AMD: Pirate Islands (R* 3** series) Speculation/Rumor Thread

PS. Since when people want corporations to make money at their expense? What world we are living in!?!
It's not that your arguments don't have some merit. But the hyperbole of questioning the world we're living simply because people like to talk about business strategy kind of devalues the whole thing.
 
PS. Since when people want corporations to make money at their expense? What world we are living in!?!
Well, in the real world, corporations have bills to pay such as the wages for their employees. Now, if a corporation doesn't make any money, they can not pay their employees. And a corporation without employees ceases to exist fairly quickly. Thus, if you enjoy your computer, phone, tablet, gpu, etc... you probably want these to remain viable businesses, because if they don't then the products produced will also go away. This really shouldn't be that hard to understand, but I digress...
 
Please define "Just fine" in real numbers.
It is well known that high-end GPUs have the highest GMs. 50%+ should never be a problem, but I expect it to be higher in today's mature 28nm. At $650, AMD would have to screw up really bad if they don't get 40% margins.
 
Isn't this AMD's first attempt in making giant GPUs? First attempts seem to have low yields at the start then improve over time with tweaks.
That's not how it works. AMD has plenty of experience with seriously large dies. Increasing the size from 440 to 600 won't magically increase the defect density.
They're probably more concerned with packaging related yield.
 
Well, in the real world, corporations have bills to pay such as the wages for their employees. Now, if a corporation doesn't make any money, they can not pay their employees. And a corporation without employees ceases to exist fairly quickly. Thus, if you enjoy your computer, phone, tablet, gpu, etc... you probably want these to remain viable businesses, because if they don't then the products produced will also go away. This really shouldn't be that hard to understand, but I digress...

I'm a business owner and I think most established businesses know their overheads and margins better than us bystanders. So for example if I want to sell my product for X dollars and ran through numbers several times to pick best price point(s) for it considering supply / demand estimates as well as market capacity for given product you think you know better from outside how much I should charge? I have never had a customer say to me they are not happy with my pricing because it's too cheap! This kind of thinking only applies to people not seriously interested in buying given product. Quickest way to put AMD out of business would be for them to price Radeon Fury X at $1999.99! It would give them enormous margins per board sold, but where would be the volume to support all that R&D?

So please don't get me wrong, I'm happy to pay $650.00 for a card approximately 50% faster than my current one to support corporation which is building new and exiting technologies to help them develop even better things in the future, but market rules dictate, certain price points will generate certain volumes and businesses must find their sweet spots for best return on investment. AMD is in difficult position as it's market share and mind share are at long time low point. They will not sell well enough at nVidia price points any time soon. They have to work on their reputation first and the best way to do that is through amazing products and good market placement. Besides, AMD will have their margin wonder product in Radeon Fury X2, so they will have that price point covered, just not now.
 
WiNvXWn.png
 
I'm a business owner and I think most established businesses know their overheads and margins better than us bystanders.
Given the massive number of businesses that fail every year along with AMD's ongoing woes, I'll trust reality over what anyone thinks, sorry.
 
Last edited:
Quickest way to put AMD out of business would be for them to price Radeon Fury X at $1999.99! It would give them enormous margins per board sold, but where would be the volume to support all that R&D?
The situation is rather odd with the Fury X. It's priced for a somewhat less super-niche segment, but it has certain features that seem to be consistent with more limited volumes that shouldn't tip AMD's fortunes either way.

Does a very new integrated package with a new-but-sunsetting HBM1, closed-loop cooler with what appears to be a bespoke mounting plate, and a card that promised to ramp to a standards-defying 500W sound like a volume play?
Maybe it's a feeler for a broader deployment of a closed-loop solution, but I don't know if Fury is what AMD is counting on to pay the bills for the better part of a decade of R&D.
 
@Jawed
Nano looks great and should still be a lot faster than R9 290X according to Dr Lisa. I'm curious how low idle power will go on Fiji based cards as with HBM there should be significant saving even comparing to downclocked GDDR5. This card might be exactly what HTPC crowd is asking for, low power, high performance, excellent idle and hopefully good acoustics.

@ninelven - never mind, enough off-topic
 
Lightman,
Nobody was proposing a $2000 price and using it as an example doesn't serve a purpose. And $1000 may be an issue indeed due to the RAM.
But if FuryX has the meaningful performance increase (say 10+%) over a GTX 980 Ti that it should have on cases that are not memory size limited, then it wouldn't hurt them one bit to ask at least $750. I think a lot of people would be cheering about what a good deal it is compared to a Titan X.
 
The situation is rather odd with the Fury X. It's priced for a somewhat less super-niche segment, but it has certain features that seem to be consistent with more limited volumes that shouldn't tip AMD's fortunes either way.

Does a very new integrated package with a new-but-sunsetting HBM1, closed-loop cooler with what appears to be a bespoke mounting plate, and a card that promised to ramp to a standards-defying 500W sound like a volume play?
Maybe it's a feeler for a broader deployment of a closed-loop solution, but I don't know if Fury is what AMD is counting on to pay the bills for the better part of a decade of R&D.

Obviously I can only speculate here, but maybe AMD has contractual obligations to Hynix for HBM1 volumes they have to meet or more likely they are hoping to regain a bit of enthusiast mind share by making their product favourable not only on price, but also performance, build quality and overclocking headroom. It's a niche product, but at the same time it's a halo product for AMD and the more positives we enthusiasts will voice on internet about Fiji the better for AMD whole line-up. I could make a car analogy, but I digress.
 
Last edited:
The crowd that buys Titan X would eat up a FuryX at $850.
What crowd? We're talking about a comparative handful of people, sum total, across the entire globe, who buy these fantasy land-priced video cards.

I don't get it.
You're complaining that the card isn't as expensive as it could have been? I don't get you.
 
Lightman,
Nobody was proposing a $2000 price and using it as an example doesn't serve a purpose. And $1000 may be an issue indeed due to the RAM.
But if FuryX has the meaningful performance increase (say 10+%) over a GTX 980 Ti that it should have on cases that are not memory size limited, then it wouldn't hurt them one bit to ask at least $750. I think a lot of people would be cheering about what a good deal it is compared to a Titan X.

I understand your point and certainly there are people who would buy at that price without thinking twice, but in case nVidia reacted and AMD had to cut prices 2 weeks after launch ... we've already been there and know how many happy customers this created.
It comes down to risk and reward. I'm happy with $650 as this is less than I expected and I will be buying this card if performance is not too far off what leaks shown. I was ready to pay up to $749 but that is my personal maximum I can justify spending on a GPU knowing there will be better one next year. I'm not super rich and there certainly are wealthier people than me who could afford higher price points. What holds Fiji top end pricing big time in my opinion is HBM1 limit of 4GB as it rules this card from most of semi-professional and professional tasks. That's where Titan X 12GB makes it easier for designers and engineers or scientists to justify expense. I know a lot of my friends in academic circles upgraded their computers with newest GPU's for research purposes (founded by government) because they needed to work with bigger data sets. AMD Fury X would not fly past commission deciding on giving grants with this argument because they already have at least 4GB cards in their research systems!
 
I think a lot of people would be cheering about what a good deal it is compared to a Titan X.

The truth is that a hell more people would be cheering about what a great deal it is, when the price is set as it is at 650$.

The truth is that the lower the price the more tempting and popular the deal will be.

There is no need for AMD in their sensitive condition to overdo things in the wrong directions and spoil the success.
Making something expensive is not the way to go. Especially not in these times when it will become even harder to more people to afford this.
 
It's impressive in a sense that 980Ti reference cooler is a jet engine. They are priced the same so comparison is valid.
It's basically an apples-to-oranges comparison, but that advantage may be necessary to remain competitive....
 
Back
Top