GPU prices — is the worst behind us ?

[DaedalusHelios]

So are NVIDIA releasing a lot of consumer GB203-400-A1/GB203-450-A! dies nor not?

And just because a FEW engineering samples was shipped is not the same a luanch level dies.

So who are "simplistic"?
Me, saying this will not happen anytime soon and a handfull of sample dies does not constitute a luanch preperation.
Or you sayingall the dies are ready, but something is blocking the launch.

If I am so simplistic, tell me how many dies have been shpped to partner, beacuse once again (a common pattern in the this) a lot of personal views are being posted, despite the facts being something else.

I explained it already, it is just the 3GB per chip density that is holding them back from a launch. They have the board designs already, GPU dies were produced, but just the chip scarcity of 3GB per chip having extremely high demand right now. The minimum amount per order required to essentially "get in line" is very high for high speed VRAM and Nvidia decoupled from providing the chips bundled with the GPU die sales. GPU board partners are therefore dealing with scarcity from these changes in the business arrangement and competing with Nvidia directly on VRAM purchasing.

 

[littlegreenmenfromhome]

I explained it already, it is just the 3GB per chip density that is holding them back from a launch. They have the board designs already, GPU dies were produced, but just the chip scarcity of 3GB per chip having extremely high demand right now. The minimum amount per order required to essentially "get in line" is very high for high speed VRAM and Nvidia decoupled from providing the chips bundled with the GPU die sales. GPU board partners are therefore dealing with scarcity from these changes in the business arrangement and competing with Nvidia directly on VRAM purchasing.

So how many dies shipped to board partners?

 

[DaedalusHelios]

So how many dies shipped to board partners?

Board partners were not worried about GPU die availability, just VRAM at 3GB chip density. Nvidia has the GPU dies available irregardless of how many were ordered so far. I hope not too many were ordered yet given they are unsure when they can get enough of that density VRAM to actually do a launch. GamersNexus and other sites were reporting on it. This is known by many, and now you know it too.

 

[littlegreenmenfromhome]

Board partners were not worried about GPU die availability, just VRAM at 3GB chip density. Nvidia has the GPU dies available irregardless of how many were ordered so far. I hope not too many were ordered yet given they are unsure when they can get enough of that density VRAM to actually do a launch. GamersNexus and other sites were reporting on it. This is known by many, and now you know it too.
View attachment 830439

So you have no clue how many dies went our rigth?

The thread is full of false claims:
"AMD marketshare went up" = Stayed flat at 8 for almost 2 years.
"AMD foucsed on on AI not consumers" = AMD has 4% of of that market.
"NVIDIA should sell consumer dies, they are just stallng = No clue about the number of alleged dies just sitting wasting $$$

 

[DaedalusHelios]

So you have no clue how many dies went our rigth?

The thread is full of false claims:
"AMD marketshare went up" = Stayed flat at 8 for almost 2 years.
"AMD foucsed on on AI not consumers" = AMD has 4% of of that market.
"NVIDIA should sell consumer dies, they are just stallng = No clue about the number of alleged dies just sitting wasting $$$

You have entered the straw man phase apparently. I made none of those claims. Under your own claims it takes massive time to produce GPU dies in a manufacturing run. If scarcity of 3GB chips happens after your production run and you end up not being able to finish a product(completed product on the market) on that design because the 3GB density chips get used on more increased revenue generating products.

It does not matter how many GPU dies are delivered if you can't get supply of the correct density of VRAM. You can get GPU dies from Nvidia. The issue is not the GPU die supply. The chief complaint is that Nvidia stopped offering VRAM bundled with the GPU die. It is really simple and you keep pretending to not understand. I wish Nvidia would bundle VRAM again and sell it with it so we can have a RTX 5080 Super to fill the market gap. Without Nvidia doing that, we will never get the RTX 5080 Super as far as I know.

But the AI race is eating the VRAM supply our dGPUs need(especially 3GB for high density VRAM designs). Unfortunately, the big memory producers(Samsung, Micron, and SK Hynix) are not building up to create enough VRAM to meet the need. They might know something on the horizon that we do not as to why.

Some say a constricted supply creates more demand and it allows for a maximum market price to sell for datacenter products. It is probably a combination of factors as it always has been in production environments that I have been in over the years.

 

[littlegreenmenfromhome]

You have entered the straw man phase apparently. I made none of those claims. Under your own claims it takes massive time to produce GPU dies in a manufacturing run. If scarcity of 3GB chips happens after your production run and you end up not being able to finish a product(completed product on the market) on that design because the 3GB density chips get used on more increased revenue generating products.

It does not matter how many GPU dies are delivered if you can't get supply of the correct density of VRAM. You can get GPU dies from Nvidia. The issue is not the GPU die supply. The chief complaint is that Nvidia stopped offering VRAM bundled with the GPU die. It is really simple and you keep pretending to not understand. I wish Nvidia would bundle VRAM again and sell it with it so we can have a RTX 5080 Super to fill the market gap. Without Nvidia doing that, we will never get the RTX 5080 Super as far as I know.

But the AI race is eating the VRAM supply our dGPUs need(especially 3GB for high density VRAM designs). Unfortunately, the big memory producers(Samsung, Micron, and SK Hynix) are not building up to create enough VRAM to meet the need. They might know something on the horizon that we do not as to why.

Some say a constricted supply creates more demand and it allows for a maximum market price to sell for datacenter products. It is probably a combination of factors as it always has been in production environments that I have been in over the years.

I did not attritbute those claims to you, I stated this thread had those claims.
The devil is in the details.

And there is no need for a 50 series Super release as AMD not only has fumbled the ball, the dropped it off field all together.
And the 8% sales numbers proves this.

It also shows that the diffenrt stance on forums was false:
- Raytracing doesn't matter!!!
- Upscaling doesn't matter!!!
- Halo card doesn't matter!!!
- FPS per $$$ matter!!!

But in reality that belief has lead to 8% sales numbers.
So I will enjoy my RTX 4090 with all the additional features unmatched by AMD or far better features than AMD has to offer, unbested by anything they (AMD) has to offer until I build a new PC with a RTX 6090 and the repeat the cycle.
Like 90% of people that buys and NVIDIA consumer GPU.
I will also keep building server STUFFED with NVIDIA GPU's and nota single AMD GPU.
Like pratically the entire industry.

And I do not feel sorry for using GPU's so consumers can not get any cheap GPU, as my job is to take care of shareholder and not whiney gamers.

 

[Marees]

A RTX 5080 Super 24GB for $2k launch would sell fine given current market conditions. I think Nvidia is too preoccupied with making ludicrous money in the datacenter market to do bother with it.
The space between RTX 5080 and RTX 5090 is widening in street price.

RTX 5080 24GB Perhaps?​

There’s also rumor about potential 24GB version of RTX 5080, which comes from hongxing2020, also very reliable NVIDIA hardware leaker. NVIDIA has reportedly notified a factory about the end of RTX 5090 production. The same message suggests that a GeForce RTX 5080 24GB will replace the current 16GB version and become the highest-end GeForce card.

https://videocardz.com/newz/nvidia-...-gb202-gpus-to-be-reserved-for-rtx-pro-series

 

[DaedalusHelios]

RTX 5080 24GB Perhaps?​

There’s also rumor about potential 24GB version of RTX 5080, which comes from hongxing2020, also very reliable NVIDIA hardware leaker. NVIDIA has reportedly notified a factory about the end of RTX 5090 production. The same message suggests that a GeForce RTX 5080 24GB will replace the current 16GB version and become the highest-end GeForce card.

https://videocardz.com/newz/nvidia-...-gb202-gpus-to-be-reserved-for-rtx-pro-series

There is the China market which is separate, so that may be a Chinese market move to constrict capability as a tightening on their AI space build out. The RTX 5090 is slightly different over there with the v2. The RTX 5080 Super or perhaps a ti branding might eventually arrive on the market. Extra VRAM on the current vanilla RTX 5080 wouldn't help with much except AI models mainly. It would future proof the RTX 5080 some though.
The issue is still VRAM until that can be solved we might not see much high VRAM cards in actual full launch volume. Unless this other RTX 5080 (24GB) gets a very high price tag.

 

[DooKey]

Again, both of you are wrong. If AMD was only Radeon GPUs and no other products, then I would agree with you, but they have a robust CPU side that is kicking Intel's butt and eating away at their dominance every quarter in Enterprise spaces and they are competing with Nvidia in the AI space, probably better than they are in the consumer discrete GPU space if comparing. They are also selling tens of millions of custom chips for Playstation, Xbox, handhelds, Steam hardware, etc. 10% in discrete GPU market share is not that bad when they are selling very well in other areas because they are a diversified company, not a one trick pony. Both Nvidia and ATI/AMD have expanded beyond being one trick pony discrete GPU designers because there isn't that much money in discrete GPUs, not compared to the deep pockets of big tech companies all fighting each other for chips.

I have been very specific about the AMD discrete video card business. That's what sucks. AMD as a whole is a great company and doing very well.

 

[erek]

“Starting This Month, AMD Implements Another Memory Price Hike for GDDR6-Based GPUs, But It's Not the Only Thing That Will Increase GPU Prices​

Memory prices must have climbed once again recently, which is why we are now hearing that AMD has implemented another price hike for memory. Board Channels reports that AMD has already raised the prices for its GDDR6 memory for its board partners, which means board partners will now have no other option but to pass the costs on to consumers.”

https://wccftech.com/amd-reportedly-raised-gddr6-memory-prices-for-its-partners/

 

[El_Capitan]

RTX 5080 24GB Perhaps?​

There’s also rumor about potential 24GB version of RTX 5080, which comes from hongxing2020, also very reliable NVIDIA hardware leaker. NVIDIA has reportedly notified a factory about the end of RTX 5090 production. The same message suggests that a GeForce RTX 5080 24GB will replace the current 16GB version and become the highest-end GeForce card.

https://videocardz.com/newz/nvidia-...-gb202-gpus-to-be-reserved-for-rtx-pro-series

I don't know if they'd do it, but it'll be really pricey. Probably shoot up to the $5K mark despite the MSRP. More if it ends up being a RTX 5080 Super 24GB.