Friday, October 14th 2022

NVIDIA Cancels GeForce RTX 4080 12GB, To Relaunch it With a Different Name

NVIDIA has decided to cancel the November 2022 launch of the GeForce RTX 4080 12 GB. The company will relaunch the card under a different name, though it didn't announce the replacement name just yet. The naming of the RTX 4080 12 GB was cause for much controversy. With the RTX 40-series "Ada," NVIDIA debuted three SKUs—the already launched RTX 4090 which is in stores right now; the RTX 4080 16 GB, and the RTX 4080 12 GB. Memory size notwithstanding, the RTX 4080 12 GB is a vastly different graphics card from the RTX 4080 16 GB.

The RTX 4080 12 GB and RTX 4080 16 GB didn't even share the same silicon. While the 16 GB model is based on the larger "AD103" silicon, has 9,728 CUDA cores, and a 256-bit wide GDDR6X memory bus; the RTX 4080 12 GB is based on the smaller "AD104" silicon, has just 7,680 CUDA cores (21% fewer CUDA cores); and a meager 192-bit wide GDDR6X memory bus. This had the potential to confuse buyers, especially given the $900 price. With criticism spanning not just social media but also bad press, NVIDIA decided to pull the plug on the RTX 4080 12 GB. The company will likely re-brand it as a successor to the RTX 3070 Ti, although then it will have a hard time justifying its $900 price-tag. The RTX 4080 16 GB, however, is on track for a November 16 availability date, with a baseline price of $1,200.
Source: NVIDIA
Add your own comment

423 Comments on NVIDIA Cancels GeForce RTX 4080 12GB, To Relaunch it With a Different Name

#76
trsttte
cvaldesAs I pointed out in another discussion thread, it makes zero sense trying to compare model numbers between GeForce generations.

This announcement pounds that concept home with brutal clarity.

Prospective customers simply need to assess the various models in the product generation and decide whether or not each SKU's features are worthy of the price being asked while disregarding the model number printed on the card and the retail packaging.

NVIDIA changes their idea of what each model number represents. Essentially, a --70 card is simply a product between a --60 model and an --80 model from the same generation's product stack. Comparing the 1070, 2070, 3070, and 4070 lacks relevancy because of NVIDIA's constant reinterpretation of their model numbers.
My answer has nothing to do with older model numbers, simply that the card bellow the (real) 4080 can't have just 8gb of vram, much less when the 4080 takes 16gb.

Nvidia continues to push their rtx narrative and also dlss and 4k gaming, 8gb is simply not enough on a high end card. The 3070 with 8bg and even the 3080 with 10gb were already starved by vram, the 4070 would be even more so.
Posted on Reply
#77
ModEl4
mahirzukic2I did. Why is it's box so huge? :kookoo:
It seems Galax is taking the GB RX6600 Eagle approach? (But probably just a trial placeholder box used just for the show)

Posted on Reply
#78
R0H1T
Fool me once shame on you, fool me twice shame on you, fool me thrice ~ ST eff up & take my mortgage JHH :roll:

Nvidia diehards probably right now :peace:

Posted on Reply
#79
cvaldes
trsttteMy answer has nothing to do with older model numbers, simply that the card bellow the (real) 4080 can't have just 8gb of vram, much less when the 4080 takes 16gb. Nvidia continues to push their rtx narrative and also dlss and 4k gaming, 8gb is simply not enough on a high end card. The 3070 with 8bg and even the 3080 with 10gb were already starved by vram, the 4070 would be even more so.
I never said that releasing a 4070 with 8GB VRAM is best. However I think that's what NVIDIA will do.

Two different concepts.

I admire NVIDIA for their engineering, both their hardware and their software (and I'm not even a developer).

I have a separate opinion about how they run their business however I choose not to provide any more details today. I've already said enough about the topic this week.
Posted on Reply
#80
ARF
cvaldesI never said that releasing a 4070 with 8GB VRAM is best. However I think that's what NVIDIA will do.

Two different concepts.
They will have to explain. Because 8 GB is a limitation and a deal-breaker for potential buyers with their own thinking process.
This is like launching the RTX 3080 with 8 GB when even 10 GB led to cheating by nvidia, lowering texture resolution and image quality to maintain framerates.
Posted on Reply
#83
cvaldes
ARFThey will have to explain. Because 8 GB is a limitation and a deal-breaker for potential buyers with their own thinking process.
This is like launching the RTX 3080 with 8 GB when even 10 GB led to cheating by nvidia, lowering texture resolution and image quality to maintain framerates.
Well, NVIDIA sold a ton of 3080 10GB cards before they released the 3080 Ti 12GB and eventual 3080 12GB cards.

I have a 3060 Ti 8GB card and for 1440p I don't recall running into VRAM limits apart from maybe one or two titles. 8GB VRAM is a factor for 4K gaming.

Clearly they could debut a 4070 8GB card and later release a 10GB or 12GB variant. It's not like NVIDIA is forced to pick a specific VRAM amount for any given GPU. And consumers will likely buy whatever is offered regardless.
Posted on Reply
#85
Ruru
Trepie
Nvidia listened to customers for once? This is incredible.
cvaldesSo clearly they could debut a 4070 8GB card and later release a 10GB or 12GB variant.
The 10GB variant then would be with 160-bit memory bus. Having a memory bus that narrow in a xx70 card would be beyond insane.
Posted on Reply
#86
ARF
cvaldesWell, NVIDIA sold a ton of 3080 10GB cards before they released the 3080 Ti 12GB and eventual 3080 12GB cards.

