far as I knew they've been selling well with little competition (on the low end anyway), have kept prices high, and the 40nm is mature and giving good returns so how come they only made $1 mil?

Surely this is the time they should be raking in the profits - if they aren't now they what's going to happen as nvidia starts to compete better (at least it has an almost full range of DX11 chips now)?

 

[Tempered81]

25 million units shipped with $1 million profit? With 9:1 DX11 marketshare? And they are carrying the company? Look at those numbers again.

Please don't get personal and call me a troll, point out where I am incorrect.

You are incorrect about almost everything you type here. Since the reason you entered this thread and made the 2nd post was to point out things that refute the op, and continue to argue with everyone in the thread, you are obviously what everyone is calling you. If you aren't that full time, you're playing the role right now.

You know, cuz i don't think you're an idiot, that AMD sells cpus, solutions, and services and it competes with intel. ATI had the best sales in it's 25 year history last year. NO denying it, 25 million graphics cards sold. Somewhere between 80% and 90% of all the DX11 graphics cards in the world are AMD/ATI and the other 10-20% are Nvidia.

Throwing in the figure for the entire AMD company profit/loss margins from your armchair Ceo point of view really takes up too much space in this enjoyable thread for me, and for the otheres here calling you names.

 

[blckgrffn]

I use Steam as it's vendor neutral...unlike PR form either side.
Infact the only cards that gained shared last month was NVIDIA's...I am sorry (not really) the world dosn't conform to "Fermi is a failure" as you had hoped *shrugs*

Glad you could take the thread there. Good work.

Anyways...

I think it will be interesting to see what AMD/ATI can do once they can get out from under all of that debt. If they can, that is. Seems like profits (and life in general) would be much better for them if that was the case.

 

[Lonyo]

I use Steam as it's vendor neutral...unlike PR form either side.
Infact the only cards that gained shared last month was NVIDIA's...I am sorry (not really) the world dosn't conform to "Fermi is a failure" as you had hoped *shrugs*

I was wondering whether you were using your faulty logic* than increasing Steam marketshare related to sales, or whether you had actual evidence not speculative crap based on something entirely non-representative of the market being discussed.

(Being 'vendor neutral' does not make Steam accurate, also it's not vendor neutral).

*Arguments against which have been presented before many, many times, you just seem to ignore them and continue on your merry way.

 

[Lonbjerg]

Glad you could take the thread there. Good work.

Anyways...

I think it will be interesting to see what AMD/ATI can do once they can get out from under all of that debt. If they can, that is. Seems like profits (and life in general) would be much better for them if that was the case.

The thread was "there" from first post, thank the OP *shrugs*

 

[blckgrffn]

The thread was "there" from first post, thank the OP *shrugs*

That's a whole lot of caring for all the shrugging. *shrugs*

So are we going to see a lot of optimization for the DX11 implementation that AMD used due to their volume or are so many of these cards so slow on DX11 that it doesn't matter?

 

[nemesismk2]

Edit: /facepalm at the people buying DX9/10 cards still, not at you posting the data.

They are probably the people who want the best performance for their money. For example:- until recently I could buy a new XFX 4890 for the same price as a 5750? Which would you buy for best performance? The same company was also selling 4870 and 4850 video cards for bargain prices.

 

[Lonbjerg]

To give another angle om this:

May: 3.7%
June: 6,19%
July: 8.6%
Aug: 12,9%
Sep: 15, 72%

NVIDIA's share on steam Survey.
DX11 cards.



NVIDIA selling more DX11 cards per month than AMD since launch...despite being late (mainly TMSC's process to blame) and consumers like performance > powerconsumption when given the choice...despite what is often posted here on the forum.

 

[Bateluer]

NVIDIA selling more DX11 cards per month than AMD since launch...despite being late (mainly TMSC's process to blame) and consumers like performance > powerconsumption when given the choice...despite what is often posted here on the forum.

I'm not the biggest fan of the Fermi cards, but the bolded statement is very true. Unless the power consumption is simply ungodly, most people don't care. When you're building/shopping for a high performance machine, you don't give a damn how much power it draws. Its performance you want.

Building HTPCs or lower end machines, thats an entirely different matter. And one where, until pretty recently, AMD had far more options for people looking for a card to fit their needs.

 

[Lonbjerg]

I'm not the biggest fan of the Fermi cards, but the bolded statement is very true. Unless the power consumption is simply ungodly, most people don't care. When you're building/shopping for a high performance machine, you don't give a damn how much power it draws. Its performance you want.

