Wednesday, February 12th 2025

AMD Radeon RX 9070 XT Could Get a 32 GB GDDR6 Upgrade

AMD's Radeon RX 9000 series GPUs are expected to come with up to 16 GB of GDDR6 memory. However, AMD is reportedly expanding its RX 9070 lineup with a new 32 GB variant, according to sources on Chiphell. The card, speculatively called the RX 9070 XT 32 GB, is slated for release at the end of Q2 2025. The current GDDR6 memory modules used in GPUs carry a capacity of 2 GB per module only, meaning that a design with 32 GB of VRAM would require as many as 16 memory modules on a single card. No 2 GB+ GDDR6 memory modules are available, meaning that the design would require memory module installation on both the front and back of the PCB. Consumers GPUs are not known for this, but it is a possibility with workstation/prosumer grade GPUs employing this engineering tactic to boost capacity,

While we don't have information on the GPU architecture, discussions point to potential modifications of the existing Navi 48 silicon. This release is positioned as a gaming card rather than a workstation-class Radeon PRO 9000 series product. AMD appears to be targeting gamers interested in running AI workloads, which typically require massive VRAM amounts to run locally. Additionally, investing in a GPU with a big VRAM capacity is essentially "future-proofing" for gamers who plan to keep their cards for longer, as recent games have been spiking VRAM usage by a large margin. The combination of gaming and AI workloads may have made AMD reconsider some of its product offerings, potentially giving us the Radeon RX 9070 XT 32 GB SKU. We have to wait for the Q2 to start, and we can expect more details by then.
Sources: Chiphell, via VideoCardz
Add your own comment

42 Comments on AMD Radeon RX 9070 XT Could Get a 32 GB GDDR6 Upgrade

#1
Quicks
Well at least AMD is not stringy with their VRAM.

Still would like AMD to try a 350 - 400W card with 24GB - 32GB VRAM and + 30% - 50% of the cores of the 9070XT and see what's the best performance they can get out of it.

Price is king, they just need to price it right and people will buy it.
Posted on Reply
#2
TPUnique
I got interested for a nanosecond, contemplating delaying my GPU purchase for something that could be very useful for ML tasks. Then I remembered that AMD was still lagging in the ML department, so I'll try to order a 5070Ti on the 20/02 as planned.

I have a 2nd computer to replace, so I'll see next year what Intel and AMD have in store by then.
Posted on Reply
#3
kondamin
That would make the card a whole lot more interesting.
cant they stack 2 ram chips without making it hbm
Posted on Reply
#5
londiste
Clamshell usually. Pro and datacenter cards have this often enough. Rarer in consumer space. 3090 and 4060Ti 16GB come to mind from recent times.
Either way, a 32GB 9070XT would be immediately bought up for AI. Given relative lack of performance improvement for gaming the SKU probably does not make much sense.
Posted on Reply
#6
W1zzard
TPUniqueI got interested for a nanosecond, [...] could be very useful for ML tasks. Then I remembered that AMD was still lagging in the ML department
Exactly what just happened to me.

No doubt, AMD is trying to sell this to AI people, but the developer experience is so terrible compared to CUDA. Unless you're buying a few hundred and can spread the developer cost among these units I doubt it will make a big difference
Posted on Reply
#7
Chomiq
I guess it will have its use for ML applications. That being said I would still prefer 9090 XTX with 32 GB of vram.
Posted on Reply
#8
Nhonho
CUDA
I would like to buy Radeon VGAs, but I am forced to buy Geforce VGAs because there are more software compatible with CUDA and also because the quality of videos encoded by Geforce NVENC have better image quality.
Posted on Reply
#9
NoneRain
doubt, but it would be awesome.
Posted on Reply
#10
InVasMani
Pricing would be my concern, but looks appealing otherwise.
Posted on Reply
#11
Onasi
Useless for gaming, kind of useful as a bone thrown to the AI crowd, but, as noted above, it would have to be priced very aggressively to compensate for the lack of NV amenities like CUDA.
Posted on Reply
#12
Sound_Card
I called it lol. I knew this was in the works somehow.
OnasiUseless for gaming, kind of useful as a bone thrown to the AI crowd, but, as noted above, it would have to be priced very aggressively to compensate for the lack of NV amenities like CUDA.
People are starting to work around CUDA, and you don't even need it to build your own LLM. A 32Gb 9070xt for half the price of a 5090 would be a major deal for the AI people.
Posted on Reply
#13
csendesmark
W1zzardExactly what just happened to me.

