Tachyonism
Junior Member
- Jan 24, 2026
- 23
- 28
- 46
The same market realities causing Nvidia to delay consumer crap would also cause AMD to delay consumer crap.
So yes, it will go without Radeon trying anything aggressive.
Panino Manino
Golden Member
- Jan 28, 2017
- 1,156
- 1,401
- 136
Isn't there any way put more cards on the shelves? Don't need to be something next gen, just "something".
No, they already ordered the amount of chips and even if they wanted to increase GPU production then from where would they acquire the GDDR6? That's rhetorical, it's all sold through for 2026. And possibly even beyond.
They are also likely to switch some N48/N44 allocation (any excess production they can't sell to partners with VRAM to pair them with) to Epyc based on reported market conditions (where there are shortages of the CPUs themselves).
Last edited:
adroc_thurston
Diamond Member
- Jul 2, 2023
- 8,570
- 11,297
- 106
AS I mentioned numerous times, AMD doesn't have the manpower to keep up in all the markets it is involved in. Which sides are going to take a cut I propositioned and it appears to be all of them aside from data center.
To make instinct a year cadence product, would have taken alot of resources from everywhere. Especially considering these products are made on the newest node. And extra highers AMD has brought in likely has to focus on software which AMD has neglected over the years and support for business to business partners which AMD has neglected over the years compared to the competition. AMD doesn't have the resources to produce simultaneous chips on the newest process year after year. Their net profit just recently reached over a billion dollars.
The consumer market is stagnant in terms of revenue and with the cost of RAM as well as the supply, will actually shrink. Everyone complaining need to understand that releasing products in the consumer market during this time is going to result in way less return in investment. These decisions where to spend much and where to put resources was made years ago and likely based on the potential growth of the CPU market and GPU market. The GPU market would have made the most sense particularly the consumer side since the market has been stuck at 4 billion quarterly for a while now and fighting Nvidia is alot harder than fighting Intel in the consumer space. If I was AMD and wanted to make the most money and getting the most return on investment, i would spend as little as possible on consumer graphics as the dollar per mm2 of die is the worst and as the current marketshare shows, getting marketshare from Nvidia is brutally difficult.
To make instinct a year cadence product, would have taken alot of resources from everywhere. Especially considering these products are made on the newest node. And extra highers AMD has brought in likely has to focus on software which AMD has neglected over the years and support for business to business partners which AMD has neglected over the years compared to the competition. AMD doesn't have the resources to produce simultaneous chips on the newest process year after year. Their net profit just recently reached over a billion dollars.
The consumer market is stagnant in terms of revenue and with the cost of RAM as well as the supply, will actually shrink. Everyone complaining need to understand that releasing products in the consumer market during this time is going to result in way less return in investment. These decisions where to spend much and where to put resources was made years ago and likely based on the potential growth of the CPU market and GPU market. The GPU market would have made the most sense particularly the consumer side since the market has been stuck at 4 billion quarterly for a while now and fighting Nvidia is alot harder than fighting Intel in the consumer space. If I was AMD and wanted to make the most money and getting the most return on investment, i would spend as little as possible on consumer graphics as the dollar per mm2 of die is the worst and as the current marketshare shows, getting marketshare from Nvidia is brutally difficult.
adroc_thurston
Diamond Member
- Jul 2, 2023
- 8,570
- 11,297
- 106
They do.
They just don't wanna focus on markets that are either miserable to penetrate (client dGFX) or are being margin squeezed from multiple directions (a solid chunk of PC).
AMD has to do exactly two things regarding client graphics:
With that, if AMD wants to build a GPU, they have all required building blocks.
From that point on, the rest are management and business decisions. And this is exactly the reason why the rumored RDNA5 ATx are looking like they do. Use each chip for multiple purposes.
- Developing a good and scalable HW architecture (PPA, features)
- Pushing their software ecosystem, e.g. push FSR like Nvidia pushes DLSS
- SW is not limited to gaming focused stuff, can be more than that (e.g. ML/AI, CAD, video rendering, ...)
With that, if AMD wants to build a GPU, they have all required building blocks.
From that point on, the rest are management and business decisions. And this is exactly the reason why the rumored RDNA5 ATx are looking like they do. Use each chip for multiple purposes.
Last edited:
Giving up on a multi-billion dollar market that is currently being vacated by the current market leader is just amazingly cowardly and short sighted. The opportunity to win goodwill in client GFX and is basically free right now. AMD can't even score an open goal now? Pathetic.
Intel are putting up a fight in client mobile solutions, so they really should try in that market instead of running the Vega playbook once again but with that trashcan architecture RDNA3, when they have better.
They are a distant second place in DCAI anyway, and if that market experiences an overdue slowdown, they'll find they'll have nothing to fall back on. Yeah compute hyperscalers will love the Italian Cities, but will that be sufficient?
adroc_thurston
Diamond Member
- Jul 2, 2023
- 8,570
- 11,297
- 106
The current market leader ships a 750mm^2 die to win.
AMD cucks outta Navi4c. that's all you gotta know.
'goodwill' is worthless garbage.
Unless you think loss leader Polaris was a sound business strategy.
The 'goal' locked itself down 18 years ago on RV770 launch.
No way back.
They don't, BMG PPA is atrocious.
They do, hence why AT3/4.
Hell no.
GPU capex down == CPU capex back up. Pretty simple.
It's been sufficient for years, and SP8 Venice will also go for enteprise share en mass. Why not?
Lol, if it takes a big chip, make one. Not making one demonstrates a severe lack of cojones, which is what I've been saying this whole time. Bunch of butt-hurt pussies.
Doesn't take a Polaris. Just takes not taking the actual piss out of whatever customers you do have. Get FSR4 working on RDNA3 for starters. We know it can be done. Fucking do it.
You can say goodwill is worthless garbage, but its pretty much the exact reason Nvidia has been sharting all over AMD/ATi, even with objectively worse products on offer. If 'goodwill' doesn't float your boat as far as vocabulary is concerned, use 'mindshare'. Same shit at the end of the day.
The thought process of the spineless.
Customers don't care about PPA. They care about products. Can Intel sustain it? Fuck if I care, but in the mean time, its just stalling AMD out of a very profitable market.
Not enough though is it. Where's the big lad? Why is RDNA3.5 still being used in APUs? Stop buying back stock and invest it in more engineers. There's an extra $6 billy there they could have used. I know you shareholders love a good stock pump, but its not actually that useful for the business.
Given Nvidia has about 4x AMD's revenue while being basically just a GPU vendor..... Yeah I dunno about that.
Still leaving billions upon billions on the table. And why? Cause they're butthurt about not succeeding in GPU? They're not the cash-strapped underdog any more. $6 billion in share buybacks, and you're telling me that they can't invest in enough validation engineers to support a complete client gfx/apu product stack AND DCAI at the same time? Lol
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 24K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter