Friday, August 21st 2026
AMD Confirms Radeon RX 9050 4 GB Has Smaller Infinity Cache, Lower Memory Bandwidth
AMD has updated its documentation page for the late-July launch of the Radeon RX 9050 graphics cards, which include two models: one with 4 GB and another with 8 GB of GDDR6 memory. Both cards feature 1,024 Streaming Processors across 16 Compute Units and can boost up to a 2.6 GHz clock speed. However, the lower-end GPU has only 4 GB of GDDR6 memory on a 64-bit bus, resulting in a memory bandwidth of 144 GB/s, which is significantly lower than the 8 GB version. The 8 GB version of the Radeon RX 9050 uses 8 GB of GDDR6 on a 128-bit bus, providing a memory bandwidth of 288 GB/s, doubling the throughput.
Another difference lies in the Infinity Cache within each GPU die. The 8 GB edition has 32 MB of Infinity Cache, while the 4 GB edition has only 16 MB. Initially, the listing for the 4 GB edition did not reveal the difference in Infinity Cache size, but AMD's documentation confirms it. Naturally, this will result in some performance differences, as larger cache buffers tend to improve performance significantly.AMD confirmed that the 4 GB version will be exclusively for OEMs and system integrators to include in their low-end PC configurations. For gamers, the 8 GB edition will be marketed as the entry-level RDNA 4 GPU and is expected to be priced at $279, making it the most affordable card AMD currently offers. There appear to be no Made by AMD (MBA) designs, as add-in board partners are handling the launch.
Sources:
AMD, via VideoCardz
Another difference lies in the Infinity Cache within each GPU die. The 8 GB edition has 32 MB of Infinity Cache, while the 4 GB edition has only 16 MB. Initially, the listing for the 4 GB edition did not reveal the difference in Infinity Cache size, but AMD's documentation confirms it. Naturally, this will result in some performance differences, as larger cache buffers tend to improve performance significantly.AMD confirmed that the 4 GB version will be exclusively for OEMs and system integrators to include in their low-end PC configurations. For gamers, the 8 GB edition will be marketed as the entry-level RDNA 4 GPU and is expected to be priced at $279, making it the most affordable card AMD currently offers. There appear to be no Made by AMD (MBA) designs, as add-in board partners are handling the launch.

118 Comments on AMD Confirms Radeon RX 9050 4 GB Has Smaller Infinity Cache, Lower Memory Bandwidth
Nvidia should have never done what they have done to advance GPU technology?
Maybe nvidia should have gave up with the TNT2 line and never introduced geforce gpus? They did everyone such a disservice by becoming successful....
Yes im critical of AMD, and wish ATI was still around I bet If ATI wasn't bought out they may have been more successful, but we will never know.
I started gaming on a voodoo1 card, and have seen how the graphics market has progressed over the years and it gets really tiring hearing how terrible nvidia is....
Crimes? Well now that you mention it, is a common citizen receiving the same justice with someone having more power?
It's so strange when someone by passes the part "RX 9050 is a bad product" and keeps replying like it was never posted.
It's so easy reading crap when someone posts crap. You just have to see it. Not rush to the rest of the post where that person probably tries to explain their thinking to find something to quote.
And someone with more "power" still doesn't use the defense of others have done this crime too.... Is there a point in there?Amen. Succinct, to the point.... Sans whataboutism.
Wow, what a nice logical jump. We are talking about misleading products and you connect this with Nvidia's advancements. You sure?
I have a TNT2. Or is it the original TNT? It's somewhere. Still functioning. Nvidia giving up, how are you even able to do such logical jumps?
ATI would be doing the same if they wanted to survive.
I also have two Voodoo cards, one is the Voodoo 2. But got those old cards the last couple years. Back then I was starting with a SiS6326 before moving to a Geforce 2 MX and then jumping to a real GPU, the GeForce 3 Ti200. Back then when you where changing GPU, you where completely changing level.
Or even the A750?
Haven't said this on TPU yet since this is my first post, but the RTX 5050 bugged me for the same reason, albeit being 130W; would probably need a blower fan. I have an oddly-specific use case for a single-slot 5050, and was debating trying to mod my own blower cooler before the prices got worse.
the 960 4GB sold for $250...
it's crazy how we are going full circle
They must be aware of the issues with the 1060 3/6GB, the 3050 6/8 GB, the 3060 8/12GB, the various DDR4 vs GDDR5 versions of otherwise identically-specced products. These ambiguous choices all resulted in the copious hand-wringing you see in this thread, yet they still decided to do this? WTF marketing department??
I have the "original" NV1 GPU, diamond edge 3d, ironically it was given to me in a system with a AMD 5x86 P75 system, but in reality it was a Trinity upgrade chip and it used to be a 486 system
G80 is 20 btw
Beyond that, the references to Nvidia and Intel—at least on my part—were necessary because some people seized the opportunity to create a completely false impression, as if AMD did something outrageous, unprecedented, something never done by anyone else before, something disastrous for the PC gamer. Creating false impressions for reasons known best to each individual only harms gamers and the broader PC gaming community. It is only right that everything is addressed in its true proportions.
The RX 9050 is a bad product that won't impact the market, given that it's an OEM product simply ordered from AMD. It won't affect the market, just like the market won't be affected by those 4GB DDR4 memory sticks we were reading about a while ago that are coming back to the market.
Other products shape what companies will offer us in the future to a far greater extent, and I have already made reference to those. I'm just sorry that some fail to see the danger in them.Oh, Diamond Edge. I think that SiS6326 card I had was a Diamond Edge model. And that's ironic. The worst 3D chip, on a graphics card made by the best brand of graphics cards in that period.
Prices really, what matters is where it lands vs the alternative, marketing msrp's not very useful lately.
It's here! And it sucks!
The 5050, 5060, 5060Ti, and the 5070, at the same time all get recommended over cheaper AMD cards, and the same people here act like Nvidia or Intel don't do the same.
Especially Intel, renaming degradation lake into something that it isn't.
www.techpowerup.com/348281/intel-to-re-launch-new-line-of-raptor-lake-refresh-lga1700-processors
Edit- Also interesting the tech news sites aren't losing their sh*t over the 5060Ti 16GB being over $700 now, a similar price the 5070Ti is supposed to be, with the 5070Ti being over $1,000 now.No this is the problem, a trillion dollar valuation doesn't mean AMD can compete on the same level as Nvidia, or Intel.
Nvidia and Intel still have a majority of market share, market valuation means nothing when AMD doesn't have a near monopoly on GPU's.It's simply because the 9060 is significantly more expensive to make, thanks to the AI greedflation affecting memory prices, and they are making an 8GB 9050.
OEM's requested the 4GB card, despite all of the outrage of AMD apparently destroying the gaming market by making a 4GB card that would show up in a few OEM prebuilts it doesn't exist for really any enthusiast user.
it all depends on where you put your metrics. power consumption over the time, low purchase price for gpu power, etc.
i doubt intel has fixed the high idle and power consumption of their graphic card. afaik years ago it was a silicion issue which can not be fixed. i did not saw any retests of the intel cards, no one has any interests in intel graphic cards anymore. no tech media post anything or any retest about those graphic cards
i am sure i will quick read any benchmark of sept 2026 which also covers or only covers intel old grpahic cards, in regards of power consumption, performance, etc.
www.intel.com/content/www/us/en/support/articles/000092564/graphics.html
With those settings the idle power is around 6 watts.
Let me ask you something, and I hope you answer honestly here. In a previous post, @Hecate91 mentioned the price increases on the RTX 5060 Ti 16GB and RTX 5070 Ti. If tomorrow AMD announces price hikes in the range of 30-40% —because that’s likely what they’d need to do to match the increases Nvidia has made so far— when you and others rush to that thread to write that AMD is a bad company punishing its customers with unreasonable price hikes, that it's hurting PC gaming, and that everyone should condemn its practices, do the rest of the forum members have the right to mention Nvidia? Can they use the name Nvidia? Will they be considered unreasonable if they mention Nvidia? Will referring to Nvidia be viewed as an attempt to justify AMD, or simply as a reference describing a specific situation that has developed in the market?