Originally Posted by tajoh111
I got confused I was thinking about Bonaire.
I am basing pricing off of pitcairns which was a smaller die on a cheaper process. It was priced at 350. You can tell something is wrong with AMD pricing just by looking at the pricing structure of everything. With the rx 480 starting at 200, it basically leaves too small of a space to sell a whole lineup of videocards. Particularly their low end cards.
That doesn't make it more legit than anything, I don't see the argument but I can tell you I don't agree and that is all I need to do.
It's not the first time AMD or ATI surprised the GPU world whit their prices.
You are basically saying that because the 7870 was priced at 350 it's impossible for AMD to design a 200$ 480, I would say that is bs and pure speculation.
The last time cards on a new node, with a similar die size had pricing like the rx 480 and rx 470 was the 3870 and 3850 which was 8.5 years ago. And these cards were $220 and $180 and 20% smaller. Considering how much wafer prices have gone up, there is no way, AMD would normally price their cards like this.
Wafer prices are not everything especially when you don't have real information of AMD's situation, everything you say is pure speculation.
Anyone familiar with business can tell their going to be problem with AMD pricing is the products are simply priced too closed together and leaves no room for their products to stack to breath. This is because the rx 480 tops out at 230 and starts at 200. With the rx470 starting at 150 which likely has 32cu, where does this leave 16cu products? The answer is too cheap for today's market. For the price/performance to make sense, these products need to be 80 dollars or less, which is way below last gen's cape verde or the 7770 which was 160 dollars.
I don't see the problem man.
If the situation would be reversed and Nvidia would have launched the 1060 at 200 and 230$ first I'm absolutely sure you would make up all king of positive scenarios regarding them.
The main problem is we are talking about AMD here.
AMD has waited too much for a new node to think that they didn't have this planned.
Designing new products at very competitive prices is what I would do, it's the best way. It already looks like it's hard for Nvidia to mach AMD's GPUs in price or specs so their strategy is clearly working.
Having pricing too close together like this also creates confusion for consumers. That's why there is distinct price points for cars for example within the same make.
Again I don't see the problem, their strategy is working in our advantage too. If the difference is too small just buy the better alternatives.
Supply is clearly not a problem.
There is no indication finfets is cheaper and every indication that its going to be more expensive.
There is also no indication nobody can design a GPU to target a certain price point.
And strictly looking at economics, GF prices can't be any better than samsungs or TSMC, based on economies of scale. High fixed cost with smaller variable = economy of scale. We have the cost it produce something at samsung and TSMC on finfet. But it would be foolish to assume GF is cheaper. The single highest expense in wafers for manufacturing is the equipment depreciation expense. This is because equipment typically costs in the billions and quickly becomes obsolete. And if you look at papers on it, it represents 60% of the expense. This is generally a fixed expense. Because GF volume is so much smaller than samsung(30+billion) and TSMC(25billion) at 5 billion and the cost of the equipment is in the billions, GF should have to price their wafers higher than samsungs and even TSMC to make money. Particularly when you add in GF is paying royalty fee's to use samsungs technology.
It's their fabs after all, AMD is their biggest costumer and Polaris already won AMD a few contracts so decent volumes will allow AMD to better negotiate certain prices or advantages.
I don't think it's a pure coincidence that they have chosen 14nm for all their new products.
If you don't think companies will sell things at no profit, look at intel. Intel has lost billions trying to put their x86 mobile chips into everything. Pricing too high vs the competition can lead to inventory build up, followed by losses when that inventory is written off. Atleast with a low price from the beginning, you can steal marketshare from your competition and create higher brand value from reviews.
Intel was willing to sacrifice some of their profits and was still hardly affected by it, AMD is simply not in the same situation.
There is absolutely no way AMD's management would approve selling Polaris at 0 profit or extremely low profits in their current situation.
The only thing I said positive about Nvidia is they have performance per mm2 advantage along with the performance per transistor advantage. If you don't know how this can effect pricing, then I can't help you there.
You have to admit, the competition of the graphic industry was much stronger when AMD could pack more performance in a given die size than Nvidia. When Nvidia out right ceased it with maxwell, AMD marketshare and graphics profit took a beating.
Yeah sure that is all.
I'm not going to do a detectives work, I've read enough posts to know what I'm talking about.
Maxwell was successful because AMD was not able to properly refresh their last gen although they were still able to compete.
Now whit Polaris they are clearly way more competitive. It will be interesting to see if Nvidia can respond in both price, performance, specs and volumes. Their last gen mid range was not very good.
The thing is I know the same information you do, I mostly read the same stuff and I simply don't agree whit you, especially based on my experience. I worked in marketing, sales, quality control and for the last few years in business2bussines handling strategic clients so I've been in many factories, I've seen a lot of stuff. Until I see the full economical effects Polaris had on AMD I'm not going to negatively speculate.