Building HTPCs or lower end machines, thats an entirely different matter. And one where, until pretty recently, AMD had far more options for people looking for a card to fit their needs.

What I meant?
I am talking about DX11 GPU (GPU's that actully can run DX11 and not turn into a DX11-decellarator) and those numbers (NVIDIA) show that (the majority of )gamers put performance above powerconsumption when buying GFX cards...

Go Figure, right? *chough*

I have been saying for a long time that looking at highend GPU and powerconsumption is a fallacy...seems the market agrees.

 

[SlowSpyder]

So even with 90% of DX11 shipments, they were still shipping a little over 50% of discreet cards in Q2? Who the hell was buying DX9/10 cards in Q2??? /facepalm

Edit: /facepalm at the people buying DX9/10 cards still, not at you posting the data.

I think you might be forgetting mobile sales? Is AMD counting those in their numbers? If I remember correctly they took the majority market share from Nvidia in mobile sales which accounted for their overall market share increase more so than discreet parts.

Would be true, if it wasn't for NVIDIA sales picking up faster than AMD's...and they(NVIDIA) have increased their DX11 market share compared ever since.

So much for a "fermi" failure...despite being 6 months late and the irrelvant "uses more power!!!1111!one" it's ganiing ground on AMD every month.

Even if AMD is selling DX11 cards, NVIDIA is selling more DX11 cards at the moment.

Yea, Fermi is a raging success with it's 10% DX11 market share. How do you figure Nvidia is selling more right now? As Russian pointed out, AMD sold a LOT of parts over the last few months. Fermi has had price drop after price drop, they are losing partners. I think Nvidia's early poor reviews of Fermi doomed it, it has had bad mojo since it launched.

AMD had 100% of the market for a long time. All Nvidia had to do was sell once card and they cut into that market share. AMD maintaining ~90% market share even seven months after Nvidia started introducing Fermi parts is a pretty stong statement. AMD has shipped 25 million DX11 parts, do you honestly believe Nvidia has come close to that?

 

[BenSkywalker]

If we take all of this as absolute fact, ATi has sold 25 Million DX 11 parts and they have 90% of the DX11 market- that would indicate that nVidia has sold 2.5 Million DX11 parts eighty times more then Charlie said was possible. You know, the guy many of you were touting as being so accurate about Fermi

Somehow the operating income of the GPU division was +$1 million which seems absurdly low

Those kinds of numbers place the 5xxx series in the realm of failure on any sort of business level. With $390 million revenue generated and only $1Million in profits made with no special charges? For a point of reference, nVidia's GPU division- the one that can't hope to compete, generated ~150 times more profit last quarter(they did have additional charges, but when looking at Fermi v 5xxx series on a business level to date, Fermi utterly dominates).

 

[Bateluer]

What I meant?
I am talking about DX11 GPU (GPU's that actully can run DX11 and not turn into a DX11-decellarator) and those numbers (NVIDIA) show that (the majority of )gamers put performance above powerconsumption when buying GFX cards...

Go Figure, right? *chough*

I have been saying for a long time that looking at highend GPU and powerconsumption is a fallacy...seems the market agrees.

Err, the 5xxx series cards run DX11 without becoming a decelerator. If you're serious about DX11 performance, you're shopping on the higher end regardless. The 5xxx series is still first gen DX11 hardware, and we go through this for all first gen hardware. The low end parts don't run the new DX well. It was the same for DX10, DX9, DX8, and DX7.

 

[Lonbjerg]

I think you might be forgetting mobile sales? Is AMD counting those in their numbers? If I remember correctly they took the majority market share from Nvidia in mobile sales which accounted for their overall market share increase more so than discreet parts.




Yea, Fermi is a raging success with it's 10% DX11 market share. How do you figure Nvidia is selling more right now? As Russian pointed out, AMD sold a LOT of parts over the last few months. Fermi has had price drop after price drop, they are losing partners. I think Nvidia's early poor reviews of Fermi doomed it, it has had bad mojo since it launched.

AMD had 100% of the market for a long time. All Nvidia had to do was sell once card and they cut into that market share. AMD maintaining ~90% market share even seven months after Nvidia started introducing Fermi parts is a pretty stong statement. AMD has shipped 25 million DX11 parts, do you honestly believe Nvidia has come close to that?

If you have 100% of the market with no competion...well DUH...that is a given.
But to loose ~3% of that marketshare month after month...when you even had a blissfull 6 months of headstart of no competion is not "stomping"...it's loosing ground, due to the performance/featires of said cards.

I marked a fallacy with red...what do you want with that fallacy?
You expect me to reply to that fallacy?

