Not surprising: nVidia Market share declining

[digitalwanderer]

Do you think they have to? Can the 9000/9200 cover the sub $100. Isn't nVidia still pushing the GeForce 4 in that space. Just because nVidia is offering ultra low performing 5200 doesn't mean ATI has to follow. How do the 5200 non-ultras compare against the 9000/9200 variants?

I'm afraid I can't fairly answer that question, I'm horribly biased against the 8500 and all derivatives there of no matter what their number.

No, I don't think they HAVE to dominate the entire market...but I do think it would be a great thing to be able to bring the 9700 technology into a sub-$100 card just because I think it's such an amazing bit-o-technology and it would allow a lot more people to try the new ATi.

 

[RussSchultz]

5200's are taking up the shelf space that the GF4 used to.

 

[Fred da Roza]

5200's are taking up the shelf space that the GF4 used to.

With Ultras under $100?

 

[RussSchultz]

5200's are taking up the shelf space that the GF4 used to.

With Ultras under $100?

Likely not. Best Buy/Fry's/Compusa, etc sell everything at list price.

But what does that matter? The 5200's appear to be replacing the GF4MX on the retail shelf.

 

[kemosabe]

Was thinking thats how it would plays out myself. How cheap is .15 at UMC compared to .13 at TSMC. Would ATI still have decent margins offering a .15 RV350 in the $100 - $150 range. If it carried a slight ($10-$20) premium over a 5200 Ultra I would expect it could still sell, provided it was clearly better. As Dave said I guess we will find out soon.

I recall reading a DigiTimes story some time ago about ATI's decision to have RV280 fabbed at UMC, and it was claimed that their 0.15u process was up to 20% cheaper than TSMC's, but I've read nothing reliable about 0.15u vs. 0.13u costs.

Do you think they have to? Can the 9000/9200 cover the sub $100. Isn't nVidia still pushing the GeForce 4 in that space. Just because nVidia is offering ultra low performing 5200 doesn't mean ATI has to follow. How do the 5200 non-ultras compare against the 9000/9200 variants?

Not sure whether they have to, but if you look at the big OEM configurations, the GF4MX slots are largely going to the FX 5200 (major exception being the Presario line which uses the 9200 more), suggesting that NVDA is succeeding in its DX9 marketing campaign. I believe that ATI will be able pick up some of those current discrete low-end slots with the RS300 chipset (e.g. see this blurb mid-page about a rumoured white box design win: http://www.nationalpost.com/financialpost/story.html?id=4924555B-23CF-4931-B43B-A1CA1867920A ), just like Intel stole lots of discrete low-end share with their integrated stuff over the last few years. Nevertheless, from what Orton said at the last CC, I doubt ATI will go into the fall without a sub-$100 standalone DX9 part. Unless I misunderstood what I've read on this board, the RV350 has a very cheap PCB design and should be easy to sell in volume under $100 if yields are as strong as the 0.13u version.

Edit: Here is the DigiTimes story from last December: http://www.digitimes.com/NewsShow/Article.asp?datePublish=2002/12/13&pages=04&seq=28

 

[Fred da Roza]

5200's are taking up the shelf space that the GF4 used to.

With Ultras under $100?

Likely not. Best Buy/Fry's/Compusa, etc sell everything at list price.

But what does that matter? The 5200's appear to be replacing the GF4MX on the retail shelf.

I just check some prices at New Egg and Compusa. The standard 9600 is about the same price as 5200 Ultras (which surprised me). So theoretically a .15 9600 would have to compete with a standard 5200. If it clearly performs better it should be able to carry a slight premium. So no I don't think it should be priced as low as a standard 5200.

 

[Entropy]

But what does that matter? The 5200's appear to be replacing the GF4MX on the retail shelf.

On Dells shelf however, they would seem not to:http://www.digitimes.com/NewsShow/Article2.asp?datePublish=2003/08/15&pages=04&seq=15

The GF4MX has remarkable tenacity. DX7 forever! Small wonder though, it is good at what it does.

Entropy

 

[Fred da Roza]

I recall reading a DigiTimes story some time ago about ATI's decision to have RV280 fabbed at UMC, and it was claimed that their 0.15u process was up to 20% cheaper than TSMC's, but I've read nothing reliable about 0.15u vs. 0.13u costs.


Not sure whether they have to, but if you look at the big OEM configurations, the GF4MX slots are largely going to the FX 5200 (major exception being the Presario line which uses the 9200 more), suggesting that NVDA is succeeding in its DX9 marketing campaign. I believe that ATI will be able pick up some of those current discrete low-end slots with the RS300 chipset (e.g. see this blurb mid-page about a rumoured white box design win: http://www.nationalpost.com/financialpost/story.html?id=4924555B-23CF-4931-B43B-A1CA1867920A ), just like Intel stole lots of discrete low-end share with their integrated stuff over the last few years. Nevertheless, from what Orton said at the last CC, I doubt ATI will go into the fall without a sub-$100 standalone DX9 part. Unless I misunderstood what I've read on this board, the RV350 has a very cheap PCB design and should be easy to sell in volume under $100 if yields are as strong as the 0.13u version.

Edit: Here is the DigiTimes story from last December: http://www.digitimes.com/NewsShow/Article.asp?datePublish=2002/12/13&pages=04&seq=28

After reading that article, if a .15 9600 is manufactured, I'm convinced it will be at UMC.

Even if it's 20% cheaper than TSMC, can a .15 9600 compete cost wise against a 5200? A selling price in the ball park should be good enough provided it's clearly better.

 

[kemosabe]

I recall reading a DigiTimes story some time ago about ATI's decision to have RV280 fabbed at UMC, and it was claimed that their 0.15u process was up to 20% cheaper than TSMC's, but I've read nothing reliable about 0.15u vs. 0.13u costs.


Not sure whether they have to, but if you look at the big OEM configurations, the GF4MX slots are largely going to the FX 5200 (major exception being the Presario line which uses the 9200 more), suggesting that NVDA is succeeding in its DX9 marketing campaign. I believe that ATI will be able pick up some of those current discrete low-end slots with the RS300 chipset (e.g. see this blurb mid-page about a rumoured white box design win: http://www.nationalpost.com/financialpost/story.html?id=4924555B-23CF-4931-B43B-A1CA1867920A ), just like Intel stole lots of discrete low-end share with their integrated stuff over the last few years. Nevertheless, from what Orton said at the last CC, I doubt ATI will go into the fall without a sub-$100 standalone DX9 part. Unless I misunderstood what I've read on this board, the RV350 has a very cheap PCB design and should be easy to sell in volume under $100 if yields are as strong as the 0.13u version.

Edit: Here is the DigiTimes story from last December: http://www.digitimes.com/NewsShow/Article.asp?datePublish=2002/12/13&pages=04&seq=28

After reading that article, if a .15 9600 is manufactured, I'm convinced it will be at UMC.

Even if it's 20% cheaper than TSMC, can a .15 9600 compete cost wise against a 5200? A selling price in the ball park should be good enough provided it's clearly better.

Don't see why ATI couldn't make it competitive, and it should outperform the 5200 (especially with DX9 pixel/vertex shaders) unless the clocks are ridiculously low. I presume they'll try to put the squeeze on Nvidia on both ends (cheap DX8.1 RS300 against legacy DX7 GF2/4MX, and 0.15u RV350 at least equivalent/marginally faster than the 5200 at a similar or slightly higher price point). Idle speculation, but the truth shouldn't be too far off.

 

[kemosabe]

I propose we continue the speculation on this thread: http://www.beyond3d.com/forum/viewtopic.php?t=7479