Nvidia shows signs in [2023]

  • Thread starter Deleted member 2197
  • Start date
Status
Not open for further replies.
I understand there are limitations. 1000% profit margins have a way making these limitations go away.

Sorry I messed up my reply and it got split in two.

That's 1000% profits (which seems to be mathematically incorrect if they are using an avg price estimate of $25k-30k given their BoM estimate) relative to the BoM (unclear how that was estimated, BoM is also not a good indicator of even just unit cost) which is not the same as 1000% profit margins per unit.
 
That's 1000% profits (which seems to be mathematically incorrect if they are using an avg price estimate of $25k-30k given their BoM estimate) relative to the BoM (unclear how that was estimated, BoM is also not a good indicator of even just unit cost) which is not the same as 1000% profit margins per unit.
I think I meant "markup" rather than "margin". The markup does seem like would be around 1000%.
 
I think NVIDIA will continue to not focus on GPUs until the HPC bubble bursts. If that ever happens it could be a long ways off. They've already priced Ada Lovelace in a manner where they won't sell very well so they won't have to make too many of them. So I'm not really saying this will happen. I'm saying it is happening.

In what way is Nvidia not focused on GPUs? They’re clearly still spending a lot of money on gaming hardware and software R&D. Low prices aren’t an indicator of “focus”.
 
They've already priced Ada Lovelace in a manner where they won't sell very well so they won't have to make too many of them. So I'm not really saying this will happen. I'm saying it is happening.
No competition to hold them off, they are pricing them high and keeping them high because people are still buying them in sufficient numbers, on the other hand AMD keeps lowering prices for their GPUs further and further, and despite lower prices, AMD's share is shrinking further and further. So NVIDIA is practically unchallenged, and is under no market pressure to drop prices yet.
 
In what way is Nvidia not focused on GPUs? They’re clearly still spending a lot of money on gaming hardware and software R&D. Low prices aren’t an indicator of “focus”.
I think they'll continue the strong R&D although they're so far ahead they really don't have to. But I don't think they'll produce a lot of GPUs or care too much how well they are selling. It's in their interest to keep prices high and volume low.
 
No competition to hold them off, they are pricing them high and keeping them high because people are still buying them in sufficient numbers, on the other hand AMD keeps lowering prices for their GPUs further and further, and despite lower prices, AMD's share is shrinking further and further. So NVIDIA is practically unchallenged, and is under no market pressure to drop prices yet.
Yes it's a shame that AMD is unable to capitalize on this.

"Selling in sufficient numbers" is technically true. It's selling rather poorly (especially given there is no competition) which is fine with NVIDIA.
 
But I don't think they'll produce a lot of GPUs or care too much how well they are selling.
You care about how something is selling by making sure that it's selling with a good enough margin for it to exist. This has nothing to do with how much this is selling and everything to do with how good this is in comparison to competition.
 
I think NVIDIA will continue to not focus on GPUs until the HPC bubble bursts. If that ever happens it could be a long ways off. They've already priced Ada Lovelace in a manner where they won't sell very well so they won't have to make too many of them. So I'm not really saying this will happen. I'm saying it is happening.

Or they raised prices on GPU because they have no competition, saw during the crypto bubble that people were ready to pay more, and they will make more money this way.
 
And they are super smart in exploiting the market. Case in point rtx 4000 family. With vram so limited, ML crowds will be forced to grab 4060 ti 16GB or jump directly to 4080.

RT Gamers on the other hand, need to jump to 4080 if they want to avoid getting vram limited or weird stutters / performance drop due to 12GB vram on 4070 ti. 4060 ti 16GB got enough vram but is already sweating hard in RT.

These two markets have no AMD alternative.
 
And they are super smart in exploiting the market. Case in point rtx 4000 family. With vram so limited, ML crowds will be forced to grab 4060 ti 16GB or jump directly to 4080.

RT Gamers on the other hand, need to jump to 4080 if they want to avoid getting vram limited or weird stutters / performance drop due to 12GB vram on 4070 ti. 4060 ti 16GB got enough vram but is already sweating hard in RT.

These two markets have no AMD alternative.
That's where Intel is trying to be imo. Good RT, enough vram, decent price ?
 
Or they raised prices on GPU because they have no competition, saw during the crypto bubble that people were ready to pay more, and they will make more money this way.
It is a confluence of things that gives us the $800 4070Ti. Aftereffects of the cryptoboom, lack of competition (RIP RDNA3 :( ), higher manufacturing costs, and now the AI boom. That last one is crucial because it puts NVIDIA in a position where they don't have to worry too much if their gaming GPUs aren't selling very well.
 
I found out after installing my RTX4070 that it is BIOS limited to 200W which is a bummer. I know it's limited by this because the GPU power is always right at 200W. I'd considered undervolting it to see if it would clock higher. Would be nice to have to option of putting a different BIOS on it, though I seriously doubt I'll ever take that risk. AFAIK it doesn't have dual BIOS or anything so if it goes wrong I'm screwed.

Also 200W is a pathetic limitation. This thing is a triple fan chonker that could probably handle 300+W. Apparently so can the 8pin PCIe cable on my PSU because it has 2 connectors on it.
 
Last edited:
Stock clocked cards are limited at stock power usually. If you want to be able to set a higher power ceiling then that's why factory OCed cards cost more than the stock clocked ones.

But tbh GPU overclocking these days nets such minuscule results that paying for them anything above the MSRP doesn't worth it usually.
I didn't know that. TBH I don't pay any attention to factory clocks since as you say it makes so little difference. I was just curious how much more performance I could get if I increased the power limit. I'm quite certain this behemoth of a card can handle well over 200W.

Anyway I ragepurchased the card after discovering the 6700XT I'd purchased was still defective after an RMA, and the only factor in determining which model 4070 I got was which one had 1 day shipping :)
 
I found out after installing my RTX4070 that it is BIOS limited to 200W which is a bummer. I know it's limited by this because the GPU power is always right at 200W. I'd considered undervolting it to see if it would clock higher. Would be nice to have to option of putting a different BIOS on it, though I seriously doubt I'll ever take that risk. AFAIK it doesn't have dual BIOS or anything so if it goes wrong I'm screwed.

Also 200W is a pathetic limitation. This thing is a triple fan chonker that could probably handle 300+W. Apparently so can the 8pin PCIe cable on my PSU because it has 2 connectors on it.
You can always try flashing with another vendors bios which has much higher power limit. As long as you stick with official bios you can't brick an nvidia 2000/3000/4000 card. What can another vendor bios mess up are fan profiles and dp/hdmi outputs (some brands have more hdmi or dp output than others). First you should try a bios from the same vendor of your card. OC versions are basically the same card that usually have higher power limits.
 
Status
Not open for further replies.
Back
Top