 

[Grooveriding]

To give another angle om this:

May: 3.7%
June: 6,19%
July: 8.6%
Aug: 12,9%
Sep: 15, 72%

NVIDIA's share on steam Survey.
DX11 cards.



NVIDIA selling more DX11 cards per month than AMD since launch...despite being late (mainly TMSC's process to blame) and consumers like performance > powerconsumption when given the choice...despite what is often posted here on the forum.

First off, you need to correct misusing commas instead of decimals, further I don't think you understand how to interpret those numbers. Because you see nvidia's DX11 cards showing increases does not mean they are selling more in any way, shape or form.

The size of the installed base is increasing as both companies sell more DX11 cards. Nvidia started at 0% seven months ago and AMD was at 100%, Nvidia had nowhere to go but up, and AMD had to nowhere to go but down. Even if AMD sold 1,000,000 cards a day and NV sold 1, AMD would still show a decrease.

Here is the full breakdown of DX11 cards from September 2010's Steam hardware survey:

Per those numbers

AMD has 84.3% of the DX11 market
Nvidia has 15.7% of the DX11 market

Not exactly the same as the numbers in the subject of this thread, but pretty darn close, enough that now having two sources saying close to the same thing, it's a safe bet AMD really does have the lion's share of DX11 and has sold a huge number of cards in the 5XXX series compared to what Nvidia has in the 4XX series.

 

[Lonyo]

If you have 100% of the market with no competion...well DUH...that is a given.
But to loose ~3% of that marketshare month after month...when you even had a blissfull 6 months of headstart of no competion is not "stomping"...it's loosing ground, due to the performance/featires of said cards.

I marked a fallacy with red...what do you want with that fallacy?
You expect me to reply to that fallacy?

PLEASE. This has been explained. OVER AND OVER in many other threads.
ATI will continue to lose marketshare in the long run until marketshare percentages level out at approximate sales ratios.

ATI had 100% of the market.
Lets assume the eventual split is 45 (NV)/55 (ATI) in the long term, with ATI and NV selling equal amounts of cards since both launched.
ATI/AMD will continue to lose marketshare month after month after month until we hit the 45/55 split representative of total sales (in this example, the headstart represents the 10% difference in total sales, this is all hypothetical).
NV gets to this 45/55 point by selling the exact same number of cards per month every month as ATI.
That means that NV will increase their marketshare from 0% to this 45% despite only selling the same number of cards each and every month, until it reaches some sort of equilibrium.
Their increasing marketshare does NOT reflect higher sales of cards vs ATI, it reflects sales of cards in a ratio exceeding their current marketshare.


Hpothetical again:
ATI has 90% marketshare, NV has 10%.
The split between sales in a month is 40 NV vs 60 ATI, NV's marketshare will increase, because they have sold more than their 10% current overall marketshare, but they still sold less than ATI.

STOP BRINGING UP NV'S INCREASING MARKETSHARE ON STEAM AS A WAY OF "SHOWING" THAT THEY ARE OUTSELLING ATI. IT DOES NOT SHOW THAT AT ALL.

The best it does is show they have more current sales than their existing marketshare.
It also doesn't show ATI outselling NV. It shows nothing with respect to either company except NV is selling more than <current marketshare> vs ATI. That could be (based on a 10~15% vs 85~90% current share) anything from NV taking say 16% of the market to taking 100% of the market. You don't know. We don't know. Stop saying it.

 

[Genx87]

Honestly who gives a crap about DX11 marketshare? What quality does that bring to the table? It doesnt appear to bring massive profit.

 

[SlowSpyder]

text

Right.

I hope my math is right, feel free to correct me if I'm wrong. Just some hypothetical numbers to try (again) to illustrate to Lonbjerg how Nvidia can gain market share by selling less parts than AMD.

Let's say AMD has sold 25 million DX11 parts, Nvidia has sold 2.5 million. So out of the 27.5 million total market share of DX11 parts, Nvidia has 9.1&#37; of the market here.

Now, let's say AMD sells another 2.5 million and Nvidia sells 500,000 parts. That brings the total market up to 30,500,000 DX11 parts. In this case Nvidia just increased their market share to 9.8%.

So in this completely made up scenario AMD would have sold 5x as many parts as Nvidia but still lost overall market share. AMD will not keep 100% market share, it just isn't going to happen.

 

[v8envy]

Honestly who gives a crap about DX11 marketshare? What quality does that bring to the table? It doesnt appear to bring massive profit.