No doubt, AMD is trying to sell this to AI people, but the developer experience is so terrible compared to CUDA. Unless you're buying a few hundred and can spread the developer cost among these units I doubt it will make a big difference
Try to run a 30GB model on a fancy 24 or 16GB RTX with CUDA and compare the experience against a Radeon with 32GB VRAM.
VRAM is valuable real-estate!
Posted on Reply
#14
AusWolf
I don't see the point. Sure, Nvidia's 12 GB lineup is anaemic as f*ck, but this is unnecessary in this class of GPU (unless it's targeted at purely AI purposes, I guess).
Posted on Reply
#15
csendesmark
AusWolfI don't see the point. Sure, Nvidia's 12 GB lineup is anaemic as f*ck, but this is unnecessary in this class of GPU (unless it's targeted at purely AI purposes, I guess).
Gaming wise, there is little point moving above 12~16GB VRAM in 2025Q1.
Posted on Reply
#16
Octopuss
AleksandarKHowever, AMD is reportedly expanding its RX 9070 lineup with a new 32 GB variant
What's the bloody point?
Posted on Reply
#17
Sound_Card
csendesmarkTry to run a 30GB model on a fancy 24 or 16GB RTX with CUDA and compare the experience against a Radeon with 32GB VRAM.
VRAM is valuable real-estate!
Ya, this will sell like hotcakes and perhaps lighten the demand on gaming cards.
AusWolfI don't see the point. Sure, Nvidia's 12 GB lineup is anaemic as f*ck, but this is unnecessary in this class of GPU (unless it's targeted at purely AI purposes, I guess).
32GB would be purely AI.
Posted on Reply
#18
Jtuck9
If it's a similar architecture hopefully this bodes well for the software suite on the other cards
Posted on Reply
#19
PaddieMayne
Fantastic news i hope they do a 24gb model.
Posted on Reply
#20
Krit
32GB is not a gaming gpu it will be artificially expensive. Mainly for work or for gamers with low iq 16GB is more than enough for gaming.
Posted on Reply
#21
AusWolf
PaddieMayneFantastic news i hope they do a 24gb model.
If they did that with 3 GB memory modules while keeping the Navi 48 core intact, that would be awesome (still not necessary for gaming, but makes a bit more sense than 32 GB).
Posted on Reply
#22
Legacy-ZA
Krit32GB is not a gaming gpu it will be artificially expensive. Mainly for work or for gamers with low iq 16GB is more than enough for gaming.
Not necessarily; remember, nGreedia moved to GDDR7, there is going to be a lot of GDDR6/X chips that still need to be sold.
Posted on Reply
#24
3valatzy
A bad decision - over engineered.

The chip is too slow to make use of so large VRAM pool, the bottleneck will be in the limited memory bandwidth, shaders count and generally poor micro-architecture performance / read IPC.
Posted on Reply
#25
yfn_ratchet
The idea's cute, but honestly I don't see a reason to reach for more than 24GB to a card for my usual use cases. 12GB is the passing grade for me as-is, running most of the tasks I want without too much trouble. If they can offer me a card with the same memory and average performance of a 7800 XT but shave $80 off the price tag, I'm sold. Nothing else is as appealing as that.
Posted on Reply
Add your own comment
Feb 12th, 2025 14:14 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts