• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

What would you regard to be fair current gen prices?

Yeah... As they're not the market innovators they really obviously just pricing their cards off the closest Nvidia card minus 5/10%.

It is obvious. But they won't gain any market or mind share with that tactic. With that huge profit margin (according to madmossy) they really should be pricing more aggressively and getting more people to go Radeon.
 
It is obvious. But they won't gain any market or mind share with that tactic. With that huge profit margin (according to madmossy) they really should be pricing more aggressively and getting more people to go Radeon.

Yeah I mentioned the same in another thread. They could drop prices 25% and grab some serious market share and goodwill.
 
  • Like
Reactions: TNA
Seeing as it's estimated to cost them something in the range of £600 in materiels alone i can't see that happening.
Is this based in material quotes from the shortage or post shortage because you pay for stuff in advance in this industry.

If it's now then it would be loads cheaper as everything has crashed in price.

R&d is covered by the whole product range, not the 1 GPU SKU.
 
Last edited:
It's old data, but it gives insight into the rough costs of gpu manufacture and also gives you an idea of the profit margins before marketing etc, while the data is over a decade old and impossible to say how relevant it would be to current prices, it is still good to know. This was still when both AMD and NVidia were using monolithic gpu chips.

FYI the MSRP for the ATi HD 6970 was $369, the GTX 580 was $499. GTX 580 was ~18% faster than the HD 6970 but it cost 35% more.

iu
 
Last edited:
It is obvious. But they won't gain any market or mind share with that tactic. With that huge profit margin (according to madmossy) they really should be pricing more aggressively and getting more people to go Radeon.

I never said anything about profit, only cost reductions, while the gpu manufactur costs may decrease you still have to factor in packaging, shipping and marketing which all eat in the profits.
 
It's old data, but it gives insight into the rough costs of gpu manufactur and also gives you an idea of the profit margins, while the data is over a decade old and impossible to say how relevant it would be to current prices, it is still good to know. This was is still when both AMD and NVidia were using monolithic gpu chips.

FYI the MSRP for the ATi HD 6970 was $369, the GTX 580 was $499. GTX 580 was ~18% faster than the HD 6970 but it cost 35% more.

iu

Over 200 dollars for a top end GPU 10 years ago without R&D, packaging, shipping, drivers, etc. So £400+ certainly seems possible today.
 
It's old data, but it gives insight into the rough costs of gpu manufacture and also gives you an idea of the profit margins before marketing etc, while the data is over a decade old and impossible to say how relevant it would be to current prices, it is still good to know. This was still when both AMD and NVidia were using monolithic gpu chips.

FYI the MSRP for the ATi HD 6970 was $369, the GTX 580 was $499. GTX 580 was ~18% faster than the HD 6970 but it cost 35% more.

iu
But those where much larger chips, tech developed loads better when it comes to efficient yields now.

But those are actual halo products at the time, I didn't specify the halo product, even the 3080 had like 5 skus above it.

Stuff now is cheaper now then back then with many things, you couldn't get ssds at the prices we see them, same with cpus.
 
We know from interviews with Lisa Sui of AMD that chiplets cost about 40% less than a monolithic design, but that doesn't translate to a 40% cost reduction as additional hardware is required like an interposer, infinity fabric connection between the die etc so while the performance is still comparable as is the efficiency its cost per actually complete GPU may not differ that much.
But those where much larger chips, tech developed loads better when it comes to efficient yields now.

But those are actual halo products at the time, I didn't specify the halo product, even the 3080 had like 5 skus above it.

Stuff now is cheaper now then back then with many things, you couldn't get ssds at the prices we see them, same with cpus.

Things like SSDs will continue to come down in price as the cost of the manufactur decreases over time, an SSD does not need the latest lithographic methods, it would be an absolute waste of money to use 5nm on them over the 24-28nm it is currently.
 
We know from interviews with Lisa Sui of AMD that chiplets cost about 40% less than a monolithic design, but that doesn't translate to a 40% cost reduction as additional hardware is required like an interposer, infinity fabric connection between the die etc so while the performance is still comparable as is the efficiency its cost per actually complete GPU may not differ that much.


Things like SSDs will continue to come down in price as the cost of the manufactur decreases over time, an SSD does not need the latest lithographic methods, it would be an absolute waste of money to use 5nm on them over the 24-28nm it is currently.

I guess there is a saving from the reduced yield loss?
 
I never said anything about profit, only cost reductions, while the gpu manufactur costs may decrease you still have to factor in packaging, shipping and marketing which all eat in the profits.

What does cost reductions equal if the savings are not passed on to the end user? If AMD are able to make their GPU's cheaper they are either pocketing the extra as profit or passing it along to their customers.

If you expect the 4090 to be £650 and they are not benefiting from chiplets, what is it you think the 7900 XT should cost?
 
Is this based in material quotes from the shortage or post shortage because you pay for stuff in advance in this industry.

If it's now then it would be loads cheaper as everything has crashed in price.

R&d is covered by the whole product range, not the 1 GPU SKU.
It's based on estimated/reported/rumoured costs. Whether they're during the supposed shortages or after is largely irrelevant as my personal estimate, if you have the time you're invited to do your own as it's quiet the eye opener/interesting IMO, my personal estimate purposefully aired on what i thought would be considered reasonably big future price increases because i wanted to air on the high side so i didn't open myself up to accusations of being unfair. :)

If you're bored and have a spare 30min I'd love for others to have a go at estimating the BoM as having multiple independent people doing it will help to esabilish an average and highlight any outliers.
 
What does cost reductions equal if the savings are not passed on to the end user? If AMD are able to make their GPU's cheaper they are either pocketing the extra as profit or passing it along to their customers.

If you expect the 4090 to be £650 and they are not benefiting from chiplets, what is it you think the 7900 XT should cost?

The thread title asked "what would you regard as a fair price", I answered, it did not require me to justify why I think that price is fair. If your happy to spend 2 grand on a GPU lucky you. A vast majority of people are not.
 
We know from interviews with Lisa Sui of AMD that chiplets cost about 40% less than a monolithic design, but that doesn't translate to a 40% cost reduction as additional hardware is required like an interposer, infinity fabric connection between the die etc so while the performance is still comparable as is the efficiency its cost per actually complete GPU may not differ that much.


Things like SSDs will continue to come down in price as the cost of the manufactur decreases over time, an SSD does not need the latest lithographic methods, it would be an absolute waste of money to use 5nm on them over the 24-28nm it is currently.
Maybe not latest however you the latest trend with really any tech is making it sustainable and you usually try to find ways to make out cheaper then more expensive.

Bespoke limited runs are the outlier to this like the dual GPUs we had in the past.
 
It's based on estimated/reported/rumoured costs. Whether they're during the supposed shortages or after is largely irrelevant as my personal estimate, if you have the time you're invited to do your own as it's quiet the eye opener/interesting IMO, my personal estimate purposefully aired on what i thought would be considered reasonably big future price increases because i wanted to air on the high side so i didn't open myself up to accusations of being unfair. :)

If you're bored and have a spare 30min I'd love for others to have a go at estimating the BoM as having multiple independent people doing it will help to esabilish an average and highlight any outliers.
Just to state, it is very relevant when you consider pre , during and post shortage as the price difference between the 2 latter eras are huge.
 
The thread title asked "what would you regard as a fair price", I answered, it did not require me to justify why I think that price is fair. If your happy to spend 2 grand on a GPU lucky you. A vast majority of people are not.

In that case I think the 4090 should be £99 and packet of chips :D
 
Just to state, it is very relevant when you consider pre , during and post shortage as the price difference between the 2 latter eras are huge.
When i said it was largely irrelevant i was trying to convey that (IMO) that would be the reasonable maximum, yes prices have come down but you expect fluctuations in materiel costs and typically those contribute to increased profits. Like i said i aired on the high side, and then some, to be as fair to Nvidia as possible. So no one could start an argument with me about it being $50 under whatever it maybe costing.
 
Back
Top Bottom