It matters quite a bit when it comes to vendor-proprietary features. As in, "Well, we can spend $ adding some physx fluff, but then 90% of our high end audience, the vocal enthusiasts with DX11 hardware, will not be wowed. Maybe we should use a more portable library instead..."

Eye candy is what sells modern games. Making decisions to limit eye candy for 90% of your high end customers will have a much greater "trickle down" sales impact than doing so for 10%.

 

[blastingcap]

I use Steam as it's vendor neutral...unlike PR form either side.
Infact the only cards that gained shared last month was NVIDIA's...I am sorry (not really) the world dosn't conform to "Fermi is a failure" as you had hoped *shrugs*

NV infinitely increased their market share in DX11 since April. Of course their market share has increased, you can hardly do worse than ZERO. But look at Steam: http://store.steampowered.com/hwsurvey/videocard/

I don't know why I bothered responding, given that you favor NV so much, but... see how the percentage declines have tailed off for 58xx and 57xx?

This indicates that the market shares are starting to stabilize. E.g., for 5700 series it goes from 40 to 38 to 36 to 33 to 33&#37;, and for 5800 it goes from 40 to 37 to 34 to 31 to 31% For 5600 series, market share has been at ~3% for a long time, ditto with 5500 series loitering at 1-2%. (Apparently most people who buy those cards don't game on Steam.)

With the 6-series debut, so long as AMD prices them right, I have little doubt that in the near term something like a 8:2 to 9:1 ratio will keep holding true unless NV comes up with something FAST, like a GTX495 dual-GF104 card, and even then, those kind of high-end cards make up a tiny fragment of the market. The GT 430 may do okay though, as it's the only solution for HTPC-3D right now.

P.S. You may find this informative: http://en.wikipedia.org/wiki/Asymptote

 

[Ares1214]

Would be true, if it wasn't for NVIDIA sales picking up faster than AMD's...and they(NVIDIA) have increased their DX11 market share compared ever since.

So much for a "fermi" failure...despite being 6 months late and the irrelvant "uses more power!!!1111!one" it's ganiing ground on AMD every month.

Even if AMD is selling DX11 cards, NVIDIA is selling more DX11 cards at the moment.

Newer products gain more ground than faster ones generally, but id like to see any proof you have of that.

 

[cusideabelincoln]

Right.

I hope my math is right, feel free to correct me if I'm wrong. Just some hypothetical numbers to try (again) to illustrate to Lonbjerg how Nvidia can gain market share by selling less parts than AMD.

Let's say AMD has sold 25 million DX11 parts, Nvidia has sold 2.5 million. So out of the 27.5 million total market share of DX11 parts, Nvidia has 9.1&#37; of the market here.

Now, let's say AMD sells another 2.5 million and Nvidia sells 500,000 parts. That brings the total market up to 30,500,000 DX11 parts. In this case Nvidia just increased their market share to 9.8%.

So in this completely made up scenario AMD would have sold 5x as many parts as Nvidia but still lost overall market share. AMD will not keep 100% market share, it just isn't going to happen.

Your math is right.

For a general thought:
We'll represent AMD's share by the variable A and Nvidia's by the variable N.

Combine them for a ration N/A (N divided by A)
Now we know A > N, afterall AMD did have a head start and has sold more.

So that means N / A is less than 1.

Now at this point let's say N and A are increased by the same constant independently, we'll call C. This would mean AMD and Nvidia are selling the same amount of cards. Well the more times you add C to the denominator and numerator, the closer the ratio (N / A) gets to 1. A ratio of 1 means they have completely equal marketshare.

If you were to plot this kind of increase on a graph then it would not form a straight line. It would be a curved line, where the slope gets gentler as the ratio approaches 1. The slope represents the ratio. So at first the ratio would have a higher degree of change than it does at the end.

The only way for the ratio to be greater than 1 is for Nvidia to increase their rate faster than AMD. This would mean N increases in size more each period than A does.

Now all of this is fine and dandy, but as you have pointed out the ratio (marketshare) does not correlate to actual units of increase.

AMD could (and probably is) sell more and Nvidia will still gain marketshare. This is because at first their marketshare percentage (ratio) will rise rapidly, but it will taper off at some point unless they actually do sell more than AMD does. However we'll probably be past Fermi and onto new video cards by the time we see how this will finally all play out, which basically means the growth of marketshare number Lonjberg likes to throw out is completely meaningless now and for the future.

Also it's not hard to imagine AMD selling more DX11 cards now anyway. They have more cards available in more performance brackets, and Nvidia just released their mid range and low end cards.

edit: Well my explanation stinks. I haven't been to bed yet and it's been five years since I've done Calculus. But I think the spirit is there...