Sales of Graphics Cards Hits the Bottom, ASP Decreases, ATI Gains Market Share

[Idontcare]

Average Selling Prices on Graphics Cards Continue to Drop ? Jon Peddie Research

Jon Peddie Research reports that the market for graphics add-in boards (AIBs) demonstrated some much-needed firmness in Q2 2009 in terms of unit sales, adding more evidence that demand has bottomed and a recovery is in the offing.

The quarter saw 16.81 million graphics cards shipped, up 3.0% sequentially and down 15% year-over-year (YoY), the latter figure presenting a substantially more moderate drop than the previous two quarters. Unfortunately, the market value dropped 4% to $2.598 billion sequentially and 28% annually.

Obviously, the sales drop of performance-mainstream graphics cards caused a decline in ASPs to $154 per card. The prices are still not as low as back in Q1 2008, when both performance and mainstream market segments were down.

JPR reports that attractive products across all market segments as well as aggressive pricing helped ATI to increase its market share from 31% in Q1 2009 to 35% in Q2 2009, while Nvidia?s share declined to 64%.

http://www.xbitlabs.com/news/v...n_Peddie_Research.html

There are some nice tables and graphs in the link. I had no idea the market value for discreet video cards had collapsed so severely over the past 3 yrs. No wonder both AMD and NV report quarter after quarter of near break-even profits/losses.

4% market share gain for AMD though, QoQ, not bad for the red team. Profit-less marketshare gains though aren't exactly what you get into the business for though, so I'm sure no one at AMD is exactly popping corks on the Dom Perignon right now.

 

[bryanW1995]

yeah, they might just be having a bud light instead. We all bitch and moan about graphics card prices, but if these guys don't go back to making lots of $$ there won't be anything left for R&D for the next gen.

 

[allies]

I'll be purchasing a next gen card, most likely from ATI if it lives up to the hype upon release.

 

[JSt0rm]

next gen doesnt matter when all games coming out are made for current gen consoles. I have never seen such a plateau in my life and I'm enjoying it.

 

[StinkyPinky]

I've had my 8800gt for two years, and it runs every game to a level I'm happy with. Crysis still has the best graphics and that's over 2 years old. Things are slowing down

 

[BTRY B 529th FA BN]

So what can we, who are buying the next gen graphics cards expect (as in price)? The usual as discussed before? Ya know, either way, it's fine. I will still be buying either the 5890 or 5870. I've been holding out with this vanilla 4870 and it's been doin enough for me to ... hold off until these new ATI/AMD cards come out...

 

[SlowSpyder]

I would expect prices to be pretty low at this point, we're very close to next gen. I would think that the board partners will be trying to clear out parts as the new AMD and Nvidia cards show up. Once the new cards show up I have a feeling that the average selling price will be back up.

 

[Demo24]

I'm wanting to get myself a 4870 (maybe a 4890 if the prices equal) here in the near future to replace my aging x800xl(just because of shader issues). I'm curious as to if the prices of these current cards are going to drop that much in the coming weeks? I haven't kept up with hardware pricing in over 4 years so I don't remember how these things played out.

 

[ChaosDivine]

Originally posted by: StinkyPinky
I've had my 8800gt for two years, and it runs every game to a level I'm happy with. Crysis still has the best graphics and that's over 2 years old. Things are slowing down

+1. Had my 8800GTS 320MB (G80) for the longest time. Then 4850 and only moved up to the 4890 because of a 1080p monitor.

 

[Idontcare]

Originally posted by: SlowSpyder
I would expect prices to be pretty low at this point, we're very close to next gen. I would think that the board partners will be trying to clear out parts as the new AMD and Nvidia cards show up. Once the new cards show up I have a feeling that the average selling price will be back up.

This looks pretty bleak IMO. And this basically shows you that an entire market segment has practically evaporated over the same timeframe (3.7B -> 1.1B in <3yrs).

Begs the question - where is the cashflow engine that is fueling tomorrow's GT200 and RV770? If it took the cashflow of 2005-2006 to support the creation of the GPU's we enjoyed in 2008...just what can we expect to see from these two GPU makers in the coming years given that their sales (and thus R&D resources) have been neutered to the tune of ~50%?

 

[BTRY B 529th FA BN]

Karma

 

[bryanW1995]

Originally posted by: Idontcare

Originally posted by: SlowSpyder
I would expect prices to be pretty low at this point, we're very close to next gen. I would think that the board partners will be trying to clear out parts as the new AMD and Nvidia cards show up. Once the new cards show up I have a feeling that the average selling price will be back up.

This looks pretty bleak IMO. And this basically shows you that an entire market segment has practically evaporated over the same timeframe (3.7B -> 1.1B in <3yrs).

Begs the question - where is the cashflow engine that is fueling tomorrow's GT200 and RV770? If it took the cashflow of 2005-2006 to support the creation of the GPU's we enjoyed in 2008...just what can we expect to see from these two GPU makers in the coming years given that their sales (and thus R&D resources) have been neutered to the tune of ~50%?

nvidia is going to have to copy amd's small-ball strategy imho. They've seen this coming a lot longer than we have, this explains their full court press on tesla, integrated graphics, mobos, etc. In 5 years nvidia's graphics division will be as small as amd's and they'll need other sources of income to remain competitive if this trend continues.

 

[BenSkywalker]

Originally posted by: Idontcare
This looks pretty bleak IMO.

This chart we could currently call 'The Crysis Factor'. Take a look at that chart and on a realistic basis you are looking at an ASP of what it will take to play the most demanding game out at 'reasonable' settings. Reasonable settings evolve, but we haven't seen much of a change in terms of displays for quite some time, 30" is going to be about as big as they get, so until we have some games that make people want to upgrade, we just aren't going to see that trend reverse itself too much(we will get a bump when the DX11 parts hit from the tech enthusiast sector, but gamers drive the market).

And this basically shows you that an entire market segment has practically evaporated over the same timeframe (3.7B -> 1.1B in <3yrs).

That actually goes back to the Crysis factor again, they are lucky they haven't seen an even larger decline however. If it wasn't for people overhauling their rigs, those numbers would likely be much lower then what they already are. Combine these factors with a massive economic downturn on a global basis and you have pretty much a 'perfect storm'.

Begs the question - where is the cashflow engine that is fueling tomorrow's GT200 and RV770?

Sony, MS and Nintendo. While the discrete market may be having issues, the consoles are not. ATi has moved ~100Million parts in the last few years in the console space, nV has moved ~25Million. At this point, those contracts are just printing cash. Margins on a doller per unit basis aren't very high, but they are pretty much pure profit at this point. The PC market is still going to provide the bulk of their revenue obvously, but the consoles give them a steady flow of profit without any significant continuing work.

If it took the cashflow of 2005-2006 to support the creation of the GPU's we enjoyed in 2008...just what can we expect to see from these two GPU makers in the coming years given that their sales (and thus R&D resources) have been neutered to the tune of ~50%?

The R&D level of nV at least hasn't declined(AMD doesn't break out R&D per division), just their profit levels have. This is one reason I see them focusing the PhysX angle very hard at this point(a good chunk of my PS3 games give the PhysX splash screen at this point, even the little PSN mini games). nV is going to push the PhysX angle to drive new sales, graphics aren't going to change in any large way until the next generation of consoles hits, but if nV can get developers to add PC specific content in the way of PhysX they can help drive sales of their parts and adding physics effects it significantly cheaper then reworkiing your entire engine and all of your art assets in order to please the, by far, smallest platform sales wise.

 

[ronnn]

Originally posted by: JSt0rm01
next gen doesnt matter when all games coming out are made for current gen consoles. I have never seen such a plateau in my life and I'm enjoying it.

Agreed, except I would enjoy new tech with games that use it for more than a visual add on.

 

[pcslookout]

Originally posted by: ChaosDivine

Originally posted by: StinkyPinky
I've had my 8800gt for two years, and it runs every game to a level I'm happy with. Crysis still has the best graphics and that's over 2 years old. Things are slowing down

+1. Had my 8800GTS 320MB (G80) for the longest time. Then 4850 and only moved up to the 4890 because of a 1080p monitor.

I still have my 8800 GTS 320MB! Sure am getting my money's worth!

 

[Idontcare]

Originally posted by: BenSkywalker

Originally posted by: Idontcare
This looks pretty bleak IMO.

This chart we could currently call 'The Crysis Factor'. Take a look at that chart and on a realistic basis you are looking at an ASP of what it will take to play the most demanding game out at 'reasonable' settings. Reasonable settings evolve, but we haven't seen much of a change in terms of displays for quite some time, 30" is going to be about as big as they get, so until we have some games that make people want to upgrade, we just aren't going to see that trend reverse itself too much(we will get a bump when the DX11 parts hit from the tech enthusiast sector, but gamers drive the market).

And this basically shows you that an entire market segment has practically evaporated over the same timeframe (3.7B -> 1.1B in <3yrs).

That actually goes back to the Crysis factor again, they are lucky they haven't seen an even larger decline however. If it wasn't for people overhauling their rigs, those numbers would likely be much lower then what they already are. Combine these factors with a massive economic downturn on a global basis and you have pretty much a 'perfect storm'.

You've probably seen the data, or are in tune with it enough to at least posit a good guess about, but is there much to say about saturation of higher screen resolutions here too?

Seems like 5-6 yrs ago gaming at 1024x768 was not the norm, but 17/19" monitors weren't the norm either...then the LCD revolution took over and 17", then 19" (and correspondingly higher resolutions) became your typical gamer specs.

But now things are about as big (screen size wise) as your average desk will handle, in the 24" area, and resolutions seem to have mostly leveled off at the 19x12 level (extreme users will be extreme of course) so I am wondering if along with the crysis effect if there has also been a sort of saturation in the demands on the output side (LCD size/resolution) of graphics as well.

Hence ATI's desire to convince gamers they need 3 screens for their 2010 games, and since they need 3 screens they need an HD5870?

Your thoughts?

 

[shangshang]

I purchased the 4870 some times ago and since then I haven't played any game that my old 4850 couldn't handle. I ain't dropping another dime in all this hype about "next gen" cards if there ain't no games out there to take advantage of them yet. Some people walk around always saying "oh i can't wait to buy a next gen card."... for what? so they can brag? or reload Crysis for another round? No thanks. I will keep my 4870 1GB and will Crossfire if it drops below $100 bux AND if there are actual games that require such power AND they are GOOD games. As it stands, I have played all the games I wish to play and my 4870 are just sitting idly sucking electricity when I turn on my PC. I'm actually thinking of buying a low powered video card to put in my PC because I'm a little bored of games right now.

side note, I wish there's a way to turn off one video card completely (via BIOS) when you have multiple video cards on your mobo! I think this would be a very big deal considering how much these video cards are sucking energy. I would definitely buy a mobo that can do this! Asus and Gigabyte, are you listening? ATI and NV, are you listening?

 

[Mr Fox]

Originally posted by: BenSkywalker

Originally posted by: Idontcare
This looks pretty bleak IMO.

This chart we could currently call 'The Crysis Factor'. Take a look at that chart and on a realistic basis you are looking at an ASP of what it will take to play the most demanding game out at 'reasonable' settings. Reasonable settings evolve, but we haven't seen much of a change in terms of displays for quite some time, 30" is going to be about as big as they get, so until we have some games that make people want to upgrade, we just aren't going to see that trend reverse itself too much(we will get a bump when the DX11 parts hit from the tech enthusiast sector, but gamers drive the market).

And this basically shows you that an entire market segment has practically evaporated over the same timeframe (3.7B -> 1.1B in <3yrs).

That actually goes back to the Crysis factor again, they are lucky they haven't seen an even larger decline however. If it wasn't for people overhauling their rigs, those numbers would likely be much lower then what they already are. Combine these factors with a massive economic downturn on a global basis and you have pretty much a 'perfect storm'.

Begs the question - where is the cashflow engine that is fueling tomorrow's GT200 and RV770?

Sony, MS and Nintendo. While the discrete market may be having issues, the consoles are not. ATi has moved ~100Million parts in the last few years in the console space, nV has moved ~25Million. At this point, those contracts are just printing cash. Margins on a doller per unit basis aren't very high, but they are pretty much pure profit at this point. The PC market is still going to provide the bulk of their revenue obvously, but the consoles give them a steady flow of profit without any significant continuing work.

If it took the cashflow of 2005-2006 to support the creation of the GPU's we enjoyed in 2008...just what can we expect to see from these two GPU makers in the coming years given that their sales (and thus R&D resources) have been neutered to the tune of ~50%?

The R&D level of nV at least hasn't declined(AMD doesn't break out R&D per division), just their profit levels have. This is one reason I see them focusing the PhysX angle very hard at this point(a good chunk of my PS3 games give the PhysX splash screen at this point, even the little PSN mini games). nV is going to push the PhysX angle to drive new sales, graphics aren't going to change in any large way until the next generation of consoles hits, but if nV can get developers to add PC specific content in the way of PhysX they can help drive sales of their parts and adding physics effects it significantly cheaper then reworkiing your entire engine and all of your art assets in order to please the, by far, smallest platform sales wise.

The thing that is forgotten here is that both Players are now playing in a Free Market Economy, instead of the Price Fixing that went on for years at our expense..... They made this bed themselves.

Remember the USDOJ slapped these guys hard ... Just not quite as hard as the DRAM guys...

The Grand Nagus Huang would be wearing jumpsuits.. If that was the case..

This sector is quickly becoming bland, and boring except the Red Team keeps slowly getting better.

Until the developers quit developing for Consoles this will continue..

Native PC, and PC Only Applications are becoming a part of history, unless the industry gets it's head out of it's ass soon.

 

[bryanW1995]

I believe that "head in the ass" has little to do with it. MS, sony, nintendo are all very powerful players, consoles are going to continue to dominate regardless of what the "industry" does.

@ SHANGSHANG: use rivatuner to underclock your gpu to ridiculous levels when its not in use.

 

[MODEL3]

Originally posted by: Idontcare
This looks pretty bleak IMO.

The TOP marks are Q1-Q2 2006 (265$ ASP)
Which is natural since we had simultaneous launches of ATI & NV products in ALL the price range (550$-70$) (NV7XXX series & ATI 1XXX series) (550$ NV+550$ ATI=550 upper high weighted average, of range)

& Q2 2007 (273$ ASP)

Again is natural, simultaneous launches of ATI & NV products in nearly ALL the price range (700$-60$) (NV8XXX series & ATI HD2XXX series) (700$ NV+400$ ATI=550 upper high weighted average, of range)

When in a quarter, we have ONLY one company launching new products across the whole price range then the ASP is always lower! (in relation with, if we had simultaneous launch)


Even for quarters like Q2-Q3 2008 when we had simultaneous launches of ATI & NV prod. (only ATI in the whole price range) (450$-40$) (NV GTXX series & ATI HD4XXX series) (450$ NV+300$ ATI=375 upper high weighted average, of range)

You can see that in Q1-Q2 2006 & Q2 2007 the indicative upper high weighted average, of range is nearly 1,5X in relation with Q2-Q3 2008, so instead of 265/270$ ASP it natural that we have 180$ ASP! (if we do the price scaling below and calculate the ASP)

As long as NV & ATI keeps releasing products ranges with price range upper highes at 450/500$ (NV) & 300$ (ATI) according to the scaling for the bellow prod. line, the ASP is going to be always much lower than 279$!

I like the ASP today, i think that, NV and ATI can survive (and execute R&D just fine) with today's ASP in Discrete Desktop Graphics Card Market! (NV & ATI have ASP in other markets also) (ASP is a problem but most serious problem is SP/chip cost ratio)

(If they like it or not is another reason!, I am not going to pay for AMD's ATI buyout or AMD's financial status or for NV's big strategic mistakes, or am I? 😕 ) 🙁

Forget the above, what I'm just trying to say is, that of cource the ASP decreased but the 270$ -> 155$ is not exactly indicative (wait to see how low will be the Q3 2009 No) for the future status of Discrete Desktop Graphics Card Market! (things will get better (for them))

Originally posted by: Idontcare
And this basically shows you that an entire market segment has practically evaporated over the same timeframe (3.7B -> 1.1B in <3yrs).

The data is wrong! (well not wrong, but the market Segmentation is based on model level correlation, not on price level correlation and i disagree with this methodology!, like JPR care... :laugh: )

I am not sure even for that, and i am too lazy to check but just for example:

Q3 2006 TOTAL MARKET VALUE: 4.97 bil. $

Q3 2006 ENTHUSIAST ONLY MARKET VALUE: 3.71 bil. $

So the ENTHUSIAST level sold 3.71 bil. $!

and the PERFORMANCE + MAINSTREAM + VALUE + WORKSTATION level sold 1.26 bil. $

Yea right!

The highest priced VGA (one unit) in Q3 2006 was the (then launched) ATI X1950XTX at 399$. (or some 7900GTX at 399$)
Even if there wasn't other lower priced VGAs in the ENTHUSIAST category:

3,71 bil. $ / 400$ = 9,275 million units in Q3 2006

If we add the lower priced VGAs in the ENTHUSIAST category, then the units are way higher than 9,275 mil.

I want say more, rest is easy!



The real problem is not what you are saying:

"(3.7B -> 1.1B in <3yrs)"

the problem is that they went from 20 billion $ market values a year (2006)
to nearly 10 billion $market values a year (2009)! (half, -10 bil. $)

Half is ASP (1,5X), half is unit volume (1,3X) (many reasons)

There may be many reasons like:

Rise of Notebook Market share!
Rise of Netbook Market share!
Rise of IGP Market share!
Rise of Total Console Market share for the gaming public!
Launch from from Q4 2006 and on of Wii & PS3! (takes away potential new PC gamers)

The above are just examples! I am to lazy to check and correlate!

 

[Idontcare]

Originally posted by: MODEL3
The real problem is not what you are saying:

"(3.7B -> 1.1B in <3yrs)"

I did not say that was the "real" problem.

I said those numbers indicated to me that an entire market segment had practically evaporated in 3 yrs.

As to statements I have made regarding the "real" problem, in the OP I stated:

Originally posted by: Idontcare
I had no idea the market value for discreet video cards had collapsed so severely over the past 3 yrs. No wonder both AMD and NV report quarter after quarter of near break-even profits/losses.

However as others have rightly pointed out this is not really "the" problem either. Sure it is "a" problem but "the" problem appears to be that neither AMD nor Nvidia have much of a compelling case to make to people to convince them to upgrade their existing and quite functional GPU's to a product that will produce only a marginally superior gaming experience.

The "real" problem it would seem is that there is a dearth of killer apps in the desktop gaming world...the same Achilles heel plaguing Intel and AMD with their cores war in the CPU biz.

Until games that people want to play require hardware that didn't exist last year there won't be much of an upgrade cycle...enter the primary purpose of DX11.

 

[BenSkywalker]

You've probably seen the data, or are in tune with it enough to at least posit a good guess about, but is there much to say about saturation of higher screen resolutions here too?

Absolutely. While right now based on sales trends I would say the majority of gamers are still in the 22" range, the current sales trends indicate that we should have the majority in the 24" range within the next year or two. I don't think the majority will ever move to 30" LCDs, OLEDs will be at mass market price points before 30" displays hit the mass gamer market sweet spot for pricing and their technical advantages are rather huge. On a realistic basis, 19x12/1080p are where we are going to be at for the 'norm' for quite some time. Of course we will have ouliers as always, but we should be pushing close to 24" for the norm within the next ~12-18 months(for the gamer market, tech segment was there a while ago) at which point resolution will be a 'solved' issue(in terms of calculating out targets for markets etc).

But now things are about as big (screen size wise) as your average desk will handle, in the 24" area, and resolutions seem to have mostly leveled off at the 19x12 level (extreme users will be extreme of course) so I am wondering if along with the crysis effect if there has also been a sort of saturation in the demands on the output side (LCD size/resolution) of graphics as well.

Yes, 1080p/4x/16x is almost certainly going to be the target for most developers for the forseeable future. 8x does look better then 4x, but not enough to warrant serious effort, and speaking for myself I would love to see 32x AF, but it just isn't realistic as the differences wouldn't be noticed by the overwhelming majority. The level of visuals gamers consider reasonable is going to be hanging in the 1080/4/16 range for quite some time, this isn't likely to go away with the next gen of consoles either, until 30" OLED displays become mass market priced I think that is where we are most likely to stay.