So clearly they could debut a 4070 8GB card and later release a 10GB or 12GB variant. It's not like NVIDIA is forced to pick a specific VRAM amount for any given GPU. And consumers will likely buy whatever is offered regardless.
I know and it's bizarre. Next is they can try a 4070 4 GB. I guess it will also sell a ton.
Posted on Reply
#87
cvaldes
LenneThe 10GB variant then would be with 160-bit memory bus. Having a memory bus that narrow in a xx70 card would be beyond insane.
I wouldn't count anything out from them. After all, they were ready to launch this 3080 12GB card at an abominable price.
Posted on Reply
#88
Ruru
Trepie
cvaldesI wouldn't count anything out from them. After all, they were ready to launch this 3080 12GB card at an abominable price.
True. But a card with 160-bit bus sold with a high-end price is beyond any sense. On the other hand, it's Ngreedia..
Posted on Reply
#90
Sisyphus
A product name change is without any importance. The price depends on demand and competition. AMD is late, nVidia has about 6 months to cash in with its AD102 and AD103. Then we will see, what price AMD calls up for GPUs with comparable performance. A secondary impact will arise from the recession, together with inflation. Difficult to forecast the impact on the chip market, pricing.
Posted on Reply
#91
AnotherReader
SisyphusA product name change is without any importance. The price depends on demand and competition. AMD is late, nVidia has at least 6 months to cash in with its AD102 and AD103. Then we will see, what price AMD calls up for GPUs with about the same performance. A secondary impact will arise from the recession, together with inflation. Difficult to forecast the impact on the chip market, pricing.
Names have meaning. For a long time, Nvidia's top tier GPU was the x80. It changed to x80 Ti when AMD surprised them with Hawaii, and remained so until Ampere. I would also like to know your source regarding a delay in RDNA3.
Posted on Reply
#92
medi01
SisyphusA product name change is without any importance.
Wrong.

Per NVs own discovery (was in their slides) they've shockingly discovered that customers stick with series (e.g. 970 => 1070 => 2070) rather than sticking with the price bracket.

4080 losing "80" would be a major hit.
Posted on Reply
#93
cvaldes
LenneTrue. But a card with 160-bit bus sold with a high-end price is beyond any sense. On the other hand, it's Ngreedia..
Due to the relationship between VRAM memory capacity and memory bus width, the 40 Series has created some tricky choices.

Most consumers will just look at the amount of VRAM printed on the retail box and figure that 12GB is better than 10GB even if the bus width is narrower on the former.
Posted on Reply
#94
AnotherReader
ZetZetIt beats a 3090 and a 3080. 4070 would have been definitely justified. If they launched it with that name right away.
We'll see if it beats a 3090. From Nvidia's own benchmarks, in at least one game, it's closer to the 3080 than the 3090 Ti. The grey bars are the native rendering, i.e. sans DLSS, numbers.

Posted on Reply
#95
pavle
Maybe, just maybe the chinese looking fellow on pictures a few posts back just wanted to give the people something to chew on to not observe the real problem: the co-existence of RTX 30xx and 40xx series on the market with the latter at much higher price just so nvidiots can prop up their stocks?
The nvidia excrement show continues...
Posted on Reply
#96
ZetZet
AnotherReaderWe'll see if it beats a 3090. From Nvidia's own benchmarks, in at least one game, it's closer to the 3080 than the 3090 Ti. The grey bars are the native rendering, i.e. sans DLSS, numbers.
Okay, maybe it doesn't, it's still within the same performance bracket. Ofc the 900 USD price is insane if it doesn't beat the 3090... :)
Posted on Reply
#97
Sisyphus
AnotherReaderNames have meaning. For a long time, Nvidia's top tier GPU was the x80. It changed to x80 Ti when AMD surprised them with Hawaii, and remained so until Ampere. I would also like to know your source regarding a delay in RDNA3.
Names are PR. Consumers need to inform themselves about the specs. I will not defend people, who buy their GPUs based on naming or nicely colored packaging.
The delay is an estimation. Fill in whatever timespan you like.
Posted on Reply
#98
ARF
SisyphusNames are PR. Consumers need to inform themselves about the specs. I will not defend people, who buy their GPUs based on naming or nicely colored packaging.
There are lawyers who will defend them, though. Because giving a 4080 brand to a 12-GB-AD104-card is like claiming a C-class mercedes to be S-class.
Posted on Reply
#99
Dimitriman
LenneNvidia listened to customers for once? This is incredible.


The 10GB variant then would be with 160-bit memory bus. Having a memory bus that narrow in a xx70 card would be beyond insane.
More like, they got caught with their pants down and are now acting desperate to fix things:

- Pretending there is a scalper/miner shortage on 4090
- "Loyalty program" for buying FE 4090, only for current nvidia owners
- Killing the LHR from 30 series
- And now, cancelling the dumb 4080 12GB.

More "unusual moves" to be expected as Jensen continues to wake up from his hubris enduced coma.
Posted on Reply
#100
ModEl4
Anyone having access (I don't) in DRAMexchange can see the GDDR6 spot price differences and have an indication.(just an indication!)
Spot price for 8Gbit GDDR6 is even lower than GDDR5 and there is a chance the 16Gb GDDR6 to be only around 1.5X the 8Gb price.(and Nvidia probably buys lower than the below spot session lows...)
So for example the actual difference between 8 8Gbit GDDR6 ICs (256bit bus case/8GB total) and 6 16Gbit GDDR6 ICs (192bit bus case/12GB) can even be only $5 total depending the 16Gbit GDDR6 IC price.
There is a reason for example ARC A770 8GB has only $20 difference with 16GB version and this could be it (8 x $5 for 8Gbit ICs vs 8 x $7.5 for 16Gbit ICs) (i don't buy Intel just want to push 16GB version)
Anyone with access can enlight us!


Posted on Reply
Add your own comment
Nov 14th, 2026 01:24 CST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts