NVIDIA shows signs ... [2008 - 2017]

Status
Not open for further replies.
It'll be interesting if any details come up about which chips are failing and how/why they're failing. DigiTimes is reporting that it's the 8500M's failing, along with a lot of "not it" responses from fab partners.

In an unrelated note, Fuad has a surprisingly coherent article about the current situation...
 
If anything, the back to school season will be dominated by HD4850 and G92b (possibly the 9800 GT variant) at the top, followed by G94, RV670, etc.
Speedy GDDR5 RV770's and GT200 high end cards will have little effect overall.

Yes, that's what I had in mind when I referred to nvidia competing with G92b in that cycle. Hence also the reference to dual slot. Of course it was 4850 I principally had in mind for OEMs.
 
It'll be interesting if any details come up about which chips are failing and how/why they're failing. DigiTimes is reporting that it's the 8500M's failing, along with a lot of "not it" responses from fab partners.

In an unrelated note, Fuad has a surprisingly coherent article about the current situation...

Bugger. I hope the Quadros got a bit more QA!

As to Faud, it's not that unusual for there to be quite a bit of difference between what the financial markets think and what competitive reality can be. nvidia had become a "momentum play" stock in recent years with those gaudy margins and consistent buttkicking of the competition. But "live by the momentum players. . . . and get a big (if temporary) tummyache when they desert you at the first sign of trouble".
 
Fuad like so many others seems to me to be confused by what drives a stock price. Too often in tech/gaming journalism the stock price is used as some sort of bizarre proxy for product merit. X company must rock or suck, because their stock price is up or down... And then in some quasi-deluded state of enlightenment, folk like Fuad speak up against that madness (because it certainly is madness), but make the mistake of simply writing off Wall Street as "irrational." They in fact justify the mentality of the price-to-merit fallacy, simply they claim that the execution is warped.

I'm not saying Wall Street can't be irrational in its own terms, but why in the world would his eventuality of NVidia lowering prices to sell cards cause investors to sleep any better at night? On the contrary, that's the very nightmare scenario every NVidia investor has been fearing... because believe me when I say that the institutions invested in NVidia care about their gross margins above all else. And why shouldn't they? Investors are in it for the company that will give them a return on investment, not esoteric reasons centering around who's got a novel architecture.

This paragraph more clearly than any states the condition of his world-view:

At the same time Intel and ATI are doing great, but AMD is down to $5.36 or 1.47 percent today and Intel is down 1.39 percent and this descent has been going on for days. Intel has never been better and AMD is supposed to be fine, at least with its GPU side.

What does he take 'great' to mean? Nice product offerings? He needs to start thinking like someone with money on the line.

Intel's quarterly report will be on July 15th, and AMD's will be on July 17th. People are extremely nervous about the inflationary environment in the US right now, and the hit to high-tech discretionary spending that might take. There's going to be a lot of anxiety leading up to that week of reporting (and probably downside in these stocks) - but whatever ends up being the revealed "trend" at that time, expect it to effect all three of the stocks we're talking about here.
 
They've stepped into muddy waters right now, but if anyone can turn things around quickly, that's Nvidia and their deep R&D pockets/wonder marketing machine.
They can put the squeeze on AMD's margins, just as AMD did it to them (especially now that G92b's die is actually slightly smaller than RV770's, and can be easily mounted onto existing PCB's).

Nvidia is TSMC's largest customer. I'd take it that this "title" comes with its little perks, like a greater price bargaining power compared to ATI/AMD, perhaps ?
 
Do you really doubt it or do you know it ?
And what makes you doubt it, BTW ?

what's nvidia going to do? threaten to take their business to UMC? and produce their new parts on 90nm or something? There's not exactly a huge number of options for producing high end silicon. Their try with IBM didn't work out so well as I recall. There's certainly going to be discounts for volume but Nvidia and AMD are both big volume with TSMC.
 
what's nvidia going to do? threaten to take their business to UMC? and produce their new parts on 90nm or something? There's not exactly a huge number of options for producing high end silicon. Their try with IBM didn't work out so well as I recall. There's certainly going to be discounts for volume but Nvidia and AMD are both big volume with TSMC.

They've used low-end/high volume production shifts to UMC before as leveraging (even some G92's are made at UMC too).
If that's not a "tour de force", i don't know what is...
 
They've used low-end/high volume production shifts to UMC before as leveraging (even G92 is made at UMC too).
If that's not a "tour de force", i don't know what is...

If anything you could make an argument for AMD getting a better price because they've helped TSMC establish the newer processes.
 
Last edited by a moderator:
ati was first at 90, 80, 65 and 55 iirc.

The Geforce 6150 motherboard IGP came out before the X1800 at the 90nm node, and the Geforce 7600 GT's equipped with the 80nm G73-B1 came out before ATI's HD2900 XT...
There was also a very low-volume 65nm version of G71 ("G78") which, obviously, came out before any ATI GPU using that process.
 
They can put the squeeze on AMD's margins, just as AMD did it to them (especially now that G92b's die is actually slightly smaller than RV770's, and can be easily mounted onto existing PCB's).
Slightly smaller? I thought the consensus was that G92b was in the neighbourhood of 260-270mm2, which would actually make RV770 slightly smaller. ;) Also isnt 4850 using the same PCB as the 3850? I doubt x is cheaper argument will fly in either direction here. The only difference is one of them is priced $30 lower .. so I guess the squeeze (for further price cuts) is on Nvidia.
 
Wow, I was just on newegg and I spotted a GT260 for 299 already, amazing. I guess Fud wasnt kidding around about those price cuts, if anything he vastly underestimated them.

http://www.newegg.com/Product/Product.aspx?Item=N82E16814127361


As well as models for 329 and 339.

4870 is still faster though. But the extra RAM on the 260 may be a factor.

280's are being slashed as well. There is already a model on newegg as low as $499 with $40 rebate, making the out the door price $459. Again amazing. So much for $650..

ZipZoomFly also has the 260 for 259 after rebate

http://www.zipzoomfly.com/jsp/ProductDetail.jsp?ProductCode=10008782&showP=1
 
Last edited by a moderator:
And significantly, the back-to-school time is a very high-volume laptop season. It should be noted that some folk eschew the desktop entirely when purchasing a school-bound PC. In an environment where NV might truly be squeezed on the mobile front, that trend would further favor AMD if their footprint in that segment of the market is considerably larger.
None of the new parts (RV770, G200) have any bearing on the mobile space though. As far as I'm aware, nothing has happened on that front that changes the competitive landscape substantially. (My personal opinion is that neither AMD nor nVidia has managed to produce a particularly attractive DX10 generation GPUs for mobile use. 40nm lithography may or may not change that.)
 
None of the new parts (RV770, G200) have any bearing on the mobile space though. As far as I'm aware, nothing has happened on that front that changes the competitive landscape substantially. (My personal opinion is that neither AMD nor nVidia has managed to produce a particularly attractive DX10 generation GPUs for mobile use. 40nm lithography may or may not change that.)

Not the new products, but Nvidia have a definite problem in the mobile space as they have an excessive failure rate with some of their mobile products. This could certainly cause the notebook manufacturers to shift their mobile products toward a competitors solution, because failures certainly reflect badly on them.
 
Not the new products, but Nvidia have a definite problem in the mobile space as they have an excessive failure rate with some of their mobile products. This could certainly cause the notebook manufacturers to shift their mobile products toward a competitors solution, because failures certainly reflect badly on them.
Of course, but an identified and rectified glitch doesn't really matter for future contracts - such things happen, and as long as it isn't part of an overall pattern of shoddy engineering and quality control, such an incident doesn't warrant changing suppliers. If there is such a pattern, or if the cost of the problem is unfairly dumped in the lap of the customer, or if the relationship is already strained for other reasons...well, that's another story.
 
Status
Not open for further replies.
Back
Top