Hence ATI's desire to convince gamers they need 3 screens for their 2010 games, and since they need 3 screens they need an HD5870?

This may be the approach that they are going to attempt to counter PhysX with in terms of driving sales. The problem is going to be this type of setup has never had any mass market appeal at all. Don't get me wrong, there absolutely is a market for it, and it could help them out in the business sector quite a bit honestly, but in terms of the broader market very few games bother taking advantage of multiple displays in no small part because the overwhelming majority of people don't have a setup that can handle it. Getting developers to support tri display would be a staggering challenge for ATi, nVidia owns 65% of the market in a bad quarter for them and they still provide most of the work to get PhysX in to the small handful of games that utilize to any real advantage. Getting the same thing done for a group that, at best, consists of ~3% of the market is going to take Herculean effort on ATi's part.

I'm not saying that isn't a market they are going to pursue, I think they see that as an oppurtunity to lock down that segment for themselves. It may be a small market, but you can charge a reasonable premium and right now 3% of the overall market is around 10% of ATi's overall marketshare, a non trivial amount for them. Given the oppurtunity for ASP and margins in that segment, I can certainly see ATi making a strong push to take over that market(and hoping nV doesn't bother countering- the market is too small to make much of a fight over for either company).

 

[MODEL3]

Originally posted by: Idontcare

Originally posted by: MODEL3
The real problem is not what you are saying:

"(3.7B -> 1.1B in <3yrs)"

I did not say that was the "real" problem.

I said those numbers indicated to me that an entire market segment had practically evaporated in 3 yrs.

As to statements I have made regarding the "real" problem, in the OP I stated:

Originally posted by: Idontcare
I had no idea the market value for discreet video cards had collapsed so severely over the past 3 yrs. No wonder both AMD and NV report quarter after quarter of near break-even profits/losses.

However as others have rightly pointed out this is not really "the" problem either. Sure it is "a" problem but "the" problem appears to be that neither AMD nor Nvidia have much of a compelling case to make to people to convince them to upgrade their existing and quite functional GPU's to a product that will produce only a marginally superior gaming experience.

The "real" problem it would seem is that there is a dearth of killer apps in the desktop gaming world...the same Achilles heel plaguing Intel and AMD with their cores war in the CPU biz.

Until games that people want to play require hardware that didn't exist last year there won't be much of an upgrade cycle...enter the primary purpose of DX11.

Correct IDC!
I am not arguing with anything you said!

I just wanted to point out how big is the real figure that the "market value for discreet video cards had collapsed so severely over the past 3 yrs"!

The real figure is not 3.7-1.1=2.6 billion $ (Enthusiast only)
The real figure will be something like -9/8 billion $ (-7 bil. in 2008) for 2009!

I don't like this JPR analysis! (If someone can tell me how the 3.71 bil. $ figure in Q3 2006 for the Enthusiast segmentation is correct, i will like it better!)

 

[nitromullet]

NVIDIA needs to make a console, and use it's existing relationships with developers via TWIMTBP to get them to create games that:

are CUDA based to run on the gpu
use PhysX
are native 1080p

The reason I say this is because consoles are killing their gpu market, and the best way for them to counter that would be to drive the console market. Don't get me wrong, it wouldn't be easy, but if they were able to get enough devs involved in writing games that ran entirely on NVIDIA gpus this could work. Not only would devs be making games for NVIDIA's "nBox", but all the games would run on NVIDIA gpus as well. ...Of course, you'd have to upgrade to whatever NVIDIA's current high end gpu was at the time because it would be the only one that could play these game adequately at 1080p.

I know this is pie in the sky type stuff, and I'm not even sure that CUDA would be sufficient to code an entire game engine in. However, as it is Sony and MS seem content to let graphics technology languish where it is right now while they try out out different market strategies on their aging consoles.

edit: oh yeah, to make sure that its console competes properly the console would come with a controller, but also support certain NVIDIA approved mouse/keyboard setups from the big names like Logitech and Razer.

 

[ChaosDivine]

Originally posted by: pcslookout
I still have my 8800 GTS 320MB! Sure am getting my money's worth!

Used to have two. Sold one in FS/FT. Using the other in my file server (great for playing Guild Wars / Perfect World in front of the boob tube).